CN118710670A - Multi-angle contour point extraction method and storage medium based on 3D engine - Google Patents

Abstract

The present application discloses a multi-angle contour point extraction method based on a 3D engine and a storage medium, the method comprising: obtaining a target model; determining a target 3D engine and a target 3D engine camera corresponding to the target model; photographing the target model at a target shooting angle through the target 3D engine camera to obtain a target image; performing pixel point traversal on the target image to obtain a first target contour point set; splitting the first target contour point set according to a target recursive algorithm to obtain n contour point sets; when n is greater than 1, using a ray crossing method to determine the inner and outer contours of the n contour point sets to obtain inner and outer contour data; calculating the area of a polygon formed by contour points in the n contour point sets based on the Green's formula to obtain n area data; integrating and deriving the n contour point sets, the inner and outer contour data and the n area data to obtain target contour point data. The embodiment of the present application improves the efficiency of contour point extraction.

Description

Multi-angle contour point extraction method and storage medium based on 3D engine

Technical Field

The present application relates to the field of image processing technology, and in particular to a multi-angle contour point extraction method and storage medium based on a 3D engine.

Background Art

With the continuous development of artificial intelligence technology, large models are increasingly used in the field of computer vision. Computer vision is a technology that simulates the human visual system and uses image processing, feature extraction, pattern recognition and other technical means to understand and analyze image content. Large models refer to deep learning models with a large number of model parameters and a huge amount of training data. Contours are useful tools for shape analysis and object detection and recognition, so a large number of contour point sets are required for large model training, which requires the use of contour point extraction technology.

Currently, most contour extraction technologies are used to extract contour points from models or images at fixed angles, such as contour extraction method, region growing method, etc. This is because they all process images, and the images are taken in advance and cannot be changed. Therefore, contour extraction technology can only process images at this angle, resulting in less contour point data and low extraction efficiency. Therefore, the problem of how to improve the extraction efficiency of contour extraction technology needs to be solved urgently.

Summary of the invention

The embodiment of the present application provides a multi-angle contour point extraction method and storage medium based on a 3D engine. By obtaining a 3D model of an object to be photographed, a 3D engine camera can be used to photograph the 3D model, thereby achieving 360-degree shooting without blind spots, and obtaining images of the 3D model at multiple angles. Then, contours are extracted from the images to obtain a large amount of contour point data, which greatly improves the extraction efficiency of the contour extraction technology.

In a first aspect, an embodiment of the present application provides a multi-angle contour point extraction method based on a 3D engine, the method comprising:

Acquire a target model; the target model is a 3D model;

Determine a target 3D engine and a target 3D engine camera corresponding to the target model;

The target model is photographed by the target 3D engine camera at a target shooting angle to obtain a target image; the target shooting angle is any angle in 360 degrees;

Perform pixel traversal on the target image to find all contour points therein to obtain a first target contour point set;

The first target contour point set is split according to the target recursive algorithm to obtain n contour point sets; n is an integer greater than or equal to 1;

When n is greater than 1, the inner and outer contours of the n contour point sets are determined by using a ray intersection method to obtain inner and outer contour data; the inner and outer contour data include contour attributes of each contour point set in the n contour point sets; the contour attribute is an inner contour or an outer contour;

Calculate the area of the polygon formed by the contour points in the n contour point set based on Green's formula to obtain n area data;

The n contour point sets, the inner and outer contour data and the n area data are integrated and exported to obtain target contour point data.

In a second aspect, an embodiment of the present application provides a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program for electronic data exchange, wherein the computer program enables a computer to execute some or all of the steps described in the first aspect of the embodiment of the present application.

In a third aspect, an embodiment of the present application provides an electronic device, comprising a processor, a memory, a communication interface, and one or more programs, wherein the one or more programs are stored in the memory and configured to be executed by the processor, and the program includes instructions for executing the steps in the first aspect of the embodiment of the present application.

In a fourth aspect, an embodiment of the present application provides a computer program product, wherein the computer program product includes a non-transitory computer-readable storage medium storing a computer program, and the computer program is operable to cause a computer to execute some or all of the steps described in the first aspect of the embodiment of the present application. The computer program product may be a software installation package.

It can be seen that the multi-angle contour point extraction method based on the 3D engine provided in the embodiment of the present application can achieve 360-degree shooting of the target model without blind spots by shooting the target model with the target 3D engine camera in the target 3D engine, and obtain the target image of the target model at multiple angles. Then, the contour is extracted from the target image to obtain a large amount of contour point data, thereby improving the extraction efficiency of the contour extraction technology.

BRIEF DESCRIPTION OF THE DRAWINGS

In order to more clearly illustrate the technical solutions in the embodiments of the present application or the background technology, the drawings required for use in the embodiments of the present application or the background technology will be described below.

FIG1 is a flow chart of a multi-angle contour point extraction method based on a 3D engine provided in an embodiment of the present application;

FIG2 is a schematic diagram of a target image provided in an embodiment of the present application;

FIG3 is a schematic diagram of a first target contour point set provided by an embodiment of the present application;

FIG4 is a schematic diagram of another first target contour point set provided by an embodiment of the present application;

FIG5 is a schematic diagram of the principle of a cross-ray method provided in an embodiment of the present application;

FIG6 is a schematic diagram of a print pop-up window provided in an embodiment of the present application;

FIG. 7 is a schematic diagram of the structure of an electronic device provided in an embodiment of the present application.

DETAILED DESCRIPTION

In order to enable those skilled in the art to better understand the solution of the present application, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

The terms "first", "second", etc. in the specification and claims of this application and the above-mentioned drawings are used to distinguish different objects, rather than to describe a specific order. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions. For example, a process, method, system, product or device that includes a series of steps or units is not limited to the listed steps or units, but optionally includes steps or units that are not listed, or optionally includes other steps or units inherent to these processes, methods, products or devices.

Reference to "embodiments" herein means that a particular feature, structure, or characteristic described in conjunction with the embodiments may be included in at least one embodiment of the present application. The appearance of the phrase in various locations in the specification does not necessarily refer to the same embodiment, nor is it an independent or alternative embodiment that is mutually exclusive with other embodiments. It is explicitly and implicitly understood by those skilled in the art that the embodiments described herein may be combined with other embodiments.

The electronic devices described in the embodiments of the present application may include smart phones (such as Android phones, iOS phones, Windows Phone phones, etc.), tablet computers, PDAs, laptops, video matrices, monitoring platforms, mobile Internet devices (mobile internet devices, MID) or wearable devices, etc. The above are only examples and not exhaustive, including but not limited to the above devices. Of course, the above electronic devices can also be servers, for example, cloud servers.

The following is an explanation of some professional terms involved in this application:

3D Engine: In simple terms, a 3D engine is a software tool used to create and render graphics in three dimensions. It allows developers to create virtual worlds and present them to users through computer screens or other devices. The 3D engine is able to manage and organize the various elements in the virtual scene, including objects, lighting, cameras, etc. The 3D engine can also use advanced graphics algorithms to project objects in the 3D scene onto a two-dimensional screen to present realistic visual effects.

Model: The model referred to in this application refers to a three-dimensional model. A three-dimensional model is a polygonal representation of an object, usually displayed by a computer or other video device. The displayed object can be a real-world entity or a fictional object. Anything that exists in the physical nature can be represented by a three-dimensional model. In a 3D engine, a model usually consists of a mesh and a texture map.

Contour points: Contours can be simply interpreted as curve polygons connecting all consecutive points (along the boundary) with the same color or intensity. Contours are useful tools for shape analysis and object detection and recognition. Contour points are a set of points on a curve polygon, usually represented by a list of two-dimensional vectors Vector2D (x, y). Contour points can be divided into outer contour points and inner contour points.

With the continuous development of artificial intelligence technology, large models are increasingly used in the field of computer vision. Computer vision is a technology that simulates the human visual system and uses image processing, feature extraction, pattern recognition and other technical means to understand and analyze image content. Large models refer to deep learning models with a large number of model parameters and a huge amount of training data. Contours are useful tools for shape analysis and object detection and recognition, so a large number of contour point sets are required for large model training, which requires the use of contour point extraction technology.

Currently, most contour extraction technologies are used to extract contour points from models or images at fixed angles, such as contour extraction method and region growing method. This is because they all process images, and the images are taken in advance and cannot be changed. Therefore, contour extraction technology can only process images at this angle, resulting in less contour point data and low extraction efficiency.

To this end, the present application provides a multi-angle contour point extraction method and storage medium based on a 3D engine. By obtaining a 3D model of the object to be photographed, the 3D model can be photographed using a 3D engine camera, which can achieve 360-degree shooting without blind spots and obtain images of the 3D model at multiple angles. Then, contour extraction is performed on the image to obtain a large amount of contour point data, which greatly improves the extraction efficiency of the contour extraction technology.

Please refer to FIG. 1 , which is a flow chart of a multi-angle contour point extraction method based on a 3D engine provided in an embodiment of the present application. The method includes but is not limited to the following steps:

S101, obtaining a target model; the target model is a 3D model.

In the embodiment of the present application, the target model may include one of the following: a person model, an animal model, an object model, etc., which is not limited here.

In a specific embodiment, it is possible to first determine in which usage scenario the contour point data extracted from the target model will be used to obtain the target usage scenario, select the target model according to the target usage scenario, and then design and produce the model through 3D modeling software, or download a model that meets the requirements from an online model library to obtain the target model. For example, the contour point data extracted from the target model will be used to train a large object recognition model. The data required by the large object recognition model is the contour point data of the object, not the contour point data of the person. In this case, the target model can be an object model, such as a 3D model of a bottle.

S102: Determine a target 3D engine and a target 3D engine camera corresponding to the target model.

In the embodiment of the present application, the target 3D engine may include at least one of the following: Unity, Unreal Engine, Cocos Creato, etc., which are not limited here; the target 3D engine may include a target 3D engine camera.

In a specific embodiment, different 3D engines can be selected according to the complexity of the target model. For example, assuming that the target model is a car model, the car model is a relatively complex model because the car itself is a complex object. In order to make the target model better simulate the appearance of a real car, it is necessary to select a 3D engine with good rendering effect. Then, the target 3D engine can be the Unity 3D engine. Then, since the Unity 3D engine generally has its own camera component, the camera component can be used as the target 3D engine camera, or a camera component can be added from the outside as the target 3D engine camera.

It should be explained that the method of selecting a 3D engine described above is only one method. Other methods can also be used in actual applications, for example, selecting a suitable 3D engine based on the cost of simulating the target model in the 3D engine, etc., which will not be repeated here.

S103, photographing the target model at a target shooting angle through the target 3D engine camera to obtain a target image; the target shooting angle is any angle within 360 degrees.

In the embodiment of the present application, the target shooting angle can be determined first, and then the image of the target model at the target shooting angle is captured by the target 3D engine camera to obtain the target image.

Optionally, step S103, photographing the target model at a target shooting angle by the target 3D engine camera to obtain a target image, may include the following steps:

31. Importing the target model into the target 3D engine, and placing the target model in a preset scene in the target 3D engine;

32. Determine the initial angle of the target 3D engine camera relative to the target model;

33. Obtain target demand data of target objects;

34. Determine the target shooting angle according to the target requirement data;

35. Determine a target adjustment parameter of the target model based on the target shooting angle and the initial angle;

36. Adjust the position of the target model according to the target adjustment parameter so that the target model is at the target shooting angle;

37. Modify the texture map of the target model to a preset color, and determine the target level of the target model in the target 3D engine;

38. Adjust the initial shooting level parameters of the target 3D engine camera according to the target level to obtain target shooting level parameters;

39. Shoot the target model at the target shooting angle using the target 3D engine camera with the target shooting level parameter to obtain the target image.

In the embodiments of the present application, the preset scenes and preset colors can be preset or defaulted in advance; the target requirement data may include at least one of the following: user preference data, requirement data for training large models, etc., which are not limited here; the target adjustment parameters may include at least one of the following: rotation matrix parameters, translation vector parameters, etc., which are not limited here.

In a specific embodiment, the target model can be first imported into the target 3D engine, and the target model can be placed in a preset scene in the target 3D engine. Specifically, the target model can be loaded into the engine through the import function of the target 3D engine, and then the above-mentioned preset scene can be found or created in the engine, and the target model can be placed in a suitable position in the preset scene (for example, the center of the scene), and the target 3D engine camera can be used to capture the full view of the target model.

It needs to be explained that in order to make the shooting effect of the target 3D engine camera better, in general, the preset scene is an empty scene. The empty scene is a scene with no other content or elements except the target model to be shot.

Next, the initial angle of the target 3D engine camera with respect to the target model can be determined. Specifically, the first position of the target model in the target 3D engine and the second position of the target 3D engine camera in the target 3D engine can be obtained first, and then the initial angle can be determined based on the first position and the second position. Further, the target requirement data of the target object can be obtained; the target shooting angle can be determined based on the target requirement data. Specifically, the target requirement data can be requirement data manually imported into the target 3D engine by a user, and the target requirement data can be requirement data for training a large model. Then, the target 3D engine camera can be determined based on the target requirement data to determine the angles from which the target 3D engine camera needs to shoot, that is, the target shooting angle. For example, assuming that the large model is an object recognition large model, in order to train it, object data of the front view of the object is generally required, and then the target shooting angle can be the front view.

Furthermore, target adjustment parameters of the target model can be determined based on the target shooting angle and the initial angle. Specifically, the difference between the target shooting angle and the initial angle can be determined to obtain the target angle difference. The mapping relationship between the preset angle difference and the adjustment parameter can be pre-stored, and the target adjustment parameters corresponding to the target angle difference can be determined based on the mapping relationship. Then, the position of the target model can be adjusted according to the target adjustment parameters so that the target model is at the target shooting angle.

It should be explained that the above description is about adjusting the position of the target model. In practical applications, the position of the target 3D engine camera can also be adjusted.

Next, the texture map of the target model can be modified to a preset color, and the target level of the target model in the target 3D engine can be determined. Specifically, the target model can be edited through the texture editor of the target 3D engine, and the texture map of the target model can be modified to a preset color, which can be pre-set or default. For example, the preset color can be black by default. Next, the level management component can be found in the target 3D engine, and the level of the target model can be viewed from the level management component to obtain the target level. Further, the initial shooting level parameters of the target 3D engine camera can be adjusted according to the target level to obtain the target shooting level parameters. Specifically, the initial shooting level parameters of the target 3D engine camera can be adjusted to a shooting level corresponding to the target level to obtain the target shooting level parameters. In this way, the target 3D engine camera can only take pictures of the level where the target model is located, eliminating interference from other levels.

Finally, the target model at the target shooting angle can be photographed by the target 3D engine camera of the target shooting level parameter to obtain the target image.

To give an example, please refer to Figure 2, which is a schematic diagram of a target image provided in an embodiment of the present application. Assuming that the target model is a bottle, the target model is placed in a preset scene (empty scene) and photographed. The obtained target image is shown in Figure 2. The target image only contains a black bottle, and the rest are blank colors. This is the empty scene.

It should be explained that the pixels of the above-mentioned target image can be fixed pixels, that is, the color value of the pixel will not change under any circumstances, that is, the target image will not change its inherent color due to lighting changes, image rotation, scaling or other image processing operations.

In this way, the target model is imported into the target 3D engine and placed in the preset scene; the initial angle of the target 3D engine camera for the target model is determined; the target requirement data of the target object is obtained; the target shooting angle is determined according to the target requirement data; the target adjustment parameters of the target model are determined based on the target shooting angle and the initial angle; the position of the target model is adjusted according to the target adjustment parameters so that the target model is at the target shooting angle; the texture map of the target model is modified to a preset color, and the target level of the target model in the target 3D engine is determined; the initial shooting angle of the target 3D engine camera is adjusted according to the target level hierarchical parameters, and obtain target shooting hierarchical parameters; the target model at the target shooting angle is photographed by the target 3D engine camera with the target shooting hierarchical parameters to obtain the target image. On the one hand, by accurately determining the angle and adjusting the position of the target model, it can be ensured that the target model is photographed at the angle that best meets the requirements, thereby accurately presenting certain specific features or states and obtaining a more targeted target image. On the other hand, by clarifying the hierarchy of the model in the target 3D engine and adjusting the initial shooting hierarchy parameters of the camera accordingly, it is helpful to reasonably arrange the hierarchical relationship of the elements in the scene, avoid occlusion or confusion, and make the shooting result more hierarchical.

S104 , performing pixel traversal on the target image to find all contour points therein and obtain a first target contour point set.

In the embodiment of the present application, pixel traversal can be performed on the target image, and contour point judgment can be performed once for each pixel in the target image, so that all contour points therein can be found to obtain a first target contour point set.

Optionally, step 104, traversing pixels of the target image to find all contour points therein to obtain a first target contour point set, may include the following steps:

Step A1: Acquire a first pixel point in the target image;

Step A2: determining i pixel points around the first pixel point; i is any number among 2, 4, and 8;

Step A3: Compare the colors of the first pixel with each of the i pixel points to obtain j pixel points with different colors from the first pixel point; j is a natural number less than or equal to i;

Step A4: when the j pixel points are greater than a specified value, the first pixel point is determined to be a contour point, and the first pixel point is added to an initial contour point set; the initial contour point set is initially an empty list starting with 0;

Based on a preset traversal algorithm, pixel points in the target image are traversed starting from the first pixel point, and the above steps A1-A4 are repeated each time a pixel point is traversed until the target image is completely traversed to obtain the first target contour point set.

In the embodiment of the present application, the designated value points may be preset or defaulted.

In a specific embodiment, the first pixel point in the target image can be obtained first. Specifically, the target image can be converted into a pixel matrix. The first pixel point can be a pixel point at a fixed position in the target image. For example, the pixel point at the top of the upper left corner of the target image can be used as the first pixel point (0,0). Then, i pixel points around the first pixel point can be determined. The points obtained are different depending on the size of i. For example, when i is 2, the i pixel points around the first pixel point can be point one (0,1) and point two (1,0). Then, the color of the first pixel point can be compared with each of the i pixel points to obtain j pixel points with a different color from the first pixel point.

Furthermore, when the number of j pixel points is greater than a specified value, the first pixel point can be determined as a contour point, and the first pixel point can be added to the initial contour point set. Specifically, the coordinates of the first pixel point can be added to the initial contour point set, that is, the initial contour point set is a two-dimensional list.

Finally, the pixel points in the target image can be traversed starting from the first pixel point based on the preset traversal algorithm, and the above steps A1-A4 are repeated each time a pixel point is traversed until the target image is traversed and the first target contour point set is obtained. Specifically, the preset traversal algorithm can be "row-first traversal", traversing the pixel points of the above target image row by row and column by column, and determining whether each pixel point is a contour point, to obtain all contour points in the above target image, that is, the first target contour point set.

Optionally, the above step A3, comparing the color of the first pixel with each of the i pixel points to obtain j pixel points having a different color from the first pixel point, may include the following steps:

A31. Obtain the RGBA color value of the first pixel to obtain a first RGBA value;

A32. Determine the color attribution of the first pixel according to the first RGBA value to obtain a first color attribution;

A33, obtaining the RGBA color value of each pixel in the i pixels, to obtain i RGBA values;

A34, determining the color attribution of each pixel in the i pixels according to the i RGBA values, to obtain i color attributions;

A35, comparing the i color attributions with the first color attribution in sequence, to obtain j color attributions different from the first color attribution;

A36. Determine the j pixel points corresponding to the j colors.

In the embodiment of the present application, the RGBA color value is a color model used to represent colors in computer graphics and image processing, wherein R represents the red channel, G represents the green channel, B represents the blue channel, and A represents the transparent channel, with a value range of 0 to 255; color attribution may include one of the following: red, green, blue, black, etc., which is not limited here.

In a specific embodiment, the RGBA color value of the first pixel may be first obtained to obtain a first RGBA value; the color attribution of the first pixel may be determined according to the first RGBA value to obtain the first color attribution. Specifically, the first pixel may be processed by an image processing library (for example, the Pillow library in Python), and the RGBA color value corresponding to the first pixel may be found according to the image processing library to obtain the first RGBA value. Then, the first RGBA value may be classified. Specifically, a "threshold method" may be used to set a threshold of an RGBA range for each color, determine within which range the first RGBA value is, determine the color corresponding to the range as the color attribution of the first pixel, and obtain the first color attribution. For example, assuming that the RGBA range of black is that the A channel value is greater than 100, and the values of other channels are all less than 100, and the first RGBA value is (0, 0, 0, 255), it is obvious that the first RGBA value is within the threshold range of black, that is, the first color attribution is black.

It should be explained that the "threshold method" described above is only a method for determining color ownership provided by this application. In practical applications, other methods can also be used. For example, the "color contrast method" can establish a color library and compare the RGBA value of the pixel with the color table to determine its color ownership.

Furthermore, the RGBA color value of each pixel in the i pixels can be obtained to obtain i RGBA values; the color attribution of each pixel in the i pixels can be determined according to the i RGBA values to obtain i color attributions; similarly, the method for obtaining the i RGBA values and the i color attributions can be the same as the acquisition method described above; then, the i color attributions can be compared with the first color attribution in turn to obtain j color attributions different from the first color attribution, and then, the j pixels corresponding to the j color attributions can be determined.

It should be explained that, in some extreme cases, no points with different color attributes are found in the entire target image, that is, all points in the target image have the same color attribute, then it is determined that the image has no contour and it is necessary to change the angle for the next shot.

In this way, the first RGBA value is obtained by obtaining the RGBA color value of the first pixel; the color attribution of the first pixel is determined according to the first RGBA value to obtain the first color attribution; the RGBA color value of each pixel in i pixels is obtained to obtain i RGBA values; the color attribution of each pixel in i pixels is determined according to the i RGBA values to obtain i color attributions; the i color attributions are compared with the first color attribution in turn to obtain j color attributions different from the first color attribution; j pixels corresponding to the j color attributions are determined, and if the j pixels are greater than a specified value, the first pixel is judged to be a contour point. On the one hand, by comparing the difference between the RGBA values of the above j pixels and the RGBA value of the first pixel, it is possible to accurately determine which pixels are different in color from the first pixel, thereby determining the contour point. On the other hand, by comparing the color attributions of multiple pixels, it is possible to reduce misjudgments caused by noise or abnormality of a single pixel, thereby improving the accuracy of contour recognition.

S105 . Split the first target contour point set according to a target recursive algorithm to obtain n contour point sets; n is an integer greater than or equal to 1.

In the embodiment of the present application, the target recursive algorithm may be preset or default.

In a specific embodiment, the first target contour point set may be split according to a target recursive algorithm to obtain n contour point sets.

To give an example, please refer to Figure 3, which is a schematic diagram of a first target contour point set provided in an embodiment of the present application. It can be seen that the figure formed by the first target contour point set 300 in Figure 3 is a circle, the color is black, wherein point 31 is a contour point in the first target contour point set 300.

It should be explained that the first target contour point set 300 only includes points on the edge of the circle in FIG. 3 , and the points inside the circle are not contour points.

For another example, please refer to Figure 4, which is a schematic diagram of another first target contour point set provided in an embodiment of the present application. It can be seen that the figure composed of the first target contour point set 300 in Figure 3 is a circle 41 and a triangle 42, the color of the circle 41 is black, and the color of the triangle 42 is white; this shows that the first target contour point set 300 is actually a contour point set containing two parts, the circle 41 and the triangle 42. Then, in the splitting process below, the first target contour point set 300 can be split into two contour point sets, and so on, it can be known how many contour point sets can be split into other different point sets.

It should be explained that the first target contour point set 300 only includes points on the outer edge of the circle 41 and the outer edge of the triangle 42 , and the interior of the circle 41 and the triangle 42 are not contour points.

Optionally, in step S105, the first target contour point set includes m points; m is a positive integer greater than or equal to n; the first target contour point set is split according to the target recursive algorithm to obtain n contour point sets, including:

Step B1: obtaining the contour point ranked first in the first target contour point set, and obtaining point p;

Step B2: Calculate the distances between the point p and all points in the first target contour point set except the point p to obtain m-1 distances;

Step B3: if the m-1 distances are all less than the preset distance, the point p is removed from the first target contour point set, and the point p is added to a new point set to obtain a first contour point set LP1; the first contour point set LP1 is one of the n contour point sets;

Step B4: repeating the above steps B1-B3 for all points except point p in the list order of the first target contour point set, repeating m-1 times, to obtain the final first contour point set LP1, and the second target contour point set after the first contour point set LP1 is split from the first target contour point set;

If the number of contour points in the second target contour point set is greater than 3, the second target contour point set is split according to the above steps B1-B4 until all the points in the second target contour point set are taken or the number of remaining points is less than 3, so as to obtain the n contour point sets.

In the embodiment of the present application, the preset distance may be pre-set or default; the preset distance is usually the distance between two discontinuous points, which is a fixed value square root of 2, for example, the distance from (0, 1) to (1, 0), square root of 2.

In a specific embodiment, the contour point ranked first in the first target contour point set can be obtained to obtain point p. Specifically, because the first target contour point set is a list, the list is ordered, and point p can be found from the list. Then, the distance between point p and all points in the first target contour point set except point p can be calculated to obtain m-1 distances. Specifically, the coordinates of point p and other points can be obtained to obtain the coordinates of point p (x1, y1), which can be calculated according to the distance calculation formula between two points:

Among them, d is any distance among m-1 distances; coordinates (x2, y2) are any point in the first target contour point set except point p; m-1 distances can be calculated m-1 times according to the above formula to obtain m-1 distances. Then, the size relationship between these m-1 distances and the preset distance is determined. If these m-1 distances are all less than the preset distance, it means that the point is one of the continuous point sets, then point p is removed from the first target contour point set, and point p is added to the new point set to obtain the first contour point set LP1. On the contrary, if these m-1 distances are not all less than the preset distance, then point p is retained in the above first target contour point set.

Further, the above steps B1-B3 can be repeated for all points except point p in the list order of the first target contour point set, and repeated m-1 times. In each repetition, points that meet the conditions (i.e., the distances between the points and the other points are less than the preset distances) are added to the first contour point set LP1 to obtain the final first contour point set LP1, and the second target contour point set after the first contour point set LP1 is split from the first target contour point set;

It should be explained that m in the above text refers to the number of points included in the first target contour point set, and m will change as the splitting process proceeds.

If the number of contour points in the second target contour point set is greater than 3, since 3 points can form a simplest polygon, the second target contour point set can be further split according to the above steps B1-B4. Each time the split is performed, a new contour point set can be obtained until all the points in the second target contour point set are taken or the number of remaining points is less than 3, so as to obtain n contour point sets.

S106. When n is greater than 1, the inner and outer contours of the n contour point sets are determined by using a ray intersection method to obtain inner and outer contour data; the inner and outer contour data include contour attributes of each contour point set in the n contour point sets; the contour attributes are inner contour or outer contour.

In the embodiment of the present application, the shot intersection method is a method for determining the inner and outer contours of an object in an image.

In a specific embodiment, when n is greater than 1, the ray intersection method can be used to determine the inner and outer contours of n contour point sets, thereby obtaining the inner and outer contour data. This is because if n is equal to 1, it means that there is only one contour point set at this time, and there is no need to distinguish between the inner and outer contours. One contour point set can only be the outer contour.

Optionally, step S106, using a ray intersection method to determine the inner and outer contours of the n contour point sets to obtain inner and outer contour data, includes:

Step C1: obtaining point A in the second contour point set LP2; the second contour point set LP2 is any point set in the n contour point sets;

Step C2: determining the positional relationship between the point A and the point sets in the n contour point sets except the second contour point set LP2 by the ray intersection method, and obtaining n-1 positional relationships;

Step C3: determining the first contour attribute data of the second contour point set LP2 according to the n-1 positional relationships;

Repeat the above steps C1-C3 once for each contour point set among the n contour point sets except the second contour point set LP2, to obtain n-1 contour attribute data;

The inner and outer contour data are determined based on the first contour attribute data and the n-1 contour attribute data.

In the embodiment of the present application, the positional relationship between a point and a contour point set may include an internal relationship or an external relationship, but does not include a boundary relationship. A boundary relationship means that the point is located on the edge of a polygon formed by the contour point set, in other words, the point is a point in the contour point set.

In a specific embodiment, point A in the second contour point set LP2 can be obtained first. Specifically, point A can be obtained in the list order of the second contour point set LP2 to obtain point A that is first in the list order. Of course, it can also be obtained randomly. Then, the positional relationship between point A and the point sets in the n contour point sets except the second contour point set LP2 can be determined by the ray intersection method to obtain n-1 positional relationships. Then, the first contour attribute data of the second contour point set LP2 can be determined based on the n-1 positional relationships. For example, the contour attributes between the second contour point set LP2 and the other n-1 contour point sets can be determined based on the n-1 positional relationships to obtain n-1 contour attributes, that is, the first contour attribute data includes the n-1 contour attributes.

To illustrate, assuming n=2, the second contour point set LP2 corresponds to one positional relationship, then a contour attribute can be directly determined based on the positional relationship to obtain first contour attribute data, that is, the first contour attribute data only includes one contour attribute; assuming n=3, the second contour point set LP2 corresponds to two positional relationships, two contour attributes can be determined based on the two positional relationships to obtain first contour attribute data, that is, the first contour attribute data only includes two contour attributes. It can be seen that the first contour attribute data includes the contour attributes of the second contour point set LP2 and each contour point set in other contour point sets.

It should be explained that the positional relationship being an internal relationship indicates that the contour attribute of the second contour point set LP2 is an inner contour, and the positional relationship being an external relationship indicates that the contour attribute of the second contour point set LP2 is an outer contour.

Furthermore, the above steps C1-C3 can be repeated once for each contour point set among the n contour point sets except the second contour point set LP2 to obtain n-1 contour attribute data; the inner and outer contour data are determined based on the first contour attribute data and the n-1 contour attribute data. Specifically, the contour attributes in the first contour attribute data and the n-1 contour attribute data can be sorted, and the duplicate data therein can be removed to obtain the above inner and outer contour data. This is because in the process of repeating the above steps C1-C3 to obtain n-1 contour attribute data, the contour point set that has already obtained the contour attribute data will be repeatedly obtained. Taking the second contour point set LP2 as an example, when the steps C1-C3 are performed for the first time, the contour attributes between the second contour point set LP2 and another contour point set A have been obtained. However, when the above steps C1-C3 are repeated for the contour point set A later, the contour attributes between the contour point set A and the second contour point set LP2 will be repeatedly obtained again, which generates duplicate data that needs to be removed.

Optionally, step C2, the ray intersection method determines the positional relationship between the point A and a third contour point set LP3, wherein the third contour point set LP3 is any point set among the n contour point sets except the second contour point set LP2, may include the following steps:

D1. Establish the target rectangular coordinate system;

D2, determining the position coordinates of the point A and the third contour point set LP3 in the target rectangular coordinate system, and obtaining the coordinate A and the coordinate set LP3; the coordinate set LP3 is a list, and the list order of the coordinate set LP3 is the same as the order of the third contour point set LP3;

D3, initialize the target counter counter to 0 and the first list index r of the first polygon to 1; the target counter counter is used to record the number of intersections between the ray A1 drawn from the point A to the positive infinity direction and the edge of the first polygon; the first polygon is the polygon formed by the coordinate set LP3;

D4. traverse the intersections of each edge of the first polygon and the ray A1 in sequence according to the first list index r, and increase the value of the target counter counter by 1 each time an intersection is detected;

D5. If the final value of the target counter counter is an even number, determine that the positional relationship between the point A and the third contour point set LP3 includes an external relationship;

D6. If the final value of the target counter counter is an odd number, determine that the positional relationship between the point A and the third contour point set LP3 includes an internal relationship.

In the embodiment of the present application, the target rectangular coordinate system may be a two-dimensional or three-dimensional coordinate system.

In a specific embodiment, a target rectangular coordinate system can be established first. Specifically, the target rectangular coordinate system can be a two-dimensional coordinate system. The center point of the target 3D engine can be used as the origin, and two mutually perpendicular coordinate axes, such as the x-axis and the y-axis, can be defined. The positive and negative directions of each coordinate axis are also defined to obtain the target rectangular coordinate system. Then, the position coordinates of point A and the third contour point set LP3 in the target rectangular coordinate system can be determined to obtain coordinate A and coordinate set LP3. Specifically, the relevant API or function provided by the target 3D engine (for example, the OpenGL library in Python) can be used to obtain the coordinates of point A to obtain coordinate A. Because the third contour point set LP3 contains multiple points, these multiple points correspond to multiple point coordinates, and these multiple point coordinates are arranged in the list order of the third contour point set LP3 to obtain coordinate set LP3.

Next, a target counter counter=0 can be initialized, and the first list index r corresponding to the first polygon can be set to 1; then, a ray A1 can be drawn from point A toward the positive infinity direction (for example, the positive direction of the x-axis), and the intersection points of each edge of the first polygon and the ray A1 can be traversed in sequence according to the first list index r, and the value of the target counter counter is increased by 1 each time an intersection is detected; if the value of the final target counter counter is an even number, then it is determined that the positional relationship between point A and the third contour point set LP3 includes an external relationship; if the value of the final target counter counter is an odd number, then it is determined that the positional relationship between point A and the third contour point set LP3 includes an internal relationship.

Please refer to FIG5 , which is a schematic diagram of the principle of a cross-ray method provided by an embodiment of the present application. As an example, it can be seen that FIG5 includes: a first polygon S1, point A, point B, point C, ray A1, etc., wherein the first polygon S1 is a polygon formed by the above-mentioned coordinate set LP3, and the coordinates of each vertex of the first polygon S1 constitute the above-mentioned coordinate set LP3. Point A, point B, point C, etc. all represent a point set in the n contour point sets. Taking point A as an example, the above-mentioned step of traversing the intersection of each edge of the first polygon and the ray A1 in turn according to the first list index r, and adding 1 to the value of the target counter counter each time an intersection is detected, includes:

First, two point vectors p1 and p2 can be defined. At the beginning, the first list index r=1. The first vertex coordinate in the coordinate set LP3 can be assigned to the point vector p1, and the second vertex coordinate in the coordinate set LP3 can be assigned to the point vector p2. In this way, the point vectors p1 and p2 are connected to form the first edge of the first polygon S1. Then, it can be checked whether the y coordinate of point A is within the y coordinate range of point vectors p1 and p2. For example, point A (x3, y3), p1 (x4, y4) and p2 (x5, y5), assuming that y4 is less than or equal to y5, determine whether y3 satisfies: y4≤y3≤y5. If this condition is met, it means that ray A1 may intersect with the above-mentioned first edge.

Secondly, check whether ray A1 intersects with the first edge. If so, obtain the coordinates of the intersection. Specifically, the straight line equation of the first edge can be calculated based on the coordinate information of p1 and p2. Then, the ordinate of point A can be substituted into the equation to solve the intersection coordinate (x6, y6). It should be noted that y6 is equal to y3. Next, it can be determined whether the abscissa x3 of point A satisfies: x3≤x6, and y5!＝y6. If all are satisfied, it means that ray A1 intersects with the first edge. The value of the target counter counter is increased by 1, and the value of the first list index r is also increased by 1. The modulus operation is performed on r through h to obtain r after the modulus operation, where h is the number of edges of the first polygon.

Finally, p1 and p2 are reassigned according to the latest first list index r, the rth vertex coordinate in the coordinate set LP3 is assigned to the point vector p1, and the (r+1)th vertex coordinate in the coordinate set LP3 is assigned to the point vector p2. In this way, p1 and p2 are connected to form the rth edge of the first polygon S1. Repeat the above steps to determine whether point A is connected to the rth edge until all edges of the first polygon S1 are traversed to obtain the final target counter counter, that is, the number of intersections between point A and the third contour point set LP3.

S107, calculating the area of the polygon formed by the contour points in the n contour point set based on Green's formula to obtain n area data.

In the embodiments of the present application, Green's formula is mainly used to convert the curve integral of the contour point on the plane into the area integral, thereby simplifying the solution of the problem.

In a specific embodiment, the area of each of the n contour point sets is calculated in turn based on Green's formula. The contour points in each contour point set will form a corresponding polygon. By calculating the sum of the areas of the polygons, n area data can be obtained.

Optionally, step S107, calculating the area of the polygon formed by the contour points in the fourth contour point set LP4 based on the Green's formula to obtain the target area data, wherein the fourth contour point set LP4 is any point set in the n contour point sets, may include the following steps:

71. Obtain the number q of contour points of the fourth contour point set LP4; q is a positive integer;

72. When the number of contour points q is greater than 2, initialize the target area S to 0; the target area S is used to accumulate the directed sum of the areas of the polygons formed by the points in the fourth contour point set LP4;

73. Determine a polygon formed by the fourth contour point set LP4 in the target rectangular coordinate system to obtain a second polygon;

74. Initialize the second list index k of the second polygon to 0; before using the second list index k, perform a modulo operation on the second list index k by the number of contour points q to prevent the second list index k from exceeding the range of the second polygon;

75. Iterate each edge of the second polygon according to the second list index k, calculate the area of a triangle formed by the current point, the previous point and the next point according to the Green's formula, and add the area of the triangle to the target area S;

76. Divide the target area S by 2 and take the absolute value to obtain the target area data; the target area data is one of the n area data.

In the embodiment of the present application,

In a specific embodiment, the fourth contour point set LP4 can be counted to obtain the number q of contour points of the fourth contour point set LP4; then, when the number q of contour points is greater than 2, the target area S is initialized to 0. This is because the necessary condition for forming a polygon is to have 3 or more points; if the number q of contour points is less than or equal to 2, there is no need to perform the following steps.

Next, the polygon formed by the fourth contour point set LP4 in the target rectangular coordinate system can be determined to obtain the second polygon. Specifically, the contour points in the fourth contour point set LP4 can be placed in the above-mentioned target rectangular coordinate system, and the contour points can be connected in sequence according to the list order of the fourth contour point set LP4 to obtain the second polygon; then, the second list index k of the second polygon can be initialized to 0; before each use, the second list index k can be modulo operated by the number of contour points q to prevent the second list index k from exceeding the range of the second polygon.

Furthermore, each edge of the second polygon can be iterated according to the second list index k, and the area of the triangle formed by the current point, the previous point and the next point can be calculated according to the Green's formula, and the area of the triangle is accumulated and added to the target area S. For example, assuming that the second list index k is 0, then the current point is the first point in the above-mentioned fourth contour point set LP4, the previous point is the last point in the above-mentioned fourth contour point set LP4, and the next point is the second point in the above-mentioned fourth contour point set LP4. These three points can form a triangle. Then, the area of the triangle can be calculated according to the Green's formula to obtain the first area, and the first area is accumulated and added to the above-mentioned target area S to obtain the latest target area S. This is repeated many times until each edge of the second polygon is traversed and the final target area S is obtained.

Finally, the final target area S can be divided by 2 and the absolute value is taken to obtain the target area data. This is because in the accumulation process, the area of each triangle is calculated twice (once in the forward direction and once in the reverse direction). The absolute value is taken because the large model training requires the size of the area rather than the direction. The target area data is one of the above n area data.

In this way, the number q of contour points of the fourth contour point set LP4 is obtained; when the number q of contour points is greater than 2, the target area S is initialized to 0; the polygon formed by the fourth contour point set LP4 in the target rectangular coordinate system is determined to obtain the second polygon; the second list index k of the second polygon is initialized to 0; each edge of the second polygon is iterated according to the second list index k, and the area of the triangle formed by the current point, the previous point and the next point is calculated according to the Green's formula, and the triangle area is added to the target area S; the target area S is divided by 2 and the absolute value is taken to obtain the target area data, and the area of the triangle formed by specific points is calculated by the Green's formula and added to the target area. The area of the polygon formed by the contour point set can be obtained more accurately, avoiding the errors caused by simple estimation.

S108, integrating and exporting the n contour point sets, the inner and outer contour data, and the n area data to obtain target contour point data.

In an embodiment of the present application, the above-mentioned n contour point sets, inner and outer contour data and n area data can be integrated and exported through a target 3D engine to obtain contour point data that is easy to use, that is, target contour point data.

Optionally, step S108, integrating and exporting the n contour point sets, the inner and outer contour data, and the n area data to obtain target contour point data, includes:

81. Convert the data formats of the n contour point sets, the inner and outer contour data, and the n area data into a preset data format;

82. Determine the corresponding relationship between the n contour point sets, the inner and outer contour data and the n area data to obtain a target corresponding relationship; one contour point set corresponds to one contour attribute data; one contour point set corresponds to one area data;

83. According to the target corresponding relationship, the n contour point sets, the inner and outer contour data and the n area data are combined and exported to obtain the target contour point data.

In the embodiment of the present application, the preset data format may be preset in advance or default.

In a specific embodiment, the data formats of the above-mentioned n contour point sets, inner and outer contour data and n area data can be converted into a preset data format. Specifically, the preset data format can be a json format, and all these data can be converted into the preset data format through the target 3D engine.

Next, the correspondence between the n contour point sets, the inner and outer contour data and the n area data can be determined to obtain the target correspondence. Specifically, each data may contain an identification information, and the identification information may contain information about the target model and the shooting angle. In this way, the correspondence between the above-mentioned n contour point sets, the inner and outer contour data and the n area data can be determined according to the identification information to obtain the target correspondence. Finally, the n contour point sets, the inner and outer contour data and the n area data can be matched one by one according to the target correspondence, and then these data can be merged and exported to a json file, that is, the target contour point data.

Please refer to Figure 6, which is a schematic diagram of a print pop-up window provided in an embodiment of the present application. After obtaining the above-mentioned target contour point data, the target 3D engine can also pop up a print pop-up window displaying printing information. As shown in Figure 6, the print pop-up window can include time information, "Found color attribution", "There is contour!!", "Area 27808.5", "End" and other information. Through the print pop-up window, the user can quickly understand the specific processing process of the target 3D engine for the above-mentioned target model.

It should be explained that the multi-angle contour point extraction method based on the 3D engine described above is all processed for a single target model. If multiple target models need to be processed, a different color can be assigned to the model texture map of each model. Assuming there are two models, one is a human model and the other is an object model, the human model can be assigned black and the object model can be assigned red. In this way, when processing the first target contour point set, the color can be used as a keyword to distinguish different models. In this way, the extracted contour point set includes the contour point sets of multiple models. In the target 3D engine, the angle and front and back rendering order of the human model and the object model can be changed, and the contour points can be extracted according to the multi-angle contour point extraction method based on the 3D engine in the above embodiment. A large amount of data of multiple models in different superposition states can be obtained, which helps to ensure that the training of large models can cope with different business scenarios and demand changes. Business scenarios can include multi-model superposition, superposition states with different front and back positions, superposition in motion, etc. Based on the target 3D engine, the model can also have various animations to meet the needs of users.

It can be seen that the multi-angle contour point extraction method based on the 3D engine provided in the embodiment of the present application can achieve 360-degree shooting of the target model without blind spots by shooting the target model with the target 3D engine camera in the target 3D engine, and obtain the target image of the target model at multiple angles. Then, the contour is extracted from the target image to obtain a large amount of contour point data, thereby improving the extraction efficiency of the contour extraction technology.

Please refer to Figure 7, which is a schematic diagram of the structure of an electronic device provided in an embodiment of the present application, the electronic device includes a processor, a memory, a communication interface and one or more programs, and the processor, the memory and the communication interface are interconnected via a bus. The one or more programs are stored in the memory and are configured to be executed by the processor. In the embodiment of the present application, the program includes instructions for executing the following steps:

Acquire a target model; the target model is a 3D model;

Determine a target 3D engine and a target 3D engine camera corresponding to the target model;

The target model is photographed by the target 3D engine camera at a target shooting angle to obtain a target image; the target shooting angle is any angle in 360 degrees;

Perform pixel traversal on the target image to find all contour points therein to obtain a first target contour point set;

The first target contour point set is split according to the target recursive algorithm to obtain n contour point sets; n is an integer greater than or equal to 1;

When n is greater than 1, the inner and outer contours of the n contour point sets are determined by using a ray intersection method to obtain inner and outer contour data; the inner and outer contour data include contour attributes of each contour point set in the n contour point sets; the contour attribute is an inner contour or an outer contour;

Calculate the area of the polygon formed by the contour points in the n contour point set based on Green's formula to obtain n area data;

The n contour point sets, the inner and outer contour data and the n area data are integrated and exported to obtain target contour point data.

Optionally, in the aspect of photographing the target model at a target shooting angle by the target 3D engine camera to obtain a target image, the program further includes instructions for executing the following steps:

Importing the target model into the target 3D engine, and placing the target model in a preset scene in the target 3D engine;

Determining an initial angle of the target 3D engine camera with respect to the target model;

Obtain target demand data of target objects;

Determine the target shooting angle according to the target requirement data;

Determining a target adjustment parameter of the target model based on the target shooting angle and the initial angle;

Adjusting the position of the target model according to the target adjustment parameter so that the target model is at the target shooting angle;

Modify the texture map of the target model to a preset color, and determine the target level of the target model in the target 3D engine;

adjusting the initial shooting level parameters of the target 3D engine camera according to the target level to obtain the target shooting level parameters;

The target model at the target shooting angle is photographed by the target 3D engine camera with the target shooting level parameter to obtain the target image.

Optionally, in the aspect of performing pixel traversal on the target image to find all contour points therein and obtain a first target contour point set, the program further includes instructions for executing the following steps:

Step A1: Acquire a first pixel point in the target image;

Step A2: determining i pixel points around the first pixel point; i is any number among 2, 4, and 8;

Step A3: Compare the colors of the first pixel with each of the i pixel points to obtain j pixel points with different colors from the first pixel point; j is a natural number less than or equal to i;

Step A4: when the j pixel points are greater than a specified value, the first pixel point is determined to be a contour point, and the first pixel point is added to an initial contour point set; the initial contour point set is initially an empty list starting with 0;

Based on a preset traversal algorithm, pixel points in the target image are traversed starting from the first pixel point, and the above steps A1-A4 are repeated each time a pixel point is traversed until the target image is completely traversed to obtain the first target contour point set.

Optionally, in the aspect of comparing the first pixel with each of the i pixels to obtain j pixels having a different color from the first pixel, the program further includes instructions for executing the following steps:

Obtaining the RGBA color value of the first pixel to obtain a first RGBA value;

Determine the color attribution of the first pixel according to the first RGBA value to obtain a first color attribution;

Obtain the RGBA color value of each pixel in the i pixels to obtain i RGBA values;

Determine the color attribution of each pixel in the i pixels according to the i RGBA values to obtain i color attributions;

Compare the i color attributions with the first color attribution in sequence to obtain j color attributions different from the first color attribution;

Determine the j pixel points corresponding to the j colors.

Optionally, the first target contour point set includes m points; m is a positive integer greater than or equal to n; in the aspect of splitting the first target contour point set according to the target recursive algorithm to obtain n contour point sets, the program further includes instructions for executing the following steps:

Step B1: obtaining the contour point ranked first in the first target contour point set, and obtaining point p;

Step B2: Calculate the distances between the point p and all points in the first target contour point set except the point p to obtain m-1 distances;

Step B3: if the m-1 distances are all less than the preset distance, the point p is removed from the first target contour point set, and the point p is added to a new point set to obtain a first contour point set LP1; the first contour point set LP1 is one of the n contour point sets;

Step B4: repeating the above steps B1-B3 for all points except point p in the list order of the first target contour point set, repeating m-1 times, to obtain the final first contour point set LP1, and the second target contour point set after the first contour point set LP1 is split from the first target contour point set;

If the number of contour points in the second target contour point set is greater than 3, the second target contour point set is split according to the above steps B1-B4 until all the points in the second target contour point set are taken or the number of remaining points is less than 3, so as to obtain the n contour point sets.

Optionally, in the aspect of using the ray intersection method to determine the inner and outer contours of the n contour point sets to obtain the inner and outer contour data, the program further includes instructions for executing the following steps:

Step C1: obtaining point A in the second contour point set LP2; the second contour point set LP2 is any point set in the n contour point sets;

Step C2: determining the positional relationship between the point A and the point sets in the n contour point sets except the second contour point set LP2 by the ray intersection method, and obtaining n-1 positional relationships;

Step C3: determining the first contour attribute data of the second contour point set LP2 according to the n-1 positional relationships;

Repeat the above steps C1-C3 once for each contour point set among the n contour point sets except the second contour point set LP2, to obtain n-1 contour attribute data;

The inner and outer contour data are determined based on the first contour attribute data and the n-1 contour attribute data.

Optionally, in determining the positional relationship between the point A and the third contour point set LP3 by the ray intersection method, the third contour point set LP3 is any point set among the n contour point sets except the second contour point set LP2, and the program further includes instructions for executing the following steps:

Establish target rectangular coordinate system;

Determine the position coordinates of the point A and the third contour point set LP3 in the target rectangular coordinate system to obtain coordinates A and a coordinate set LP3; the coordinate set LP3 is a list, and the list order of the coordinate set LP3 is the same as the order of the third contour point set LP3;

Initialize the target counter counter to 0 and the first list index r of the first polygon to 1; the target counter counter is used to record the number of intersections between the ray A1 drawn from the point A to the positive infinity direction and the edge of the first polygon; the first polygon is a polygon formed by the coordinate set LP3;

Traversing the intersections of each edge of the first polygon and the ray A1 in sequence according to the first list index r, and adding 1 to the value of the target counter counter each time an intersection is detected;

If the value of the target counter counter is an even number, it is determined that the positional relationship between the point A and the third contour point set LP3 includes an external relationship;

If the final value of the target counter counter is an odd number, it is determined that the positional relationship between the point A and the third contour point set LP3 includes an internal relationship.

Optionally, in the aspect of calculating the area of the polygon formed by the contour points in the fourth contour point set LP4 based on the Green's formula to obtain the target area data, the fourth contour point set LP4 is any point set in the n contour point sets, and the above program further includes instructions for executing the following steps:

Obtain the number q of contour points of the fourth contour point set LP4; q is a positive integer;

When the number of contour points q is greater than 2, the target area S is initialized to 0; the target area S is used to accumulate the directed sum of the areas of the polygons formed by the points in the fourth contour point set LP4;

Determine a polygon formed by the fourth contour point set LP4 in the target rectangular coordinate system to obtain a second polygon;

Initialize the second list index k of the second polygon to 0; perform a modulo operation on the second list index k by the number of contour points q before using the second list index k to prevent the second list index k from exceeding the range of the second polygon;

Iterate each edge of the second polygon according to the second list index k, calculate the area of a triangle formed by the current point, the previous point and the next point according to the Green's formula, and add the area of the triangle to the target area S;

The target area S is divided by 2 and the absolute value is taken to obtain the target area data; the target area data is one of the n area data.

Optionally, in the aspect of integrating and deriving the n contour point sets, the inner and outer contour data and the n area data to obtain target contour point data, the program further includes instructions for executing the following steps:

Converting the data formats of the n contour point sets, the inner and outer contour data, and the n area data into a preset data format;

Determine the correspondence between the n contour point sets, the inner and outer contour data and the n area data to obtain a target correspondence; one contour point set corresponds to one contour attribute data; one contour point set corresponds to one area data;

The n contour point sets, the inner and outer contour data and the n area data are combined and exported according to the target corresponding relationship to obtain the target contour point data.

An embodiment of the present application also provides a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program for electronic data exchange, wherein the computer program enables a computer to execute part or all of the steps of any method recorded in the above method embodiments, and the above computer includes an electronic device.

The present application also provides a computer program product, which includes a non-transitory computer-readable storage medium storing a computer program, and the computer program is operable to cause a computer to execute some or all of the steps of any method described in the method embodiment. The computer program product may be a software installation package, and the computer includes an electronic device.

It should be noted that, for the above-mentioned method embodiments, for the sake of simplicity, they are all expressed as a series of action combinations, but those skilled in the art should be aware that the present application is not limited by the order of the actions described, because according to the present application, certain steps can be performed in other orders or simultaneously. Secondly, those skilled in the art should also be aware that the embodiments described in the specification are all preferred embodiments, and the actions and modules involved are not necessarily required by the present application.

In the above embodiments, the description of each embodiment has its own emphasis. For parts that are not described in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

In the several embodiments provided in the present application, it should be understood that the disclosed devices can be implemented in other ways. For example, the device embodiments described above are only schematic, such as the division of the above-mentioned units, which is only a logical function division. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, and the indirect coupling or communication connection of devices or units can be electrical or other forms.

The units described above as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

If the above-mentioned integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable memory. Based on this understanding, the technical solution of the present application is essentially or the part that contributes to the prior art or all or part of the technical solution can be embodied in the form of a software product, and the computer software product is stored in a memory, including a number of instructions to enable a computer device (which can be a personal computer, server or network device, etc.) to perform all or part of the steps of the above-mentioned methods of each embodiment of the present application. The aforementioned memory includes: U disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), mobile hard disk, disk or CD-ROM and other media that can store program codes.

The embodiments of the present application are introduced in detail above. Specific examples are used in this article to illustrate the principles and implementation methods of the present application. The description of the above embodiments is only used to help understand the method of the present application and its core idea. At the same time, for general technical personnel in this field, according to the idea of the present application, there will be changes in the specific implementation method and application scope. In summary, the content of this specification should not be understood as a limitation on the present application.

Claims (10)

1. A multi-angle contour point extraction method based on a 3D engine, characterized in that the method comprises: Acquire a target model; the target model is a 3D model; Determine a target 3D engine and a target 3D engine camera corresponding to the target model; The target model is photographed by the target 3D engine camera at a target shooting angle to obtain a target image; the target shooting angle is any angle in 360 degrees; Perform pixel traversal on the target image to find all contour points therein to obtain a first target contour point set; The first target contour point set is split according to the target recursive algorithm to obtain n contour point sets; n is an integer greater than or equal to 1; When n is greater than 1, the inner and outer contours of the n contour point sets are determined by using a ray intersection method to obtain inner and outer contour data; the inner and outer contour data include contour attributes of each contour point set in the n contour point sets; the contour attribute is an inner contour or an outer contour; Calculate the area of the polygon formed by the contour points in the n contour point set based on Green's formula to obtain n area data; The n contour point sets, the inner and outer contour data and the n area data are integrated and exported to obtain target contour point data. 2. The method according to claim 1, wherein photographing the target model at a target shooting angle by the target 3D engine camera to obtain a target image comprises: Importing the target model into the target 3D engine, and placing the target model in a preset scene in the target 3D engine; Determining an initial angle of the target 3D engine camera with respect to the target model; Obtain target demand data of target objects; Determine the target shooting angle according to the target requirement data; Determining a target adjustment parameter of the target model based on the target shooting angle and the initial angle; Adjusting the position of the target model according to the target adjustment parameter so that the target model is at the target shooting angle; Modify the texture map of the target model to a preset color, and determine the target level of the target model in the target 3D engine; adjusting the initial shooting level parameters of the target 3D engine camera according to the target level to obtain the target shooting level parameters; The target model at the target shooting angle is photographed by the target 3D engine camera with the target shooting level parameter to obtain the target image. 3. The method according to claim 1 or 2, characterized in that the step of traversing pixels of the target image to find all contour points therein and obtaining a first target contour point set comprises: Step A1: Acquire a first pixel point in the target image; Step A2: determining i pixel points around the first pixel point; i is any number among 2, 4, and 8; Step A3: Compare the colors of the first pixel with each of the i pixel points to obtain j pixel points with different colors from the first pixel point; j is a natural number less than or equal to i; Step A4: when the j pixel points are greater than a specified value, the first pixel point is determined to be a contour point, and the first pixel point is added to an initial contour point set; the initial contour point set is initially an empty list starting with 0; Based on a preset traversal algorithm, pixel points in the target image are traversed starting from the first pixel point, and the above steps A1-A4 are repeated each time a pixel point is traversed until the target image is completely traversed to obtain the first target contour point set. 4. The method according to claim 3, wherein the step of comparing the first pixel with each of the i pixel points to obtain j pixel points having a different color from the first pixel point comprises: Obtaining the RGBA color value of the first pixel to obtain a first RGBA value; Determine the color attribution of the first pixel according to the first RGBA value to obtain a first color attribution; Obtain the RGBA color value of each pixel in the i pixels to obtain i RGBA values; Determine the color attribution of each pixel in the i pixels according to the i RGBA values to obtain i color attributions; Compare the i color attributions with the first color attribution in sequence to obtain j color attributions different from the first color attribution; Determine the j pixel points corresponding to the j colors. 5. The method according to claim 1 or 2, characterized in that the first target contour point set includes m points; m is a positive integer greater than or equal to n; the first target contour point set is split according to the target recursive algorithm to obtain n contour point sets, comprising: Step B1: obtaining the contour point ranked first in the first target contour point set, and obtaining point p; Step B2: Calculate the distances between the point p and all points in the first target contour point set except the point p to obtain m-1 distances; Step B3: if the m-1 distances are all less than the preset distance, the point p is removed from the first target contour point set, and the point p is added to a new point set to obtain a first contour point set LP1; the first contour point set LP1 is one of the n contour point sets; Step B4: repeating the above steps B1-B3 for all points except point p in the list order of the first target contour point set, repeating m-1 times, to obtain the final first contour point set LP1, and the second target contour point set after the first contour point set LP1 is split from the first target contour point set; If the number of contour points in the second target contour point set is greater than 3, the second target contour point set is split according to the above steps B1-B4 until all the points in the second target contour point set are taken or the number of remaining points is less than 3, so as to obtain the n contour point sets. 6. The method according to claim 1 or 2, characterized in that the step of using a ray intersection method to determine the inner and outer contours of the n contour point sets to obtain inner and outer contour data comprises: Step C1: obtaining point A in the second contour point set LP2; the second contour point set LP2 is any point set in the n contour point sets; Step C2: determining the positional relationship between the point A and the point sets in the n contour point sets except the second contour point set LP2 by the ray intersection method, and obtaining n-1 positional relationships; Step C3: determining the first contour attribute data of the second contour point set LP2 according to the n-1 positional relationships; Repeat the above steps C1-C3 once for each contour point set among the n contour point sets except the second contour point set LP2, to obtain n-1 contour attribute data; The inner and outer contour data are determined based on the first contour attribute data and the n-1 contour attribute data. 7. The method according to claim 6, wherein the ray intersection method determines the positional relationship between the point A and a third contour point set LP3, wherein the third contour point set LP3 is any point set among the n contour point sets except the second contour point set LP2, comprising: Establish target rectangular coordinate system; Determine the position coordinates of the point A and the third contour point set LP3 in the target rectangular coordinate system to obtain coordinates A and a coordinate set LP3; the coordinate set LP3 is a list, and the list order of the coordinate set LP3 is the same as the order of the third contour point set LP3; Initialize the target counter counter to 0 and the first list index r of the first polygon to 1; the target counter counter is used to record the number of intersections between the ray A1 drawn from the point A to the positive infinity direction and the edge of the first polygon; the first polygon is a polygon formed by the coordinate set LP3; Traversing the intersections of each edge of the first polygon and the ray A1 in sequence according to the first list index r, and adding 1 to the value of the target counter counter each time an intersection is detected; If the value of the target counter counter is an even number, it is determined that the positional relationship between the point A and the third contour point set LP3 includes an external relationship; If the final value of the target counter counter is an odd number, it is determined that the positional relationship between the point A and the third contour point set LP3 includes an internal relationship. 8. The method of claim 7, wherein the area of the polygon formed by the contour points in the fourth contour point set LP4 is calculated based on the Green's formula to obtain the target area data, wherein the fourth contour point set LP4 is any point set in the n contour point sets, and comprises: Obtain the number q of contour points of the fourth contour point set LP4; q is a positive integer; When the number of contour points q is greater than 2, the target area S is initialized to 0; the target area S is used to accumulate the directed sum of the areas of the polygons formed by the points in the fourth contour point set LP4; Determine a polygon formed by the fourth contour point set LP4 in the target rectangular coordinate system to obtain a second polygon; Initialize the second list index k of the second polygon to 0; perform a modulo operation on the second list index k by the number of contour points q before using the second list index k to prevent the second list index k from exceeding the range of the second polygon; Iterate each edge of the second polygon according to the second list index k, calculate the area of a triangle formed by the current point, the previous point and the next point according to the Green's formula, and add the area of the triangle to the target area S; The target area S is divided by 2 and the absolute value is taken to obtain the target area data; the target area data is one of the n area data. 9. The method according to claim 1 or 2, characterized in that the step of integrating and deriving the n contour point sets, the inner and outer contour data and the n area data to obtain target contour point data comprises: Converting the data formats of the n contour point sets, the inner and outer contour data, and the n area data into a preset data format; Determine the correspondence between the n contour point sets, the inner and outer contour data and the n area data to obtain a target correspondence; one contour point set corresponds to one contour attribute data; one contour point set corresponds to one area data; The n contour point sets, the inner and outer contour data and the n area data are combined and exported according to the target corresponding relationship to obtain the target contour point data. 10. A computer-readable storage medium, characterized in that it stores a computer program for electronic data exchange, wherein the computer program enables a computer to execute the method according to any one of claims 1 to 9.