CN114266816B - Dense depth map construction method, device, equipment and medium - Google Patents

Abstract

The application discloses a dense depth map construction method, a dense depth map construction device, dense depth map construction equipment and a dense depth map construction medium, which relate to the technical field of image processing, and the method comprises the following steps: acquiring an initial depth map generated by a speckle structured light system; performing expansion operation on the initial depth map by using a preset structural element to obtain a corresponding expansion depth map; the expansion operation logic corresponding to the preset structural element is that when the preset structural element traverses to any pixel point on the initial depth map, the depth value of a target depth point closest to the central pixel point of the preset structural element is determined based on the depth value of the central pixel point; carrying out corrosion operation on the expansion depth map by using the preset structural elements to obtain a dense depth map; and processing the dense depth map by using a smoothing filtering technology to obtain a dense smooth depth map. Through the scheme, the depth value of the expanded depth point can be more accurate. Thereby, the problems of depth quality degradation and depth distortion in the depth map can be avoided.

Description

A dense depth map construction method, device, equipment and medium

technical field

The present invention relates to the technical field of image processing, in particular to a dense depth map construction method, device, equipment and medium.

Background technique

At present, 3D reconstruction technology is widely used in production and life. The 3D reconstruction technology needs to perform depth measurement through camera calibration to establish an effective imaging model, and combine the image matching results to obtain the 3D point coordinates in space, so as to achieve the purpose of 3D reconstruction. In the process of depth measurement, speckle structured light can be used to calculate the depth of discrete points instead of calculating the depth of the entire image to reduce the amount of calculation. When speckle structured light is used, discrete depth points are obtained, and image processing is required to obtain a depth map of the entire image.

However, on the one hand, traditional interpolation algorithms are based on fixed-spaced pixels applied to image scaling. Commonly used interpolation algorithms for discrete points include inverse distance weighted interpolation, Kriging interpolation, and radial basis function. Except for the simple inverse distance weighted interpolation method, other discrete point interpolation functions have a large amount of calculation and are not suitable for real-time structured light depth algorithms. However, the inverse distance weighted interpolation method is easy to deal with large or small data in the interpolation process. The bull's-eye phenomenon with the interpolation point as the center is formed, causing the objects in the depth map to deform and the depth quality to decline; on the other hand, the traditional closing operation will cause depth distortion due to the coverage of depth points. Therefore, the problem of how to avoid the depth quality degradation and depth distortion of objects in the depth image in the process of constructing a dense depth image needs to be further resolved.

Contents of the invention

In view of this, the object of the present invention is to provide a dense depth map construction method, device, equipment and medium, which can avoid the problems of object depth quality degradation and depth distortion in the depth map during the process of constructing a dense depth image. The specific plan is as follows:

In the first aspect, the present application discloses a dense depth map construction method, including:

Obtain the initial depth map generated by the speckle structured light system;

Perform an expansion operation on the initial depth map by using a preset structural element to obtain a corresponding expanded depth map; the expansion operation logic corresponding to the preset structural element is when the preset structural element traverses to the initial depth map When any pixel, determine the depth value of the central pixel based on the depth value of the target depth point closest to the central pixel of the preset structure element;

performing an erosion operation on the expanded depth map by using the preset structural elements to obtain a dense depth map;

The dense depth map is processed by using a smoothing filtering technique to obtain a dense smooth depth map.

Optionally, before performing the dilation operation on the initial depth map by using the preset structural elements, the method further includes:

calculating an average distance between speckle pixels in the initial depth map;

A structural element size parameter is determined based on the average distance, and a corresponding structural element is constructed according to the structural element size parameter, so as to obtain the preset structural element.

Optionally, said constructing corresponding structural elements according to said structural element size parameters to obtain said preset structural elements includes:

Screen out the target structure element shape from the preset structure element shape library based on the preset structure element shape requirements, and construct the corresponding structure element according to the structure element size parameter and the target structure element shape, so as to obtain the preset structure element shape Structural element; the shape library of the structural element includes any one or a combination of rectangles, rhombuses, crosses and circles.

Optionally, performing an erosion operation on the expansion depth map by using the preset structural elements includes:

Use the preset structural element to traverse the expanded depth map, and when traversing to any pixel on the expanded depth map, if any pixel in the depth map area currently located inside the preset structural element If the pixel point is a blank point, then mark the central pixel point as a blank point.

Optionally, after processing the dense depth map using a smoothing filtering technique to obtain a dense smooth depth map, the method further includes:

Performing three-dimensional reconstruction based on the dense smooth depth map to obtain corresponding three-dimensional reconstruction results.

Optionally, performing an expansion operation on the initial depth map by using a preset structural element includes:

Use the preset structural element to traverse the initial depth map, and when traversing to any pixel on the initial depth map, according to the distance between the central pixel of the preset structural element is small To the largest traversal order, traversing the pixels currently located in the depth map area inside the preset structural element, if the target depth point closest to the central pixel point is traversed in the depth map area, then Determining the depth value of the central pixel point based on the depth value of the target depth point, and marking the central pixel point as a blank point if no depth point can be traversed in the depth map area.

Optionally, if the target depth point closest to the central pixel point is traversed in the depth map area, determining the depth value of the central pixel point based on the depth value of the target depth point includes:

If the target depth point closest to the central pixel point is traversed in the depth map area, the depth value of the first traversed target depth point is used as the depth value of the central pixel point;

Or, if the target depth point closest to the central pixel point is traversed in the depth map area, the average value corresponding to the depth values of all the target depth points is used as the depth value of the central pixel point;

Or, if the target depth point closest to the central pixel point is traversed in the depth map area, a depth value with the highest frequency of occurrence among the target depth points is used as the depth value of the central pixel point.

In the second aspect, the present application discloses a dense depth map construction device, including:

A depth map acquisition module, configured to obtain an initial depth map generated by a speckle structured light system;

The expansion module is used to perform an expansion operation on the initial depth map by using a preset structural element to obtain a corresponding expanded depth map; the expansion operation logic corresponding to the preset structural element is that when the preset structural element traverses to the When any pixel on the initial depth map is selected, the depth value of the central pixel is determined based on the depth value of the target depth point closest to the central pixel of the preset structural element;

An erosion module, configured to use the preset structural elements to perform an erosion operation on the expanded depth map to obtain a dense depth map;

The smoothing processing module is used to process the dense depth map by smoothing filtering technology to obtain a dense smooth depth map.

In a third aspect, the present application discloses an electronic device, comprising:

memory for storing computer programs;

The processor is configured to execute the computer program, so as to realize the steps of the dense depth map construction method disclosed above.

In a fourth aspect, the present application discloses a computer-readable storage medium for storing a computer program; wherein, when the computer program is executed by a processor, the steps of the dense depth map construction method disclosed above are implemented.

When constructing a dense depth map, the present application first obtains the initial depth map generated by the speckle structured light system, and uses preset structural elements to perform an expansion operation on the initial depth map to obtain a corresponding expanded depth map; Let the expansion operation logic corresponding to the structure element be when the preset structure element traverses to any pixel point on the initial depth map, based on the depth of the target depth point closest to the central pixel point of the preset structure element value to determine the depth value of the central pixel point, and then use the preset structure element to corrode the expanded depth map to obtain a dense depth map, and finally process the dense depth map using smoothing filtering technology to obtain A dense smooth depth map is obtained. It can be seen that when constructing a dense depth map in this application, the traditional interpolation algorithm is not used, but the morphological image processing method is used, and the expansion operation in the traditional morphological closing operation is improved. In the process of the improved expansion operation Considering the distance relationship between the central pixel of the preset structural element and the depth point to determine the expanded depth value of the pixel at the central pixel position of the preset structural element, the depth value of the expanded depth point is more accurate. Therefore, the present application avoids the bull's-eye phenomenon caused by the traditional interpolation algorithm and the problem of object deformation and depth quality degradation in the depth map when constructing a dense depth map; The problem of depth distortion caused by depth point coverage after the dilation operation in the traditional closing operation is avoided.

Description of drawings

In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings that need to be used in the description of the embodiments or the prior art. Obviously, the accompanying drawings in the following description are only It is an embodiment of the present invention, and those skilled in the art can also obtain other drawings according to the provided drawings without creative work.

Fig. 1 is a flow chart of a method for constructing a dense depth map provided by the present application;

Figure 2 is a flow chart of a specific method for constructing a dense depth map provided by the present application;

Fig. 3 is a schematic diagram of preset internal traversal order of a preset structural element provided by the present application;

FIG. 4 is a flow chart of a specific dense depth map construction method provided by the present application;

FIG. 5 is a flow chart of a specific dense depth map construction method provided by the present application;

FIG. 6 is a schematic structural diagram of a dense depth map construction device provided by the present application;

FIG. 7 is a structural diagram of an electronic device provided by the present application.

Detailed ways

The following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only some, not all, embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by persons of ordinary skill in the art without making creative efforts belong to the protection scope of the present invention.

At present, traditional interpolation algorithms are based on fixed-spaced pixels applied to image scaling. Commonly used interpolation algorithms for discrete points include inverse distance weighted interpolation, Kriging interpolation, and radial basis function. Except for the simple inverse distance weighted interpolation method, other discrete point interpolation functions have a large amount of calculation and are not suitable for real-time structured light depth algorithms. However, the inverse distance weighted interpolation method is easy to deal with large or small data in the interpolation process. The bull's-eye phenomenon with the interpolation point as the center is formed, causing the objects in the depth map to deform and the depth quality to decline; on the other hand, the traditional closing operation will cause depth distortion due to the coverage of depth points. To this end, the present application provides a dense depth map construction method that can effectively avoid the problems of object depth quality degradation and depth distortion in the depth map.

The embodiment of the present invention discloses a dense depth map construction method, as shown in Figure 1, the method includes:

Step S11: Obtain the initial depth map generated by the speckle structured light system.

In this embodiment, the speckle structured light system refers to a system composed of a camera and a speckle light projector. When performing depth measurement, a speckle light projector projects speckle structured light to the object to be measured, and the camera receives the imaging of the object to be measured under the speckle structured light. It can be understood that the depth points in the depth map generated in the speckle structured light system are composed of discrete speckle pixels. After the initial depth map of the discrete depth points is obtained, the initial depth map needs to be imaged subsequently. processed to obtain a dense depth map.

Step S12: Use the preset structural element to perform an expansion operation on the initial depth map to obtain the corresponding expanded depth map; the expansion operation logic corresponding to the preset structural element is that when the preset structural element traverses to the initial For any pixel on the depth map, the depth value of the central pixel is determined based on the depth value of the target depth point closest to the central pixel of the preset structural element.

In this embodiment, before performing the expansion operation on the initial depth map by using preset structural elements, it further includes: calculating the average distance between speckle pixels in the initial depth map; based on the distance The average value determines the size parameter of the structural element, and constructs the corresponding structural element according to the size parameter of the structural element, so as to obtain the preset structural element. It should be pointed out that after determining the structural element size parameters, the shape of the structural element needs to be determined, and the target structural element shape is selected from the preset structural element shape library based on the preset structural element shape requirements, and according to the structural element size parameters and the shape of the target structural element to construct the corresponding structural element to obtain the preset structural element; the structural element shape library includes any one or a combination of rectangles, rhombuses, crosses and circles. The above method is used to calculate the size parameters of the structural elements, so that the size of the structural elements corresponds to the initial depth map, so as to avoid the depth of the subsequent expansion and erosion treatment due to random selection of the structural element size parameters due to random selection. On the other hand, by selecting the target structure element shape from the preset structure element shape library based on the preset structure element shape requirements, the flexibility of the shape of the structure element can be guaranteed, and the preset structure element shape can be selected according to the demand. The more suitable structure element shape for this dense depth map construction is selected from the library.

Further, in this embodiment, the initial depth map is traversed by using the preset structural element, and when traversing to any pixel on the initial depth map, according to the preset structural element The distance between the central pixels of the traversal sequence from small to large, traversing the pixels currently located in the depth map area inside the preset structural element, if the distance from the center to the depth map area is traversed in the depth map area The nearest target depth point of the pixel point, then determine the depth value of the center pixel point based on the depth value of the target depth point, if the depth point cannot be traversed in the depth map area, then mark the center pixel point for the blank point. The original depth map is expanded by improving the traditional closing operation. When assigning the depth value to the central pixel of the preset structural element during the expansion process, the relationship between the central pixel of the preset structural element and the depth point in the structural element is considered. positional relationship to determine the target depth point, and determine the depth value of the central pixel based on the depth value of the target depth point, which ensures the reliability of the depth value marked by the central pixel of the preset structural element. It can be understood that the expansion operation is used to fill the gaps between the discrete pixel points in the initial depth map, and assign depth values to the gaps around the discrete depth points, so as to obtain the expanded depth map after the discrete depth map is expanded.

Step S13: using the preset structural elements to perform an erosion operation on the expanded depth map to obtain a dense depth map.

In this embodiment, the preset structural element is used to traverse the expanded depth map, and when traversing to any pixel on the expanded depth map, if the currently located inside the preset structural element If any pixel in the depth map area is a blank point, the central pixel is marked as a blank point. On the contrary, if all the pixels in the depth map area currently located inside the preset structure element are depth points, keep the depth value of the central pixel unchanged. It can be understood that the expansion depth map is corroded by an erosion operation to restore the size of the initial depth map to obtain a dense depth map.

Step S14: Process the dense depth map using a smoothing filter technique to obtain a dense smooth depth map.

In this embodiment, the dense and smooth depth map is processed by using a smoothing filtering technique to obtain a dense and smooth depth map, and three-dimensional reconstruction is performed based on the dense and smooth depth map to obtain a corresponding three-dimensional reconstruction result.

It can be seen that when constructing a dense depth map in this embodiment, instead of using the traditional interpolation algorithm, a morphological image processing method is used, and the expansion operation in the traditional morphological closing operation is improved. After the improved expansion operation In the process, the distance relationship between the center pixel of the preset structure element and the depth point is considered to determine the depth value after expansion of the pixel point at the center pixel point of the preset structure element, so that the depth value of the expanded depth point is more accurate. Therefore, the present application avoids the bull's-eye phenomenon caused by the traditional interpolation algorithm and the problem of object deformation and depth quality degradation in the depth map when constructing a dense depth map; The problem of depth distortion caused by depth point coverage after the dilation operation in the traditional closing operation is avoided.

Referring to FIG. 2 , the embodiment of the present invention discloses a specific method for constructing a dense depth map. Compared with the previous embodiment, this embodiment further explains and optimizes the technical solution.

Step S21: Obtain an initial depth map generated by the speckle structured light system.

Step S22: Use the preset structural element to traverse the initial depth map, and when traversing to any pixel on the initial depth map, follow the distance between the central pixel of the preset structural element The traversal order of the distance from small to large is to traverse the pixels currently located in the depth map area inside the preset structural element. If the target depth closest to the center pixel is traversed in the depth map area point, the depth value of the first traversed target depth point is used as the depth value of the central pixel point.

In this embodiment, during the traversal process inside the structure element, the depth value of the target depth point traversed first is used as the depth value of the central pixel point. Taking the rectangular structural element as an example, as shown in Figure 3, after the preset structural element is determined, the traversal order of the pixel points within the structural element from small to large from the center pixel of the preset structural element is set in the preset structural element, Point 1 in the figure is the center pixel point of the preset structural element, starting from point 1 according to the preset numbering sequence, then traversing through point 2 and point 3 in sequence, until point 25 is the end of the traversal. When traversing within each structural element, it is traversing according to a preset order, and the depth value of the first target depth point encountered during the traversal process is used as the depth value of the central pixel point, for example, when point 1, Points 2 and 3 are blank points, and when point 4 is a depth point, use point 4 as the target depth point, and determine the depth value of the target depth point where point 4 is located as the depth value of the central pixel point . It can be understood that when traversing according to the traversal order shown in FIG. 3 , the first encountered depth point is the depth point closest to the central pixel point.

Step S23: using the preset structural elements to perform an erosion operation on the expanded depth map to obtain a dense depth map.

Step S24: Process the dense depth map using a smoothing filter technique to obtain a dense smooth depth map.

It can be seen that in this embodiment, when assigning the central pixel point of the preset structural element during the expansion operation, the depth value of the first target depth point encountered during the traversal process is used as the depth of the central pixel point The value can speed up the traversal process, improve the efficiency of each traversal process within the structural element, and speed up the assignment of the center pixel of the preset structural element.

Referring to FIG. 4 , the embodiment of the present invention discloses a specific method for constructing a dense depth map. Compared with the previous embodiment, this embodiment further explains and optimizes the technical solution.

Step S31: Obtain an initial depth map generated by the speckle structured light system.

Step S32: Use the preset structural element to traverse the initial depth map, and when traversing to any pixel on the initial depth map, follow the distance between the central pixel of the preset structural element The traversal order of the distance from small to large is to traverse the pixels currently located in the depth map area inside the preset structural element. If the target depth closest to the center pixel is traversed in the depth map area point, the average value corresponding to the depth values of all the target depth points is used as the depth value of the central pixel point.

In this embodiment, the average value corresponding to the depth values of all the target depth points is used as the depth value of the central pixel point. Taking the rectangular structural element as an example, as shown in Figure 3, after the preset structural element is determined, the traversal order of the pixel points within the structural element from small to large from the center pixel of the preset structural element is set in the preset structural element, When traversing in each structural element, it traverses according to the preset order. If point 1, point 2 and point 5 are all blank points, and the first depth point encountered is point 6, then it is the same as point 6. Point 7, Point 8, and Point 9, whose distances are equal to the center of the preset structural element, are all judged, and the depth points in Points 6, 7, 8, and 9 are taken as the target depth points, and the average depth value is taken as the target depth point. The depth value of the center pixel.

Step S33: using the preset structural elements to perform an erosion operation on the expanded depth map to obtain a dense depth map.

Step S34: Process the dense depth map using a smoothing filter technique to obtain a dense smooth depth map.

It can be seen that, in this embodiment, the depth value of the central pixel of the preset structural element is made more accurate by using the average value corresponding to the depth values of all the target depth points as the depth value of the central pixel.

Referring to FIG. 5 , the embodiment of the present invention discloses a specific method for constructing a dense depth map. Compared with the previous embodiment, this embodiment further explains and optimizes the technical solution.

Step S41: Obtain an initial depth map generated by the speckle structured light system.

Step S42: Use the preset structural element to traverse the initial depth map, and when traversing to any pixel on the initial depth map, follow the distance between the central pixel of the preset structural element The traversal order of the distance from small to large is to traverse the pixels currently located in the depth map area inside the preset structural element. If the target depth closest to the center pixel is traversed in the depth map area point, use the depth value with the highest frequency of occurrence among the target depth points as the depth value of the central pixel point.

In this embodiment, the average value corresponding to the depth values of all the target depth points is used as the depth value of the central pixel point. Taking the rectangular structural element as an example, as shown in Figure 3, after the preset structural element is determined, the traversal order of the pixel points within the structural element from small to large from the center pixel of the preset structural element is set in the preset structural element, When traversing in each structural element, it is traversed according to the preset order. If point 1, point 2 and point 5 are all blank points, and the first depth point encountered is point 6, then it is the same as point 6. Points No. 7, No. 8, and No. 9, which have the same distance from point No. 1 to the center of the preset structural element, are all judged, and the depth points in points 6, 7, 8, and 9 are used as the target depth points, and the frequency of occurrence among them is taken The largest depth value is used as the depth value of the central pixel point. For example, the depth values of point 7, point 8 and point 9 are the same but the depth value of point 6 is different, then point 7, point 8 or The depth value corresponding to point 9 is used as the depth value of the central pixel point.

Step S43: using the preset structural elements to perform an erosion operation on the expanded depth map to obtain a dense depth map.

Step S44: Process the dense depth map using smoothing filtering technology to obtain a dense smooth depth map.

It can be seen that in this embodiment, by taking the target depth point and taking the depth value with the highest frequency of occurrence as the depth value of the central pixel point, it can be avoided that the depth value of the first encountered depth point is the same as the others. The depth value assignment of the central pixel point is inaccurate due to the large gap of the depth value from the depth point.

Referring to Figure 6, the embodiment of the present application discloses a dense depth map construction device, including:

Depth map acquisition module 11, configured to acquire the initial depth map generated by the speckle structured light system;

The expansion module 12 is configured to perform an expansion operation on the initial depth map by using a preset structural element to obtain a corresponding expanded depth map; the expansion operation logic corresponding to the preset structural element is as follows: when the preset structural element traverses to For any pixel on the initial depth map, determine the depth value of the center pixel based on the depth value of the target depth point closest to the center pixel of the preset structure element;

An erosion module 13, configured to use the preset structural elements to perform an erosion operation on the expanded depth map to obtain a dense depth map;

The smoothing processing module 14 is configured to process the dense depth map using a smoothing filtering technique to obtain a dense smooth depth map.

It can be seen that when constructing a dense depth map in this application, the traditional interpolation algorithm is not used, but the morphological image processing method is used, and the expansion operation in the traditional morphological closing operation is improved. In the process of the improved expansion operation Considering the distance relationship between the central pixel of the preset structural element and the depth point to determine the expanded depth value of the pixel at the central pixel position of the preset structural element, the depth value of the expanded depth point is more accurate. Therefore, the present application avoids the bull's-eye phenomenon caused by the traditional interpolation algorithm and the problem of object deformation and depth quality degradation in the depth map when constructing a dense depth map; The problem of depth distortion caused by depth point coverage after the dilation operation in the traditional closing operation is avoided.

In some specific embodiments, the device for constructing a dense depth map further includes:

An average distance calculation module, configured to calculate an average distance between speckle pixels in the initial depth map;

A structural element determination module, configured to determine a structural element size parameter based on the average distance, and construct a corresponding structural element according to the structural element size parameter, so as to obtain the preset structural element.

In some specific embodiments, the expansion module 12 specifically includes:

A structural element central point pixel depth value determination unit, configured to use the preset structural element to traverse the initial depth map, and when traversing to any pixel on the initial depth map, follow the preset Set the distance between the central pixels of the structural elements in the order of traversal from small to large, and traverse the pixels currently located in the depth map area inside the preset structural element, if the distance is traversed in the depth map area The nearest target depth point of the center pixel point, then determine the depth value of the center pixel point based on the depth value of the target depth point, if no depth point can be traversed in the depth map area, then the center Pixels are marked as blank points.

In some specific embodiments, the corrosion module 13 specifically includes:

A blank point marking unit, configured to use the preset structural element to traverse the expanded depth map, and when traversing to any pixel on the expanded depth map, if the pixel currently located inside the preset structural element If any pixel in the depth map area is a blank point, the central pixel is marked as a blank point.

FIG. 7 shows an electronic device 20 provided by an embodiment of the present application. Specifically, the electronic device 20 may further include: at least one processor 21 , at least one memory 22 , a power supply 23 , a communication interface 24 , an input/output interface 25 and a communication bus 26 . Wherein, the memory 22 is used to store a computer program, and the computer program is loaded and executed by the processor 21 to implement relevant steps in the method for constructing a dense depth map disclosed in any of the foregoing embodiments. In addition, the electronic device 20 in this embodiment may specifically be an electronic computer.

In this embodiment, the power supply 23 is used to provide voltage for each hardware device on the electronic device 20; the communication interface 24 can create a data transmission channel between the electronic device 20 and external devices, and the communication protocol it follows is applicable to Any communication protocol of the technical solution of the present application is not specifically limited here; the input and output interface 25 is used to obtain external input data or output data to the external, and its specific interface type can be selected according to specific application needs, and is not described here. Make specific restrictions.

In addition, the memory 22, as a resource storage carrier, can be a read-only memory, random access memory, magnetic disk or optical disk, etc., and the resources stored thereon can include operating system 221, computer program 222, etc., and the storage method can be temporary storage or permanent storage. .

Wherein, the operating system 221 is used to manage and control each hardware device on the electronic device 20 and the computer program 222, which may be Windows Server, Netware, Unix, Linux and so on. In addition to computer programs that can be used to complete the dense depth map construction method performed by the electronic device 20 disclosed in any of the foregoing embodiments, the computer program 222 may further include computer programs that can be used to complete other specific tasks.

It can be seen that when constructing a dense depth map in this application, the traditional interpolation algorithm is not used, but the morphological image processing method is used, and the expansion operation in the traditional morphological closing operation is improved. In the process of the improved expansion operation Considering the distance relationship between the central pixel of the preset structural element and the depth point to determine the expanded depth value of the pixel at the central pixel position of the preset structural element, the depth value of the expanded depth point is more accurate. Therefore, the present application avoids the bull's-eye phenomenon caused by the traditional interpolation algorithm and the problem of object deformation and depth quality degradation in the depth map when constructing a dense depth map; The problem of depth distortion caused by depth point coverage after the dilation operation in the traditional closing operation is avoided.

Furthermore, the present application also discloses a computer-readable storage medium for storing a computer program; wherein, when the computer program is executed by a processor, the aforementioned dense depth map construction method disclosed above is realized. Regarding the specific steps of the method, reference may be made to the corresponding content disclosed in the foregoing embodiments, and details are not repeated here.

Finally, it should also be noted that in this text, relational terms such as first and second etc. are only used to distinguish one entity or operation from another, and do not necessarily require or imply that these entities or operations, any such actual relationship or order exists. Furthermore, the term "comprises", "comprises" or any other variation thereof is intended to cover a non-exclusive inclusion such that a process, method, article or apparatus comprising a set of elements includes not only those elements, but also includes elements not expressly listed. other elements of or also include elements inherent in such a process, method, article, or device. Without further limitations, an element defined by the phrase "comprising a ..." does not exclude the presence of additional identical elements in the process, method, article or apparatus comprising said element.

The method, device, equipment and medium for constructing a dense depth map provided by the present invention have been described above in detail. In this paper, specific examples are used to illustrate the principle and implementation of the present invention. The description of the above embodiments is only for To help understand the method of the present invention and its core idea; at the same time, for those of ordinary skill in the art, according to the idea of the present invention, there will be changes in the specific implementation and scope of application. In summary, the content of this specification It should not be construed as limiting the invention.

Claims (8)

1. A method for constructing a dense depth map, comprising: Obtain the initial depth map generated by the speckle structured light system; Perform an expansion operation on the initial depth map by using a preset structural element to obtain a corresponding expanded depth map; the expansion operation logic corresponding to the preset structural element is when the preset structural element traverses to the initial depth map When any pixel, determine the depth value of the central pixel based on the depth value of the target depth point closest to the central pixel of the preset structure element; performing an erosion operation on the expanded depth map by using the preset structural elements to obtain a dense depth map; Processing the dense depth map by using a smoothing filtering technique to obtain a dense smooth depth map; Wherein, the eroding operation on the expansion depth map by using the preset structural elements includes: traversing the expansion depth map by using the preset structural elements, and traversing to the expansion depth map For any pixel, if any pixel currently located in the depth map area inside the preset structure element is a blank point, mark the central pixel as a blank point; Moreover, the dilation operation on the initial depth map by using the preset structural elements includes: traversing the initial depth map by using the preset structural elements, and traversing to any of the initial depth maps pixels, according to the traversal order of the distance from the central pixel of the preset structural element from small to large, traverse the pixels currently located in the depth map area inside the preset structural element, if Traversing to the target depth point closest to the center pixel point in the depth map area, then determining the depth value of the center pixel point based on the depth value of the target depth point, if traversing in the depth map area does not to the depth point, then mark the central pixel point as a blank point. 2. The method for constructing a dense depth map according to claim 1, wherein, before performing an expansion operation on the initial depth map using preset structural elements, further comprising: calculating an average distance between speckle pixels in the initial depth map; A structural element size parameter is determined based on the average distance, and a corresponding structural element is constructed according to the structural element size parameter, so as to obtain the preset structural element. 3. The dense depth map construction method according to claim 2, wherein said constructing corresponding structural elements according to said structural element size parameters to obtain said preset structural elements comprises: Screen out the target structure element shape from the preset structure element shape library based on the preset structure element shape requirements, and construct the corresponding structure element according to the structure element size parameter and the target structure element shape, so as to obtain the preset structure element shape Structural element; the shape library of the structural element includes any one or a combination of rectangles, rhombuses, crosses and circles. 4. The method for constructing a dense depth map according to claim 1, wherein, after the dense depth map is obtained by using a smoothing filter technique to obtain a dense smooth depth map, further comprising: Performing three-dimensional reconstruction based on the dense smooth depth map to obtain corresponding three-dimensional reconstruction results. 5. The dense depth map construction method according to claim 1, wherein if the target depth point closest to the central pixel point is traversed in the depth map region, then based on the target depth point The depth value determines the depth value of the center pixel, including: If the target depth point closest to the central pixel point is traversed in the depth map area, the depth value of the first traversed target depth point is used as the depth value of the central pixel point; Or, if the target depth point closest to the central pixel point is traversed in the depth map area, the average value corresponding to the depth values of all the target depth points is used as the depth value of the central pixel point; Or, if the target depth point closest to the central pixel point is traversed in the depth map area, a depth value with the highest frequency of occurrence among the target depth points is used as the depth value of the central pixel point. 6. A device for constructing a dense depth map, comprising: A depth map acquisition module, configured to obtain an initial depth map generated by a speckle structured light system; The expansion module is used to perform an expansion operation on the initial depth map by using a preset structural element to obtain a corresponding expanded depth map; the expansion operation logic corresponding to the preset structural element is that when the preset structural element traverses to the When any pixel on the initial depth map is selected, the depth value of the central pixel is determined based on the depth value of the target depth point closest to the central pixel of the preset structural element; An erosion module, configured to use the preset structural elements to perform an erosion operation on the expanded depth map to obtain a dense depth map; A smoothing processing module, configured to process the dense depth map using a smoothing filtering technique to obtain a dense smooth depth map; Wherein, the corrosion module is specifically configured to use the preset structural element to traverse the expansion depth map, and when traversing to any pixel on the expansion depth map, if the current location is at the preset Any pixel in the depth map area inside the structural element is a blank point, then mark the central pixel as a blank point; Moreover, the expansion module is specifically configured to use the preset structure element to traverse the initial depth map, and when traversing to any pixel on the initial depth map, follow the preset structure The distance between the central pixels of the element is traversed from small to large, and the pixels currently located in the depth map area inside the preset structure element are traversed. If the distance between the depth map area is traversed to the The nearest target depth point of the central pixel point, then determine the depth value of the central pixel point based on the depth value of the target depth point, if the depth point cannot be traversed in the depth map area, then the central pixel point Mark as blank spots. 7. An electronic device, characterized in that it comprises: memory for storing computer programs; A processor, configured to execute the computer program, so as to realize the steps of the method for constructing a dense depth map according to any one of claims 1 to 5. 8. A computer-readable storage medium, characterized in that it is used to store a computer program; wherein, when the computer program is executed by a processor, the dense depth map construction method according to any one of claims 1 to 5 is implemented step.

Images