TWI818715B - A method for visual inspection of curved objects - Google Patents

Abstract

The method for visual inspection of curved objects, which is executed by a visual inspection that includes a robotic arm；a camera；a fixing unit； and a control unit to perform the following steps: S101: fixing an object to be inspected with the fixing unit; S102: photographing the object to be inspected with preset parameters to obtain images of the object with the camera and counting the number of pixels occupied by each grayscale value in the image of the object with the control unit; S103: calculating a photographing parameter for inspection based on the number of pixels occupied by the grayscale value in the image of the object with the control unit; S104: performing visual inspection of a large number of objects to be inspected with the photographing parameter for inspection with the camera.

Description

Methods for visual inspection of curved objects

The present invention relates to a method for visual inspection of objects, and in particular, to a visual inspection method for visual inspection of curved objects and capable of reducing the need to re-adjust the shooting angle of the object.

Please refer to the fifth figure. When visually inspecting an object, if the surface of the object 20 to be inspected is a curved surface, because different parts of the object surface have slightly different distances from the light source and the normal vectors of the parts are different, or the surface of the object uses Different materials mean that even if you shoot with the same exposure settings, the final exposure results will be very different. Therefore, it often happens that the various surfaces of the object to be inspected 20 have extremely different exposure values, and it is impossible to clearly identify whether the surface of the object to be inspected 20 has defects in the same photo.

For example, when the camera 30 and the light source of an object to be detected 20 shown in this figure are placed above and illuminated and photographed downward, the normal vector of the object to be detected 20 points to the light source and the flat part receives The reflected light is stronger, while the light received by the inclined part is weaker. Therefore, the flat portion has higher exposure, but the inclined portion is prone to underexposure.

And since underexposure or overexposure will make it difficult for the visual inspection system to identify the structural details in this part, in order to clearly identify various parts of the object to be inspected 20, the conventional visual inspection system must frequently adjust the camera 30, The positions and angles between the light source and the object to be detected 20 must be photographed separately to ensure that each surface of the object to be detected 20 is correctly exposed. However, re-adjusting the positions and angles between the camera 30, the light source and the object to be detected 20 is very time-consuming. Therefore, visual inspection of curved objects requires a lot of time in re-adjusting the shooting angle. In summary, it is necessary to provide a visual inspection method that can reduce the need to readjust the shooting angle of objects.

The object of the present invention is to provide a method for visual inspection of objects, and in particular, to a method for visual inspection of curved surface objects that can reduce the need to readjust the shooting angle of the object.

In order to achieve the above objectives, a method of visual inspection of curved surface objects of the present invention is executed through a visual inspection system. The visual inspection system includes a robotic arm; a camera, and the camera is installed on the robotic arm. the end; a fixed unit, the fixed unit is arranged below the camera; and a control unit, the control unit is electrically connected to the robotic arm and the camera, and the curved surface object is visually inspected The method performs the following steps when executed by the control unit: S101: the fixing unit fixes an object to be detected; S102: the camera shoots the object to be detected with a plurality of sets of preset parameters to obtain a plurality of sets of object images, The control unit counts the number of pixels occupied by each gray-scale value in the object image; S103: The control unit calculates an optimal shooting parameter for detection based on the number of pixels occupied by the gray-scale value in the object image; S104: The camera performs visual inspection in batches with the optimal shooting parameters for inspection.

In a preferred embodiment, the step S102 further includes the following steps: S201: The camera shoots the object to be detected with a first exposure setting to obtain a first object image, and the control unit counts the first object image. The number of pixels occupied by each grayscale value in an object image; S202: The camera maintains the same shooting angle and position and shoots the object to be detected with an exposure different from the first exposure setting to obtain object images with different exposures, and The control unit counts the number of pixels occupied by each grayscale value in the object images with different exposures.

In a preferred embodiment, the step S202 includes the camera capturing a lower-exposed object image with an exposure lower than the first exposure setting, and the control unit counts the The number of pixels occupied by each grayscale value in the lower-exposed object image; and a higher-exposed object image obtained by the camera shooting with an exposure higher than the first exposure setting, and the control unit counts the The number of pixels occupied by each grayscale value in the image of a higher-exposure object.

In a preferred embodiment, step S103 further includes the following steps: S301: The control unit presets a target grayscale value; S302: The control unit compares the target grayscale value in the first object image. The number of pixels occupied by the level value is the same as the number of pixels occupied by the target gray level value in the object images with different exposures, and the exposure setting with a larger number of pixels occupied by the target gray level value is set as the optimal detection shooting parameter.

In order to explain in detail the technical content, structural features, objectives and effects of the present invention, the following examples are given in conjunction with the drawings.

Please refer to the first and fifth figures. The present invention is a visual inspection method. The visual inspection method of the present invention can be executed in different visual inspection equipment, but in this preferred embodiment, the method of visual inspection of curved objects of the present invention is executed by a visual inspection system with a robotic arm. The visual inspection system includes a robotic arm 40; a camera 30, the camera 30 is installed at the end of the robotic arm 40; a fixing unit 50, the fixing unit 50 is installed on the end of the camera 30 The object to be detected 20 is fixed below; and a control unit (not shown) is electrically connected to the robotic arm 40 and the camera 30 and controls the robotic arm 40 To adjust the shooting angle and control the camera 30 to adjust the exposure parameters.

In this preferred embodiment, a method for visual inspection of curved surface objects of the present invention is executed through the control unit and then the following steps are performed: S101: Fix an object to be inspected 20; S102: Use multiple groups with the same shooting position However, the object to be detected 20 is photographed with the default parameters of different exposure settings to obtain a plurality of groups of object images 10, and the control unit counts the number of pixels occupied by each grayscale value in the object image 10; S103: According to the above The number of pixels occupied by the grayscale value in the object image 10 is used to calculate a better detection photography parameter; S104: Visually detect a large number of objects 20 to be detected using the detection photography parameters.

Please refer to the first and second figures again. Step S102 further includes the following steps: S201: Photograph the object to be detected 20 with the first exposure setting to obtain the first object image 11, and collect statistics on the first object image 11. The number of pixels occupied by each grayscale value; S202: Keep the same shooting angle and position and shoot the object to be detected 20 several times with an exposure different from the first exposure setting to obtain object images 10 with different exposures, and make statistics The number of pixels occupied by each grayscale value in the object image 10 with different exposures.

Please refer to the second and fourth figures. In this preferred embodiment, in step S201, a first object image 11 is captured with an exposure time of 300 milliseconds. Then in step S202, the object to be detected 20 is photographed using the same shooting angle, position, light source, aperture and sensitivity with exposure times of 100 milliseconds, 200 milliseconds, 400 milliseconds and 500 milliseconds respectively to obtain the second to fifth object images 12 The image 15 is used to count the number of pixels occupied by each grayscale value in the first object image 11 to the fifth object image 15 .

When the exposure time is shortened, the object image 10 will be underexposed and the overall grayscale value of the object image 10 will be reduced. However, the direct light-receiving surface or light-colored areas in the object image 10 can also be avoided. Overexposed. On the other hand, when the exposure time is lengthened, the object image 10 will be overexposed and the overall grayscale value of the object image 10 will increase. However, it can also avoid the backlight or dark side of the object image 10 . Color areas are underexposed. As shown in the figure, the overall grayscale value in the second object image 12 and the third object image 13 with shorter exposure time is lower, but the turning point 21 at the top of the handle is also avoided. The cross button on the handlebar is overexposed. The overall grayscale value in the fourth object image 14 and the fifth object image 15 with a longer exposure time is higher, and underexposure of the handle surface 23 and the handle rocker cap 24 is avoided.

Please refer to the first, third and fourth pictures again. After obtaining the first object image 11 to the fifth object image 15 with different shooting parameters, the following steps are further performed: S301: Preset a target grayscale value. , in this embodiment, the target grayscale value is set to the middle value of the grayscale value range in the object image to avoid falling into the underexposure or overexposure range during sampling. However, the selection of this target grayscale value It still needs to be adjusted depending on the material or color of the object to be inspected. S302: Compare the number of pixels occupied by the target grayscale value in the first object image 11 to the fifth object image 15, and set the exposure setting with a larger number of pixels occupied by the target grayscale value as the detection Use shooting parameters.

In order to expand the area of the object image that can be visually detected, in this preferred embodiment, any two object images 10 from the first object image 11 to the fifth object image 15 are selected as a group. method, and count and compare the number of pixels falling within the target grayscale value in different object image groups. And after statistics and comparison, it is obtained that combining the third object image 13 and the fourth object image 14 can maximize the number of pixels of the target grayscale value. Therefore, in step S103, the third object image is The combination of two sets of shooting parameters 13 and the fourth object image 14 are used as the shooting parameters for detection.

Next, in step S104, the subsequent objects 20 to be detected are tested in batches using the detection photography parameters obtained in step S103, that is, each subsequent object 20 to be detected uses an exposure time of 200 milliseconds and 400 milliseconds respectively. An object image 10 is photographed and the object image 10 obtained is visually inspected.

The visual detection method in the present invention determines the shooting parameters for detection by maximizing the pixels occupied by the target grayscale value in the object image, thereby maximizing the observable object characteristics at each shooting angle. And reduce the number of times you have to re-adjust the shooting angle of objects.

Although the present invention has been disclosed in the foregoing embodiments, they are not intended to limit the present invention. Anyone skilled in the similar art can make some modifications and modifications without departing from the spirit and scope of the present invention. Therefore, the present invention is The scope of patent protection shall be determined by the scope of the patent application attached to this specification.

S101~S302: steps

10:Object image

11:First object image

12:Second object image

13:Third object image

14: The fourth object image

15:Fifth object image

20: Object to be inspected

21:The turning point

22: Cross button

23: Handle surface

24: rocker cap

30:Camera

40:Robotic arm

50: Fixed unit

The first figure is a flow chart for visual inspection of curved surface objects in the present invention.

The second figure is a flow chart of obtaining an object image by shooting with preset parameters in the present invention.

The third figure is a flow chart for calculating the shooting parameters for detection based on the number of pixels occupied by the grayscale value in the object image in the present invention.

The fourth figure is a schematic diagram of object images captured with different shooting parameters in the present invention.

The fifth picture is a schematic diagram of visual inspection of objects.

S101~S104: Steps

Claims (3)

A method of visual inspection of curved surface objects is performed through a visual inspection system. The visual inspection system includes a robotic arm; a camera installed at the end of the robotic arm; and a fixed unit. The fixing unit is disposed below the camera; and a control unit is electrically connected to the robotic arm and the camera, and the method of visual inspection of curved objects is controlled by the control unit. The following steps are performed during execution: S101: The fixing unit fixes an object to be detected; S201: The camera photographs the object to be detected with a first exposure setting to obtain a first object image, and the control unit counts the first object The number of pixels occupied by each grayscale value in the image; S202: The camera maintains the same shooting angle and position and shoots the object to be detected with an exposure different from the first exposure setting to obtain images of the object with different exposures, and the The control unit counts the number of pixels occupied by each grayscale value in the object image with different exposures; S103: The control unit calculates an optimal shooting parameter for detection based on the number of pixels occupied by the grayscale value in the object image; S104: The camera performs visual inspection in batches with the optimal shooting parameters for inspection. The method for visual inspection of curved surface objects as described in item 2 of the patent application, wherein the step S202 includes a lower exposure obtained by shooting with the camera at an exposure lower than the first exposure setting. Object image, and the control unit counts the number of pixels occupied by each grayscale value in the lower exposure object image; and the camera uses A higher-exposed object image is captured at an exposure higher than the first exposure setting, and the control unit counts the number of pixels occupied by each grayscale value in the higher-exposed object image. The method for visual inspection of curved surface objects as described in item 2 of the patent application scope, wherein the step S103 further includes the following steps: S301: the control unit presets a target grayscale value; S302: the control The unit compares the number of pixels occupied by the target grayscale value in the first object image with the number of pixels occupied by the target grayscale value in the object images with different exposures, and sets the exposure setting with a larger number of pixels occupied by the target grayscale value. Set as the optimal detection shooting parameters.