Roughness recognition device based on multi-light field view angle and algorithm detection method

Abstract

The application provides a roughness recognition device based on a multi-light field view angle and an algorithm detection method, and has the characteristics of including a light source A, a light source B and an industrial camera, the industrial camera is arranged directly above a product to be measured, the light source A is arranged on the front side of the product to be measured, and the light source B is arranged on the rear side of the product to be measured; the device can quickly and accurately detect the roughness of the material, and can recognize the problematic material according to the algorithm, and has the advantages of low cost, high speed and easy implementation.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of intelligent manufacturing of automatic processing and furniture, and particularly relates to a roughness recognition device based on multiple light field perspectives and an algorithm detection method. BACKGROUND

[0002] With the improvement of industrial automation, in the wood and furniture manufacturing industry, the automation degree is getting higher and higher, and the division of processing is getting more and more detailed. In the production process of solid wood boards, the obtained raw wood needs to be cut and planed first, and then a series of procedures such as finger jointing are performed to process solid wood boards, which can be provided to furniture factories for furniture production and manufacturing.

[0003] In the planing process, the surface should be smooth after planing. However, some incoming wood has different sizes, is curved or has internal defects, and after planing, the surface is still rough and unqualified, so roughness inspection is needed for the planed products. The current industry mainly adopts two methods for this inspection work, one is manual detection, and the other is detection by high-speed 3D laser profiler. The manual detection has low efficiency, non-uniform judgment standard, and is easy to miss detection. The 3D laser profiler has high cost and high precision requirement for the conveying equipment. When the precision is not met, false detection or missed detection may occur, so the current industry still mainly relies on manual detection. SUMMARY

[0004] In view of the above technical problems, the present application provides a roughness recognition device based on multiple light field perspectives and an algorithm detection method for improving detection speed and accuracy.

[0005] The present application provides a roughness recognition device based on multiple light field perspectives, characterized by comprising a light source A, a light source B and an industrial camera, wherein the industrial camera is arranged directly above the product to be detected, the light source A is arranged on the front side of the product to be detected, and the light source B is arranged on the rear side of the product to be detected.

[0006] The light source A and the light source B are alternately lit.

[0007] The light emitting direction of the light source A and the light source B forms an acute angle with the advancing direction of the planed product to be detected, and the acute angle is preferably 30-45 degrees.

[0008] The light source A and the light source B are preferably parallel light sources.

[0009] The present application provides a roughness algorithm detection method based on multiple light field perspectives, and the steps are as follows:

[0010] Step one, light up light source A, light source B in turn, take a picture with an industrial camera after each time, so you get images under different light field view, recorded as Img1 and Img2;

[0011] Step two, respectively, the two images are subjected to gray scale conversion, the conversion idea is to change the average gray scale to 127, calculate the average gray value avgGray for each image, and the gray offset value is grayOffset = 127-avgGray;

[0012] Step three, the gray offset value of each pixel in the image is calculated, newValue = oldValue + grayOffset, if newValue is less than 0, it is set to 0, if it is greater than 255, it is set to 255, so the two processed images proImg1, proImg2 are obtained;

[0013] Step four, then compare the pixels, set the pixel value greater than the judgment threshold as foreground color 255, and set the other as background color 0, so the contrast binary image is obtained;

[0014] Step five, according to the area size of the foreground color in the surrounding area of each pixel as a index to judge the roughness of the point, the unqualified area can be located, so as to judge whether the roughness of each module is qualified product.

[0015] The specific judgment method of step five is: when the area of the foreground color in the positioning area is large, it means that the roughness of the positioning area is large, if the roughness of the large area positioning area is large, the product is unqualified.

[0016] The specific value of the rough area can be set according to the needs.

[0017] The beneficial effects of the present application are: the present application is a roughness recognition device and algorithm detection method based on multi-light field view, which can quickly and accurately detect the roughness of the product to be tested, and can recognize the problem of the product to be tested according to the algorithm, which is low in cost, fast in speed and easy to realize. The present application can not only be applied to the roughness detection of wood, but also can be used for detecting the materials such as wall surface or ground surface which need roughness detection, and has wide application range. BRIEF DESCRIPTION OF DRAWINGS

[0018] Figure 1 It is a schematic diagram of the device of the present application;

[0019] Figure 2 It is a schematic diagram of the detection algorithm flow of the present application;

[0020] Figure 3 It is a schematic diagram of the sample with smaller roughness of the present application;

[0021] Figure 4 The roughness of the sample of the present application is large.

[0022] Reference signs: 1-light source A; 2-light source B; 3-industrial camera; 4-product to be measured. DETAILED DESCRIPTION

[0023] The technical solutions of the present application will be described in detail below in combination with the drawings of the present application.

[0024] The present application provides a roughness algorithm detection method based on multi-light field view, and the steps are:

[0025] Step one, light up light source A and light source B in turn, take pictures with an industrial camera after each time, so that images under different light field views are obtained, denoted as Img1 and Img2;

[0026] Step two, perform gray scale transformation on the two images respectively, and the transformation idea is to change the average gray scale to 127, calculate the average gray scale value avgGray for each image, and the gray scale offset value is grayOffset = 127-avgGray;

[0027] Step three, perform gray scale offset calculation on each pixel in the image, newValue = oldValue + grayOffset, if newValue is less than 0, set it to 0, if it is greater than 255, set it to 255, so that the two processed images proImg1 and proImg2 are obtained;

[0028] Step four, then perform pixel comparison, set the pixel value with a difference greater than the judgment threshold to foreground color 255, and set the others to background color 0, so that the contrast binary image is obtained;

[0029] Step five, according to the foreground color area size in the surrounding area of each pixel as a criterion to judge the roughness of the point, the unqualified area can be located, so as to judge whether the roughness of each module is a good product.

[0030] The specific judgment method of step five is: when the foreground color area of the positioning area is large, it means that the roughness of the positioning area is large, if a large area of positioning area has large roughness, the product is unqualified.

[0031] The specific numerical value of the rough area can be set according to the needs.

[0032] The application provides a roughness recognition device based on multi-light field view, which is characterized by comprising a light source A1, a light source B2 and an industrial camera 3, the industrial camera 3 is arranged above a product 4 to be detected, the light source A1 is arranged on the front side of the product 4 to be detected, and the light source B2 is arranged on the rear side of the product 4 to be detected.

[0033] The light source A1 and the light source B2 are alternately lighted.

[0034] The light emitting directions of the light source A1 and the light source B2 form an acute angle with the advancing direction of the product 4 to be detected, and the acute angle is preferably 30-45 degrees.

[0035] The light source A1 and the light source B2 are preferably parallel light sources.

[0036] Embodiment 1

[0037] When the product 4 to be detected is a wood block, first, the wood block planed from a planer is conveyed to a designated position, and the wood block is clamped left and right, the light source A1 is lighted, the light source B2 is turned off, and then the industrial camera 3 is used to take an image Img1; then the light source A1 is turned off, the light source B2 is lighted, and then the industrial camera 3 is used to take an image Img2; finally, the light source B2 is turned off, and the image collection of different light fields is completed.

[0038] Then, the collected images are processed by an algorithm, and the detected unqualified part is located to a specific wood block, if there is such a wood block, it is picked out by a robot, and a deeper planing process is performed.

[0039] After picking out the unqualified wood block, the remaining surface is a qualified product, for a double-sided planing device, an additional device is needed to turn it over so that the back is up, and then the detection is performed again. Then the unqualified product is picked out.

[0040] Embodiment 2

[0041] When the product 4 to be detected is a wall surface or a floor, the light source A1, the light source A2 and the industrial camera 3 are installed on a mobile device, the mobile device is started, and when a certain detection point is reached, the mobile device is stopped and detection is performed. During detection, the light source A1 and the light source A2 are alternately lighted, and the industrial camera 3 is used to take a picture for each time of lighting. Then, according to the picture taking result, the roughness detection is performed by using the algorithm detection method described above, whether the area meets the roughness requirement can be detected, and the specific area with problems can be located, and then feedback is given to an upper computer for recording or other operation processing.

[0042] The above shows and describes the basic principles and main features of the present application and the advantages of the present application. The various components mentioned in the present application are common techniques in the prior art, and those skilled in the art should understand that the present application is not limited to the above embodiments, and the above embodiments and descriptions in the specification are only to illustrate the principles of the present application. Without departing from the spirit and scope of the present application, various changes and improvements can be made to the present application, and these changes and improvements all fall within the scope of the claimed present application. The scope of protection of the present application is defined by the appended claims and their equivalents.

Claims

1. A roughness detection method based on multi-light field view angle, comprising the following steps: Step one, light up light source A and light source B in turn, take pictures with an industrial camera after each time, thus obtaining images under different light field view angles, denoted as Img1 and Img2; Step two, perform gray scale transformation on the two images respectively, the transformation idea is to change the average gray scale to 127, calculate the average gray scale value avgGray for each image, and the gray scale offset value is grayOffset = 127-avgGray; Step three, calculate the gray scale offset for each pixel in the image, newValue = oldValue +grayOffset, if newValue is less than 0, set it to 0, if it is greater than 255, set it to 255, thus obtaining the processed two images proImg1 and proImg2; Step four, then perform pixel comparison, set the pixel value with a difference greater than the judgment threshold to foreground color 255, and set the others to background color 0, thus obtaining the contrast binary image; Step five, according to the foreground color area size in the surrounding area of each pixel as a criterion to judge the roughness of the point, locate the unqualified area, and thus judge whether the roughness of each module is good.

2. The multi-light-field view-based roughness algorithm detection method according to claim 1, characterized in that The specific judgment method of step five is: when the foreground color area of the located area is large, it represents that the roughness of the located area is large, if a large area of the located area has large roughness, the product is unqualified.

3. The method of claim 2, wherein the method is based on a multi-light-field view angle-based roughness algorithm. The specific numerical value of the rough area is set according to the requirement.

4. The multi-light-field view-based roughness algorithm detection method according to claim 1, wherein The device using the method comprises light source A, light source B and an industrial camera, the industrial camera is arranged above the product to be tested, the light source A is arranged on the front side of the product to be tested, and the light source B is arranged on the rear side of the product to be tested.

5. The multi-light-field view-based roughness algorithm detection method according to claim 4, characterized in that The light source A and the light source B are alternately lighted up.

6. The multi-light-field view-based roughness algorithm detection method according to claim 4, characterized in that The light emitting direction of the light source A and the light source B is an acute angle with the advancing direction of the product to be tested, and the acute angle is 30-45 degrees.

7. The multi-light-field view-based roughness algorithm detection method according to claim 4, characterized in that The light source A and the light source B are parallel light sources.

Images