Automobile paint color difference identification method based on color appearance model
Through the method based on the color and appearance model, the color difference of automobile paint is automatically calculated, which solves the problem of subjectivity and low efficiency of traditional detection methods, and realizes high-precision and efficient color difference detection of automobile paint, supporting the development of intelligent manufacturing.
Patent Information
- Application Number
- CN202411163747.9
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-23
- Publication Date
- 2024-11-15
- Estimated Expiration
- Not applicable · inactive patent
AI Technical Summary
The traditional automotive paint color difference detection methods have problems such as strong subjectivity, low efficiency and insufficient accuracy, making it difficult to achieve automated and quantitative detection.
Using a color-like model-based method, by obtaining the car paint image under the same lighting conditions, converting it to the CIEXYZ color space, calculating the color-like properties, and calculating the color-like properties based on these properties to determine whether there are color-like defects, providing automation and quantitative Testing scheme.
It realizes automated inspection, reduces errors from human factors, improves detection accuracy and efficiency, and can provide quantitative data to support the development of intelligent manufacturing and production line networking, ensuring the consistency and accuracy of automotive paint coatings.
Smart Images

Figure CN118960963A_ABST
Abstract
Description
Technical Field
[0001] The invention relates to the technical field of defect detection, and in particular to a method for identifying color difference of automobile paint surfaces based on a color appearance model. Background Art
[0002] Car paint is the outermost coating in the car painting process. It not only plays a role in aesthetics, but also has the function of protecting the car body from environmental factors, such as weather resistance and fading resistance. Identifying the color difference of car paint is an important technical means to improve vehicle consistency and enhance the quality of cars. Traditional color difference detection methods usually rely on manual visual inspection or simple color measurement tools, which may have problems such as strong subjectivity, low efficiency, and insufficient precision. Summary of the invention
[0003] The purpose of the present invention is to provide a method for identifying color difference of automobile paint surface based on color appearance model to solve the problems existing in the prior art.
[0004] In order to achieve the above-mentioned purpose, the technical solution of the present invention provides a method for identifying color difference of automobile paint surface based on color appearance model, comprising the following steps: S1, under the same lighting conditions, obtaining an automobile paint surface image and a standard automobile paint surface image in a measured area; S2, converting the automobile paint surface image in the measured area into a CIEXYZ color space after preprocessing to obtain an XYZ value, and converting the standard automobile paint surface image into a CIEXYZ color space to obtain an XYZ value; S3, calculating the color appearance attributes of the automobile paint surface image in the measured area and the standard automobile paint surface image; S4, based on the color appearance attributes, calculating the color difference of the automobile paint surface image in the measured area and the standard automobile paint surface image, and comparing the color difference value with a preset value range, if it exceeds the preset range, it is determined that the automobile paint surface image in the measured area has a color difference defect.
[0005] Furthermore, in step S1, the lighting condition is a D65 light source with the same brightness and illumination angle.
[0006] Furthermore, if the standard automobile paint image is collected from a standard color card, the size of a single measured area automobile paint image unit is determined according to the size of the standard automobile paint image, and multiple measured area automobile paint image units are collected at different positions of the automobile paint surface.
[0007] Furthermore, if the standard automobile paint surface image is collected for the entire vehicle, the automobile paint surface image of the measured area is consistent with the standard automobile paint surface image.
[0008] Furthermore, the color space conversion process of the measured area car paint image and the standard car paint image in step 2 is as follows: the RGB value needs to be linearly mapped from [0, 255] to [0, 1], and the formula is as follows: ; Calculate the XYZ value based on the normalized RGB value, the formula is as follows:
[0009] Furthermore, the image color attributes in step S3 at least include the red-green component a, the yellow-blue component b and the brightness J Furthermore, the color difference calculation process based on the color appearance attribute in step S4 is as follows:
[0010] In the formula: They are the brightness difference between the car paint in the tested area and the standard car paint, the difference between the red-green component, the difference between the yellow-blue component, It is the color difference between the car paint surface in the tested area and the standard car paint surface.
[0011] The present invention adopts a color difference evaluation method based on the color appearance model, which can realize automation and quantification, reduce the error of human factors, and effectively improve the working efficiency and detection accuracy of the process. At the same time, color difference recognition based on the color appearance model can provide quantitative data support, which is in line with the current development direction of intelligent manufacturing and production line networking, and is helpful for further analysis and optimization of the production process.
[0012] In terms of specific technical means, the use of color appearance models and comprehensive consideration of the impact of environmental factors can better reflect the differences in consumers' color vision, improve the accuracy and reliability of color difference evaluation, and help ensure the consistency and accuracy of automotive paint coatings. BRIEF DESCRIPTION OF THE DRAWINGS
[0013] Figure 1 It is a schematic diagram of an overall scheme of an embodiment of the present invention. DETAILED DESCRIPTION
[0014] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.
[0015] As attached Figure 1 As shown, the method for detecting color difference defects of automobile paint involved in the present invention comprises the following steps: S1. Under the same lighting conditions, obtain the car paint surface image of the measured area and the standard car paint surface image; S2, converting the measured area automobile paint surface image into CIEXYZ color space after preprocessing to obtain XYZ value, converting the standard automobile paint surface image into CIEXYZ color space to obtain XYZ value; S3, calculating the color appearance attributes of the automobile paint surface image of the measured area and the standard automobile paint surface image; S4. Based on the color appearance attribute, the color difference between the automobile paint surface image in the measured area and the standard automobile paint surface image is calculated, and the color difference value is compared with a preset value range. If it exceeds the preset range, it is determined that the automobile paint surface image in the measured area has a color difference defect.
[0016] In step S1, the lighting condition is D65 light source with the same brightness and illumination angle.
[0017] For example, the standard car paint image is collected from a standard color card. Since the car paint area is large, in order to ensure the comprehensiveness and accuracy of the comparison, the size of a single car paint image unit in the measured area is determined according to the size of the standard car paint image, and multiple car paint image units in the measured area are collected at different positions on the car paint surface. In one embodiment of the present applicant, the resolution of the car paint image in the measured area and the standard car paint image is 300 PPI, and the size of the standard car paint image and the car paint image unit in the measured area is 640 640 pixels.
[0018] For example, the standard car paint image captures the entire vehicle, and the car paint image of the measured area is consistent with the standard car paint image. In practice, it is impossible to complete the image capture in one go. In order to improve the convenience of subsequent comparison, the captured objects need to be spliced. In addition, non-painted surfaces (background) will inevitably be introduced into the image during the acquisition process. Image splicing includes the following processes: (1) The threshold method is used to segment the image. The threshold can be set manually or automatically determined based on the grayscale distribution of the pixels. That is, the grayscale value corresponding to the maximized inter-class variance in the grayscale distribution of the pixels in the image is the optimal threshold. (2) Stitch the segmented images together. You can set up repeated areas in the image acquisition and use the SIFT feature detection algorithm to select the feature vector with the minimum distance to achieve automatic stitching. For the color space conversion of the measured area automobile paint image (including the measured area automobile paint image unit and the measured area automobile paint image obtained by splicing) and the standard automobile paint image in step 2, in one embodiment of the present application, the conversion is from RGB to CIEXYZ, and the conversion process is as follows: (2) The RGB value needs to be linearly mapped from [0, 255] to the range [0, 1] using the following formula:
[0019] (2) Calculate the XYZ value based on the normalized RGB value. The formula is as follows:
[0020] The image color attribute calculation process in step S3 is as follows: (1) The environmental parameters and background parameters are determined. The calculation formulas for the color adaptation factor (F), the surrounding environment influence factor (c) and the color induction factor (Nc) of the environmental parameters are as follows:
[0021] Among them, Yb is the background brightness factor and Yw is the reference white field brightness factor.
[0022] The calculation formulas of the background induction factor n, nonlinear index z, brightness background factor Nbb, chroma background factor Ncb and brightness level adaptation factor FL of the background parameters are as follows:
[0023] Among them, L A is the adaptation field brightness.
[0024] (2) Input the tristimulus values, and calculate the input tristimulus values RGB according to the following formula using the XYZ values obtained in step 2, that is, convert the target tristimulus values XYZ into the human eye cone response signal RGB:
[0025] Where M16 is the color adaptation transformation matrix of the CAM16 color appearance model.
[0026] (3) The fitness factor D is determined. The fitness factor D is used to set the brightness perception of the adjusted color. The calculation formula is as follows:
[0027] The value range of D is 0-1. The higher the D value, the more suitable the color is for a bright background, while the lower the D value, the more suitable the color is for a dark background.
[0028] (4) Calculation of the color adaptation of the human eye to the light source, which represents the phenomenon that the human eye automatically adapts to the color as the chromaticity of the light source changes. The calculation formula is as follows:
[0029]
[0030] (5) Calculation of the post-adaptive cone response, characterizing the dynamic nonlinear characteristics of the cone response of the human visual system. The calculation formula is as follows:
[0031] If R C If is a negative value,
[0032] (6) Calculation of color appearance attribute variables Calculate the red-green component a, yellow-blue component b, hue angle h, and eccentricity et using the following formula:
[0033]
[0034] Calculate the colorless response factor A and the colorless response factor AW of the visual perception white point. The calculation formula is as follows:
[0035] Calculate the lightness J, apparent lightness Q, chroma C, apparent chroma M and saturation s. The calculation formula is as follows:
[0036]
[0037] Another embodiment of the present application provides a relatively simple calculation method as follows: In the CAM16 color appearance model, the calculation process of the red-green component a, the yellow-blue component b and the brightness J is as follows: 1. Normalize the XYZ values of the image to the range of [0, 100]. The calculation formula is as follows:
[0038] 2. Use the chromatic adaptation transformation matrix M CAM16 Convert the XYZ value to the XYZ value after color adaptation, that is, Xw, Yw, Zw:
[0039] Among them, MCAM16 involves multiple parameters and steps, refer to the detailed specifications and standards of CAM16.
[0040] 3. Calculate the red-green component a, yellow-blue component b and brightness J values in the CAM16 color model according to the following formula:
[0041]
[0042] The color difference calculation process based on the color appearance attribute in step S4 is as follows.
[0043]
[0044] In the formula: They are the brightness difference between the car paint in the tested area and the standard car paint, the difference between the red-green component, the difference between the yellow-blue component, It is the color difference between the car paint surface in the tested area and the standard car paint surface. The lightness Q, chroma C, chroma M and saturation s can also be introduced into the color difference calculation.
[0045] In terms of hardware system, this application requires the use of light source, color camera system, and machine vision processing system. The light source is D65 light source, the color camera system can be a commercially available high-definition digital camera in the field, and the machine vision processing system. You can use a professional image acquisition card, or you can choose a processing system based on a PC or industrial host. The specific selection of the hardware system does not fall within the scope of protection of this application.
[0046] It should be noted that, unless otherwise clearly specified and limited, the terms such as "install", "connect", "connect", "fix", "set" and the like should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral one; it can be a mechanical connection or an electrical connection; it can be a direct connection or an indirect connection through an intermediate medium, it can be the internal connection of two elements or the interaction relationship between two elements. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.
Claims
1. A method for identifying color difference of automobile paint based on color appearance model, characterized in that: The steps include: S1. Under the same lighting conditions, obtain the car paint surface image of the measured area and the standard car paint surface image; S2, converting the measured area automobile paint surface image into CIEXYZ color space after preprocessing to obtain XYZ value, converting the standard automobile paint surface image into CIEXYZ color space to obtain XYZ value; S3, calculating the color appearance attributes of the automobile paint surface image of the measured area and the standard automobile paint surface image; S4. Based on the color appearance attribute, the color difference between the automobile paint surface image in the measured area and the standard automobile paint surface image is calculated, and the color difference value is compared with a preset value range. If it exceeds the preset range, it is determined that the automobile paint surface image in the measured area has a color difference defect.
2. The method for detecting color difference defects on automobile paint surfaces according to claim 1, characterized in that: In step S1, the lighting condition is D65 light source with the same brightness and illumination angle.
3. The method for detecting color difference defects on automobile paint surfaces according to claim 1, characterized in that: For example, a standard automobile paint surface image is collected from a standard color card, and the size of a single automobile paint surface image unit in the measured area is determined according to the size of the standard automobile paint surface image, and multiple automobile paint surface image units in the measured area are collected at different positions on the automobile paint surface.
4. The method for detecting color difference defects of automobile paint according to claim 1, characterized in that: For example, if the standard automobile paint surface image is collected for the entire vehicle, the automobile paint surface image of the measured area is consistent with the standard automobile paint surface image.
5. The method for detecting color difference defects on automobile paint surfaces according to claim 1, characterized in that: The color space conversion process of the measured area car paint image and the standard car paint image in step 2 is as follows: (1) The RGB value needs to be linearly mapped from [0, 255] to [0, 1] using the following formula: (2) Calculate the XYZ value based on the normalized RGB value. The formula is as follows.
6. The method for detecting color difference defects on automobile paint surfaces according to claim 1, characterized in that: In step S3, the image color attributes at least include red-green component a, yellow-blue component b and brightness J.
7. The method for detecting color difference defects on automobile paint surfaces according to claim 1, characterized in that: The color difference calculation process based on the color appearance attribute in step S4 is as follows: In the formula: They are the brightness difference between the car paint in the tested area and the standard car paint, the difference between the red-green component, the difference between the yellow-blue component, It is the color difference between the car paint surface in the tested area and the standard car paint surface.
Citation Information
Patent Citations
Door and window color difference detection device and color difference detection algorithm
CN114184278A
Textile color difference detection method based on image data
CN117372334A
Image processing method and device, storage medium and electronic equipment
CN117611509A
Lateral face color measurement device and control method
JP2021135098A
