Chicken skin color evaluation method and system based on machine vision

Abstract

The invention discloses a chicken skin color evaluation method and system based on machine vision, and the method comprises the steps: obtaining a plurality of original images through employing a multi-angle multi-light-source collection device, and obtaining a comprehensive image based on the original images; preprocessing the comprehensive image to obtain a processed image; performing feature extraction on the processed image to obtain a skin color feature image; and based on the skin color feature image, constructing a machine vision model to complete chicken skin color evaluation. According to the method, through a method of combining multi-stage image processing and machine learning, key technical problems in chicken skin color evaluation are effectively solved, the accuracy and reliability of evaluation are improved, and important technical support is provided for chicken quality control.

Description

Technical Field

[0001] The present invention relates to the field of information technology, and in particular to a chicken skin color assessment method and system based on machine vision. Background Art

[0002] In the chicken processing process, the assessment of chicken skin color is an important quality control link. Traditional manual assessment methods have problems such as strong subjectivity and low efficiency, which are difficult to meet the needs of large-scale production. In order to solve this problem, it is urgent to develop a chicken skin color assessment method based on machine vision. However, factors such as the uneven surface of chicken skin and variable lighting conditions have brought great challenges to image acquisition and processing. First, the wrinkles and hair follicles on the surface of chicken skin will form a large number of shadows and highlights, resulting in local overbrightness or overdarkness of the image, affecting the accuracy of feature extraction. Secondly, the color distribution of chicken skin is uneven, the skin color of different parts of the same chicken is quite different, and the skin color of chickens of different breeds and different growth stages is also significantly different, which brings difficulties to skin color assessment. Furthermore, the complex lighting conditions at industrial sites, such as cloudy days, sunny days, lights, etc., will cause the color of chicken skin images to shift, affecting the accuracy of skin color assessment. Finally, there may be impurities such as chicken feathers, blood spots, stains, etc. in chicken skin images, which puts higher requirements on image processing algorithms. Therefore, how to overcome the above technical difficulties and develop a chicken skin color assessment method with strong robustness and wide adaptability is a key issue that needs to be solved urgently. Summary of the invention

[0003] To solve the above problems, the present invention proposes a chicken skin color assessment method and system to address the problems of uneven chicken skin surface, variable lighting conditions, uneven color distribution, and differences in breeds and growth stages.

[0004] To achieve the above object, the present invention provides a chicken skin color assessment method based on machine vision, the steps comprising:

[0005] A multi-angle and multi-light source acquisition device is used to acquire a plurality of original images, and a composite image is obtained based on the original images;

[0006] Preprocessing the composite image to obtain a processed image;

[0007] Performing feature extraction on the processed image to obtain a skin color feature image;

[0008] Based on the skin color feature image, a machine vision model is constructed to complete the chicken skin color assessment.

[0009] Preferably, the method for obtaining the composite image includes: using a multi-angle, multi-light source image acquisition device to obtain a plurality of original images, and fusing the original images into the composite image through image fusion technology.

[0010] Preferably, the method for performing the preprocessing includes: processing the composite image using an adaptive histogram equalization algorithm for shadow and highlight areas, adjusting the brightness and contrast of the image, and obtaining the processed image, which is used for subsequent extraction of skin color features.

[0011] Preferably, the method for extracting the feature includes: using color space conversion technology to convert the processed image from RGB space to Lab color space in view of uneven color distribution and skin color differences, extracting a and b channels of the processed image in Lab color space as skin color features, and obtaining the skin color feature image, which is used for skin color assessment.

[0012] Preferably, the method for completing the chicken skin color assessment includes: using a classification algorithm based on a support vector machine to classify the skin color feature image to obtain a skin color assessment result, and comparing the skin color assessment result with a pre-established skin color standard library to complete the assessment of the chicken skin color.

[0013] The present invention also provides a chicken skin color assessment system based on machine vision, the system is used to implement the above method, including: a collection module, a processing module, an extraction module and an assessment module;

[0014] The acquisition module is used to acquire multiple original images and obtain a comprehensive image based on the original images;

[0015] The processing module is used to pre-process the comprehensive image to obtain a processed image;

[0016] The extraction module is used to extract features from the processed image to obtain a skin color feature image;

[0017] The assessment module is used to complete the chicken skin color assessment based on the skin color feature image using a support vector machine.

[0018] Preferably, the workflow of the acquisition module includes: using a multi-angle, multi-light source image acquisition device to acquire multiple original images, and fusing the original images into the composite image through image fusion technology.

[0019] Preferably, the workflow of the processing module includes: processing the composite image with an adaptive histogram equalization algorithm for shadow and highlight areas, adjusting the brightness and contrast of the image, and obtaining the processed image, which is used for subsequent extraction of skin color features.

[0020] Compared with the prior art, the present invention has the following beneficial effects:

[0021] The present invention effectively solves the key technical problems in chicken skin color assessment through a method combining multi-stage image processing and machine learning, improves the accuracy and reliability of the assessment, and provides important technical support for chicken quality control. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] In order to more clearly illustrate the technical solution of the present invention, the following briefly introduces the drawings required for use in the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.

[0023] Figure 1 The figure is a schematic diagram of a method flow of an embodiment of the present invention. DETAILED DESCRIPTION

[0024] 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.

[0025] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the present invention is further described in detail below with reference to the accompanying drawings and specific embodiments.

[0026] Embodiment 1

[0027] like Figure 1 FIG. 1 is a schematic diagram of the method flow of this embodiment, and the steps include:

[0028] S1. Use multi-angle and multi-light source acquisition equipment to obtain multiple original images, and obtain a comprehensive image based on the original images.

[0029] In the image acquisition stage, in view of the uneven surface of the chicken skin and the changeable lighting conditions, a multi-angle, multi-light source image acquisition device is used to obtain multiple original images, and the original images are fused into a comprehensive image through image fusion technology. In this embodiment, 8 light source positions with different angles and directions are selected, with an angle range of 0° to 360° and an interval of 45°. For each light source position, a high-resolution camera is used to obtain the original image of the chicken skin surface.

[0030] S2. Preprocess the comprehensive image to obtain a processed image.

[0031] In the image preprocessing stage, the adaptive histogram equalization algorithm is used to process the composite image for the shadow and highlight areas, adjust the brightness and contrast of the image, and obtain a brightness balanced image to ensure the accuracy of subsequent feature extraction.

[0032] Specifically, the composite image is first preprocessed by adaptive histogram equalization. The brightness and contrast parameters of the image are determined by analyzing the histogram distribution of the image. The brightness range of the image is set to 0 to 255. By calculating the histogram of the image, it is found that the pixel values ​​of the shadow area are concentrated between 0 and 50, while the pixel values ​​of the highlight area are concentrated between 200 and 255. For the shadow area, its brightness is increased to between 50 and 100 through adaptive adjustment, so that the details of the shadow area can be displayed. For the highlight area, the brightness is adjusted to between 150 and 200 by reducing the brightness to avoid loss of details. Then, the adjusted image is processed by histogram equalization to make the brightness distribution of the image more balanced and retain more image detail information. By evaluating the contrast and brightness of the original image and the processed image, if the preset brightness balance threshold is not reached, the parameters are returned to be readjusted.

[0033] S3. Perform feature extraction on the processed image to obtain a skin color feature image.

[0034] In the feature extraction stage, color space conversion technology is used to convert the processed image from RGB space to Lab color space to address uneven color distribution and skin color differences. The a and b channels in the processed image are extracted as skin color features to obtain a skin color feature image for skin color assessment.

[0035] Specifically, first, according to the pixel values ​​of the a channel and the b channel of the Lab color space, the chromaticity value and the brightness value of each pixel are calculated to form a chromaticity brightness image. Among them, the chromaticity value can be obtained by calculating the square root of the sum of the squares of the a channel and the b channel, and the brightness value is directly taken from the L channel. Next, for the chromaticity brightness image, the Gaussian filter algorithm is used to perform image smoothing to remove noise interference in the image. The standard deviation of the Gaussian filter is set to 5 to ensure the smoothing effect while retaining the detailed features of the image. After smoothing, the obtained chromaticity brightness image is clearer. Then, the skin color area is extracted from the smoothed chromaticity brightness image. By judging whether the chromaticity value of each pixel falls within the preset skin color range; in the embodiment, the a channel is set between 20 and 30, and the b channel is set between 10 and 20, to obtain the skin color area image. According to the skin color area image, the area proportion of the skin color area is calculated. If the area proportion is greater than 5, it is judged that the skin color area proportion is large and the skin color is more uniform.

[0036] S4. Based on the skin color feature image, a machine vision model is constructed to complete the chicken skin color assessment.

[0037] In this embodiment, a classification algorithm based on a support vector machine is used to classify the skin color feature image to obtain a skin color assessment result, and the skin color assessment result is compared with a pre-established skin color standard library to complete the assessment of the chicken skin color.

[0038] The color distribution feature vector is extracted by acquiring the skin color feature image; the extracted skin color distribution feature vector is input into a pre-trained support vector machine model, and the input skin color feature vector is classified by the support vector machine model.

[0039] The support vector machine model classifies the input skin color feature vector through the learned classification boundary and outputs the quality level of the skin color. In this implementation, the quality level is divided into 1 to 5 levels, of which level 3 and above are qualified; the skin color quality level output by the support vector machine is compared with the pre-established skin color standard library. If the determined quality level is higher than or equal to level 3, it is marked as qualified; otherwise, it is marked as unqualified. Finally, according to the qualified or unqualified mark of the skin color sample, the results of all samples are comprehensively evaluated, and the skin color evaluation report is output, which includes the quality level and qualified status of each sample.

[0040] Embodiment 2

[0041] The present embodiment also provides a chicken skin color assessment system based on machine vision, including: an acquisition module, a processing module, an extraction module and an assessment module; the acquisition module is used to obtain multiple original images and obtain a comprehensive image based on the original images; the processing module is used to pre-process the comprehensive image to obtain a processed image; the extraction module is used to extract features from the processed image to obtain a skin color feature image; the assessment module is used to complete the chicken skin color assessment based on the skin color feature image using a support vector machine.

[0042] The following will describe in detail how the present invention solves technical problems in real life based on the present embodiment.

[0043] First, the acquisition module uses a multi-angle and multi-light source acquisition device to obtain multiple original images, and obtains a comprehensive image based on the original images.

[0044] In the image acquisition stage, in view of the uneven surface of the chicken skin and the changeable lighting conditions, a multi-angle, multi-light source image acquisition device is used to obtain multiple original images, and the original images are fused into a comprehensive image through image fusion technology. In this embodiment, 8 light source positions with different angles and directions are selected, with an angle range of 0° to 360° and an interval of 45°. For each light source position, a high-resolution camera is used to obtain the original image of the chicken skin surface.

[0045] The processing module pre-processes the comprehensive image to obtain a processed image.

[0046] In the image preprocessing stage, the adaptive histogram equalization algorithm is used to process the composite image for the shadow and highlight areas, adjust the brightness and contrast of the image, and obtain a brightness balanced image to ensure the accuracy of subsequent feature extraction.

[0047] Specifically, the composite image is first preprocessed by adaptive histogram equalization. The brightness and contrast parameters of the image are determined by analyzing the histogram distribution of the image. The brightness range of the image is set to 0 to 255. By calculating the histogram of the image, it is found that the pixel values ​​of the shadow area are concentrated between 0 and 50, while the pixel values ​​of the highlight area are concentrated between 200 and 255. For the shadow area, its brightness is increased to between 50 and 100 through adaptive adjustment, so that the details of the shadow area can be displayed. For the highlight area, the brightness is adjusted to between 150 and 200 by reducing the brightness to avoid loss of details. Then, the adjusted image is processed by histogram equalization to make the brightness distribution of the image more balanced and retain more image detail information. By evaluating the contrast and brightness of the original image and the processed image, if the preset brightness balance threshold is not reached, the parameters are returned to be readjusted.

[0048] The extraction module performs feature extraction on the processed image to obtain a skin color feature image.

[0049] In the feature extraction stage, color space conversion technology is used to convert the processed image from RGB space to Lab color space to address uneven color distribution and skin color differences. The a and b channels in the processed image are extracted as skin color features to obtain a skin color feature image for skin color assessment.

[0050] Specifically, first, according to the pixel values ​​of the a channel and the b channel of the Lab color space, the chromaticity value and the brightness value of each pixel are calculated to form a chromaticity brightness image. Among them, the chromaticity value can be obtained by calculating the square root of the sum of the squares of the a channel and the b channel, and the brightness value is directly taken from the L channel. Next, for the chromaticity brightness image, the Gaussian filter algorithm is used to perform image smoothing to remove noise interference in the image. The standard deviation of the Gaussian filter is set to 5 to ensure the smoothing effect while retaining the detailed features of the image. After smoothing, the obtained chromaticity brightness image is clearer. Then, the skin color area is extracted from the smoothed chromaticity brightness image. By judging whether the chromaticity value of each pixel falls within the preset skin color range; in the embodiment, the a channel is set between 20 and 30, and the b channel is set between 10 and 20, to obtain the skin color area image. According to the skin color area image, the area proportion of the skin color area is calculated. If the area proportion is greater than 5, it is judged that the skin color area proportion is large and the skin color is more uniform.

[0051] Finally, the assessment module builds a machine vision model based on the skin color feature image to complete the chicken skin color assessment.

[0052] In this embodiment, a classification algorithm based on a support vector machine is used to classify the skin color feature image to obtain a skin color assessment result, and the skin color assessment result is compared with a pre-established skin color standard library to complete the assessment of the chicken skin color.

[0053] The color distribution feature vector is extracted by acquiring the skin color feature image; the extracted skin color distribution feature vector is input into a pre-trained support vector machine model, and the input skin color feature vector is classified by the support vector machine model.

[0054] The support vector machine model classifies the input skin color feature vector through the learned classification boundary and outputs the quality level of the skin color. In this implementation, the quality level is divided into 1 to 5 levels, of which level 3 and above are qualified; the skin color quality level output by the support vector machine is compared with the pre-established skin color standard library. If the determined quality level is higher than or equal to level 3, it is marked as qualified; otherwise, it is marked as unqualified. Finally, according to the qualified or unqualified mark of the skin color sample, the results of all samples are comprehensively evaluated, and the skin color evaluation report is output, which includes the quality level and qualified status of each sample.

[0055] The embodiments described above are only descriptions of the preferred embodiments of the present invention and are not intended to limit the scope of the present invention. Without departing from the design spirit of the present invention, various modifications and improvements made to the technical solutions of the present invention by ordinary technicians in this field should all fall within the protection scope determined by the claims of the present invention.

Claims

1. A method for evaluating chicken skin color based on machine vision, characterized in that the steps include: A multi-angle and multi-light source acquisition device is used to acquire a plurality of original images, and a composite image is obtained based on the original images; Preprocessing the composite image to obtain a processed image; Performing feature extraction on the processed image to obtain a skin color feature image; Based on the skin color feature image, a machine vision model is constructed to complete the chicken skin color assessment.

2. The chicken skin color assessment method based on machine vision according to claim 1, characterized in that: The method for obtaining the composite image includes: using a multi-angle, multi-light source image acquisition device to obtain a plurality of original images, and fusing the original images into the composite image through an image fusion technology.

3. The chicken skin color assessment method based on machine vision according to claim 1, characterized in that: The method for performing the preprocessing includes: processing the composite image using an adaptive histogram equalization algorithm for the shadow and highlight areas, adjusting the brightness and contrast of the image, and obtaining the processed image, which is used for subsequent extraction of skin color features.

4. The chicken skin color assessment method based on machine vision according to claim 1, characterized in that: The method for extracting the feature includes: using color space conversion technology to convert the processed image from RGB space to Lab color space in view of uneven color distribution and skin color differences, extracting a and b channels of the processed image in Lab color space as skin color features, and obtaining the skin color feature image, which is used for skin color assessment.

5. The chicken skin color assessment method based on machine vision according to claim 1, characterized in that: The method for completing the chicken skin color assessment includes: using a classification algorithm based on a support vector machine to classify the skin color feature image to obtain a skin color assessment result, and comparing the skin color assessment result with a pre-established skin color standard library to complete the chicken skin color assessment.

6. A chicken skin color assessment system based on machine vision, the system being used to implement the method according to any one of claims 1 to 5, characterized in that: include: Acquisition module, processing module, extraction module and assessment module; The acquisition module is used to acquire multiple original images and obtain a comprehensive image based on the original images; The processing module is used to pre-process the comprehensive image to obtain a processed image; The extraction module is used to extract features from the processed image to obtain a skin color feature image; The assessment module is used to complete the chicken skin color assessment based on the skin color feature image using a support vector machine.

7. The machine vision-based chicken skin color assessment system according to claim 6, characterized in that: The workflow of the acquisition module includes: using a multi-angle, multi-light source image acquisition device to acquire multiple original images, and fusing the original images into the comprehensive image through image fusion technology.

8. The machine vision-based chicken skin color assessment system according to claim 6, characterized in that: The workflow of the processing module includes: for the shadow and highlight areas, using an adaptive histogram equalization algorithm to process the composite image, adjusting the brightness and contrast of the image, and obtaining the processed image, and the processed image is used for subsequent extraction of skin color features.

Images