A jaundice detection method and system based on visual recognition

Abstract

The present invention discloses a jaundice detection method and system based on visual recognition, and the detection method comprises: obtaining first image data of a person to be detected under white light shooting illumination, and obtaining second image data of the person to be detected under blue light shooting illumination; performing image color extraction on the first image data based on HSV color space to obtain color data of the first image data; performing comparison on the color data of the first image data with the image chromaticity data of a standard image to obtain first comparison data; extracting the proportion of green component and yellow component from the second image data based on Lab color space recognition, and obtaining second comparison data based on shooting parameter settings; obtaining an image yellow chromaticity change evaluation value of the person to be detected based on the first comparison data and the second comparison data, and outputting a detection result based on the evaluation value. By performing image analysis on the yellow component change value, the image detection accuracy is improved, thereby improving the evaluation reliability of the occurrence of jaundice symptoms.

Description

Technical Field

[0001] The present invention relates to the field of morning checkup technology, and in particular to a jaundice detection method and system based on visual recognition. Background Art

[0002] Jaundice is a phenomenon characterized by elevated serum bilirubin levels, which leads to yellowing of the sclera, skin, and mucous membranes. Jaundice symptoms can be caused by a variety of related diseases. Conducting image detection on relevant personnel in densely populated places such as campuses and parks, and assessing the presence of jaundice in the tested personnel, can help to monitor the health status of personnel.

[0003] Due to the existing image analysis and processing methods, which are generally based on deep neural network models to identify and detect jaundice symptoms on the human body surface, the image processing requires a large amount of computational effort, and the main manifestation of jaundice symptoms is the change of the yellow color component. The detection method based on deep learning is not effective in color recognition and differentiation, resulting in insufficient reliability of the detection equipment in detecting jaundice symptoms of the people being tested. Summary of the Invention

[0004] The purpose of the present invention is to overcome the shortcomings of the existing technology. The present invention provides a jaundice detection method and system based on visual recognition. By performing image analysis on the change value of the yellow component of the person to be detected, the accuracy of image detection of jaundice symptoms is improved, thereby improving the reliability of the assessment of the occurrence of jaundice symptoms.

[0005] The present invention provides a jaundice detection method based on visual recognition, the detection method comprising:

[0006] Acquire first image data of the person to be detected under white light, and acquire second image data of the person to be detected under blue light;

[0007] Performing image color extraction on the first image data based on the HSV color space to obtain color data of the first image data;

[0008] Comparing the color data of the first image data with the image chromaticity data of a standard image to obtain first comparison data;

[0009] Identify the second image data based on the Lab color space to extract the proportions of the green component and the yellow component, and obtain second contrast data based on the shooting parameter settings;

[0010] An evaluation value of yellow chromaticity change of the image of the person to be detected is obtained based on analysis of the first comparison data and the second comparison data, and a detection result is output based on the evaluation value.

[0011] Furthermore, the acquiring of first image data of the person to be detected under white light and the acquiring of second image data of the person to be detected under blue light includes:

[0012] The face of the person to be detected is photographed under white light to obtain iris image data of the person to be detected, which is marked as first image data;

[0013] The hand of the person to be detected is photographed under blue light, and palm image information of the person to be detected is obtained and marked as second image data.

[0014] Furthermore, performing image color extraction on the first image data based on the HSV color space to obtain color data of the first image data includes:

[0015] Inputting the first image data into the HSV color space for image segmentation, extracting the target area in the first image data, and obtaining an image of the target area;

[0016] Color recognition is performed on the target area image to obtain color data of the first image data.

[0017] Furthermore, performing color recognition on the target area image to obtain color data of the first image data includes:

[0018] Set fixed value parameters and single variables in HSV space, adjust the value of the single variable with preset intervals, and generate color recognition frames;

[0019] Color extraction is performed on the target area image based on the color recognition frame to obtain the color chromaticity of the target area image.

[0020] Furthermore, setting a fixed value parameter and a single variable in the HSV space, adjusting the value of the single variable with a preset interval, and generating a color recognition frame includes:

[0021] In the HSV space, set the S and V values ​​to fixed values, set the H value to a variable, and generate the H value color recognition box;

[0022] The H value of the target area image is obtained based on the H value color recognition frame.

[0023] Furthermore, the comparing the color data of the first image data with the image chromaticity data of the standard image to obtain the first comparison data includes:

[0024] Converting the acquired color data of the first image data from the HSV format to the RGB format;

[0025] A standard image corresponding to the detection area is extracted from the database, and the color data of the first image data is compared with the color data of the standard image to obtain first comparison data.

[0026] Furthermore, the extracting the proportions of the green component and the yellow component from the second image data based on the Lab color space identification and obtaining the second comparison data based on the shooting parameter settings includes:

[0027] The second image data is input into a Lab color space model to extract color components, the second image data is adjusted based on the Lab color space, and color component data in the second image data is extracted.

[0028] Furthermore, the extracting of the proportions of the green component and the yellow component from the second image data based on the Lab color space identification and the obtaining of the second comparison data based on the shooting parameter settings further include:

[0029] Extracting the original color component of the hand position of the second image data of the person to be detected according to the hand image data of the person to be detected recorded in the database;

[0030] The original color component of the shooting environment of the second image data is obtained according to the blue light illumination parameter setting of the second image data.

[0031] Furthermore, obtaining an evaluation value of yellow chromaticity change of the image of the person to be detected based on the first comparison data and the second comparison data, and outputting a detection result based on the evaluation value includes:

[0032] The calculation formula of the chromaticity change evaluation value is:

[0033]

[0034] Among them, θ is the yellow chromaticity change evaluation value, μ1 is the weight of the first comparison data, μ2 is the weight of the second comparison data, d1 is the first comparison data, d2 is the second comparison data, and γ0 is the initial yellow chromaticity value of the person to be detected.

[0035] The present invention also provides a jaundice detection system based on visual recognition, the detection system comprising:

[0036] Image acquisition component: used to acquire first image data of the person to be detected under white light shooting illumination, and acquire second image data of the person to be detected under blue light shooting illumination;

[0037] An HSV model processing component is used to extract the image color of the first image data based on the HSV color space to obtain the color data of the first image data;

[0038] A first comparison component is configured to compare the color data of the first image data with the image chromaticity data of a standard image to obtain first comparison data;

[0039] A Lab model processing component is used to identify the second image data based on the Lab color space, extract the proportion of the green component and the yellow component, and obtain the second contrast data based on the shooting parameter settings;

[0040] Evaluation component: used to obtain an evaluation value of the yellow chromaticity change of the image of the person to be detected based on the first comparison data and the second comparison data, and output a detection result based on the evaluation value.

[0041] The present invention provides a jaundice detection method and system based on visual recognition. The method obtains first image data and second image data of a person to be detected through a detection device, combines the color data of the standard image data of the person to be detected with color data comparison, and analyzes the iris color data and hand skin color data of the person to be detected. The method can accurately obtain the change in the yellow color component of the detection image of the person to be detected, thereby improving the reliability of the assessment of jaundice symptoms in the person to be detected. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] Figure 1 is a flow chart of a jaundice detection method based on visual recognition in an embodiment of the present invention;

[0043] Figure 2 1 is a schematic diagram of the workflow of the jaundice detection method based on visual recognition in an embodiment of the present invention;

[0044] Figure 3 Schematic diagram of the structure of a jaundice detection system based on visual recognition in an embodiment of the present invention. DETAILED DESCRIPTION

[0045] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. 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 making any creative efforts shall fall within the scope of protection of the present invention.

[0046] Example 1:

[0047] Figure 1 The figure shows a flow chart of a jaundice detection method based on visual recognition in an embodiment of the present invention. The detection method includes:

[0048] S11: Acquire first image data of the person to be detected under white light, and acquire second image data of the person to be detected under blue light.

[0049] A voice prompt is set in the detection device to guide the person to be detected to cooperate with the detection device to perform image capture detection. The camera component of the detection device is equipped with a contour recognition module, so that the detection device can obtain the shooting contour of the person to be detected and perform a shooting operation on the person to be detected.

[0050] Furthermore, the detection equipment can be an access control device, a morning check-up device, a physical examination device, etc., to meet the needs of different usage scenarios.

[0051] Specifically, when the detection device is in operation, the infrared detection module of the detection device detects whether the person to be detected enters the image shooting area. If so, the person to be detected is guided by voice prompts to face the shooting camera with his face, and the shooting module of the detection device performs preliminary contour detection on the person to be detected, and identifies that the person to be detected is in the correct shooting position and shooting posture. The detection device can illuminate the person to be detected by soft white light, so that the detection device can shoot the face of the person to be detected under white light conditions to obtain the first image data of the person to be detected.

[0052] Specifically, after the detection device obtains the first image data of the person to be detected, the detection device guides the person to be detected to place his hand within the detection area of ​​the detection device based on the voice prompt of the detection device, and preliminarily identifies the hand outline of the person to be detected based on the shooting module, and determines that the hand of the person to be detected is located in the shooting area. The shooting component of the detection device illuminates the user's hand with blue light, so that the detection device can shoot the hand of the person to be detected under the condition of blue light, thereby obtaining the second image data of the person to be detected.

[0053] S12: Perform image color extraction on the first image data based on the HSV color space to obtain color data of the first image data.

[0054] The first image data is input into the HSV color space for image processing, and the H value, S value, and V value of the first image data are identified and extracted in the HSV color space, thereby obtaining the HSV value of the target area in the first image data.

[0055] The HSV color space is a spatial model for representing the color state of an image. In the HSV color space, each color is represented by hue (H), saturation (S), and brightness (V). That is, the image color parameters in the HSV color space model are hue (H), saturation (S), and brightness (V). The color information, that is, the position of the spectral color, is represented by the hue H parameter, and the value range of H is 0° to 360°.

[0056] Furthermore, in the HSV color space, starting from red and counting in a counterclockwise direction, red is 0°, green is 120°, blue is 240°, yellow is 60°, cyan is 180°, and purple is 300°.

[0057] Specifically, the first image data is input into the HSV color space for image segmentation, a target area in the first image data is extracted to obtain a target area image, and color recognition is performed on the target area image to obtain color data of the first image data.

[0058] Color extraction is performed on the target area image, and a color extraction frame is set in the HSV color space. By setting the S value and V value in the extraction frame to fixed values ​​and the H value to an increasing variable, the recognition range of the color recognition frame is refreshed by adjusting the specific numerical value of the H value inside the color recognition frame. By performing multiple color extractions on the target area image, several color extraction images with the target area outline are obtained, and the one with the highest overlap rate with the target area image outline is extracted from the several color extraction images as the color extraction image that completes the matching, and the corresponding H value is recorded as the H value of the target area image in the HSV color space.

[0059] Furthermore, the S value and the V value are sequentially set as variables through the same operation mode, and the S value and the V value of the target area image are extracted, thereby determining the HSV value of the target area image.

[0060] S13: Compare the color data of the first image data with the image chromaticity data of the standard image to obtain first comparison data.

[0061] Specifically, the color data of the acquired first image data is converted from the HSV format to the RGB format, and the RGB pixel values ​​of the color data of the first image data are acquired, so as to compare the color data of the first image data with the standard data in the database.

[0062] A standard image corresponding to the detection area is extracted from the database, and the color data of the first image data is compared with the color data of the standard image to obtain first comparison data.

[0063] The standard data corresponding to the first image data detection area is queried in the database, that is, the shooting detection position of the first image data is retrieved, the corresponding standard color image data is extracted, and the color data of the first image data is compared with the color data of the standard image to obtain the first comparison data.

[0064] Furthermore, based on the facial shooting information of the person to be detected, the iris color data of the person to be detected is obtained, and the iris color data of the person to be detected is compared and analyzed with the standard image data to obtain the difference between the iris color range of the person to be detected and the color range of the standard image data.

[0065] Specifically, the database inside the detection device stores basic image data corresponding to the person to be detected, that is, the basic image data of the person to be detected taken and stored in a healthy state, including facial information, iris images, hand images and other image data of various parts of the body of the person to be detected. The basic images pre-entered into the set of people to be detected are used as standard image data of the database, so that the body image data of the person to be detected can be accurately analyzed and judged.

[0066] Furthermore, an access control system equipped with the above-mentioned detection equipment is installed at the entrance and exit locations of campuses and industrial parks. While performing entry and exit management registration for personnel in the campus or park, it can also capture images and perform image detection and analysis for personnel in the campus or park, thereby improving the reliability of the access control system in health management of personnel in the campus and park.

[0067] S14: extracting the proportions of the green component and the yellow component from the second image data based on the Lab color space identification, and obtaining second comparison data based on the shooting parameter settings.

[0068] The second image data is input into a Lab color space model to extract color components, the second image data is adjusted based on the Lab color space, and color component data in the second image data is extracted.

[0069] Furthermore, the Lab color space is based on an international standard for color measurement established by the Commission International Eclairage (CIE). The Lab color model consists of three elements: lightness (L), color channel (a), and color channel (b). Colors in a range from dark green to gray to bright pink, and colors in b range from bright blue to gray to yellow. Based on the Lab color space, color component data of the second image data can be obtained, thereby obtaining color detection data of the person to be detected.

[0070] Specifically, according to the identity information of the person to be detected, the original color components of the person to be detected are marked in the second image data, and by querying the hand image data corresponding to the person to be detected in the database, the original color components of the hand position of the second image data of the person to be detected are extracted according to the hand image data of the person to be detected recorded in the database, and according to the blue light illumination parameter setting of the second image data, the original color components of the shooting environment of the second image data are obtained. Based on the color component setting of the shooting subject and the shooting environment of the second image data, the accuracy of the color component extraction of the second image data can be improved, thereby improving the accuracy of the second image data in detecting the physical condition of the person to be detected.

[0071] Furthermore, the initial value of the green color component in the second image data is converted based on the color components of the shooting subject and the shooting environment of the second image data. According to the green color component detection data fed back by the second image data in the Lab color space, the initial value of the green color component and the green color component detection data are compared to obtain the second comparison data.

[0072] S15: Analyze the first comparison data and the second comparison data to obtain an evaluation value of the yellow chromaticity change of the image of the person to be detected, and output a detection result based on the evaluation value.

[0073] The calculation formula of the chromaticity change evaluation value is:

[0074]

[0075] Among them, θ is the yellow chromaticity change evaluation value, μ1 is the weight of the first comparison data, μ2 is the weight of the second comparison data, d1 is the first comparison data, d2 is the second comparison data, and γ0 is the initial yellow chromaticity value of the person to be detected.

[0076] Furthermore, the values ​​of μ1 and μ2 are set according to the yellow chromaticity of the initial image of the person to be detected. Taking the HSV color space as an example, the hue value H of the yellow component ranges from 26 to 34, and the hue H value of the middle value yellow color component is 29. When the initial hue H value of the yellow component of the second image data of the person to be detected is less than or equal to 29, μ1 is taken as 0.5 and μ2 is taken as 0.5; if the initial hue H value of the yellow component of the second image data of the person to be detected is greater than 29, μ1 is taken as 0.7 and μ2 is taken as 0.3.

[0077] Furthermore, when θ≤0.2, the physical condition of the person to be tested is marked as normal; when θ>0.2, the physical condition of the person to be tested is marked as abnormal, and relevant detection information is generated and sent to relevant staff so that the relevant staff can conduct further inspections on the person to be tested.

[0078] An embodiment of the present invention provides a jaundice detection method based on visual recognition. The first image data and the second image data of the person to be detected are obtained through a detection device. The iris color data and the hand skin color data of the person to be detected are analyzed by combining the color data comparison of the standard image data of the person to be detected. The change in the yellow color component of the detection image of the person to be detected can be accurately obtained, thereby improving the reliability of the assessment of jaundice symptoms in the person to be detected.

[0079] Example 2:

[0080] Figure 2 The following is a schematic diagram of the workflow of a jaundice detection method based on visual recognition according to an embodiment of the present invention. The detection method is implemented based on a detection device and includes:

[0081] S101: Identify the person to be detected.

[0082] The infrared recognition module of the detection device detects the position of the person to be detected. When the person to be detected approaches the detection device, the detection device obtains the position of the person to be detected based on the infrared recognition module and guides the person to be detected to perform photo detection based on voice prompts.

[0083] Furthermore, the detection device can be connected to the access control system of the campus or park, so that when the person to be detected enters or leaves the campus or park, the identity information of the person to be detected can be identified based on the detection device to complete the access control punch-in operation.

[0084] S102: guiding the person to be tested to take a photo and perform the test.

[0085] Specifically, the detection equipment is provided with a voice module and a contour recognition module. Based on the voice module, voice reminders are given to the persons to be detected who are in the detection area of ​​the detection equipment. The language reminders guide the persons to be detected to face the camera of the construction equipment for shooting and detection.

[0086] Based on the contour recognition module, the facial contour of the person to be detected is preliminarily recognized to detect whether the person to be detected is in the correct shooting position, thereby ensuring that the facial information of the person to be detected is obtained.

[0087] When the person to be detected completes the shooting of facial information, a voice reminder is given to the person to be detected based on the voice module, so that the person to be detected displays his hands in the shooting area of ​​the detection device, so that the detection device can obtain the hand image information of the person to be detected.

[0088] S103: Acquire first image data.

[0089] Specifically, the face of the person to be detected is photographed under white light to obtain iris image data of the person to be detected, which is marked as first image data;

[0090] When the detection device is in operation, it detects whether a person to be detected enters the image shooting area based on the infrared detection module of the detection device. If so, the person to be detected is guided by voice prompts to face the shooting camera with his face, and the shooting module of the detection device performs preliminary contour detection on the person to be detected, and identifies that the person to be detected is in the correct shooting position and shooting posture. The detection device can illuminate the person to be detected by soft white light, so that the detection device can shoot the face of the person to be detected under white light conditions to obtain the first image data of the person to be detected.

[0091] Specifically, due to the symptoms of jaundice, the human iris may be yellowed, and the human iris is located at the front of the middle membrane of the eyeball. Based on the comparison of the white area of ​​the eye, the detection equipment can quickly locate the iris in the first detection image, thereby obtaining the target area image.

[0092] Based on the soft white light, the shooting brightness of the first image data can be increased to improve the accuracy of the color analysis of the target area image.

[0093] S104: Based on HSV color space analysis.

[0094] Performing image color extraction on the first image data based on the HSV color space to obtain color data of the first image data includes:

[0095] Inputting the first image data into the HSV color space for image segmentation, extracting the target area in the first image data, and obtaining an image of the target area;

[0096] Color recognition is performed on the target area image to obtain color data of the first image data.

[0097] Set fixed value parameters and single variables in HSV space, adjust the value of the single variable with preset intervals, and generate color recognition frames;

[0098] Color extraction is performed on the target area image based on the color recognition frame to obtain the color chromaticity of the target area image.

[0099] Specifically, the yellow color component of the target area image is identified and extracted. Based on the value range of the yellow color component in the HSV space, including: 26≤H≤34, 43≤S≤255, and 46≤V≤255, the S and V values ​​are set as fixed values ​​in the HSV space, and the H value is set as a variable. An H value color recognition frame is generated, and the H value of the target area image is obtained based on the H value color recognition frame. In this embodiment, the S value is set to 43 and the V value is set to 46, so that the H value increases sequentially within the range of 26 to 34, that is, the starting value of the H value is 26, the increment step is 1, and a plurality of color recognition frames are generated to obtain the H value of the yellow color component of the target area image.

[0100] Furthermore, the obtained H value is used as a constant, and the S value and the V value are set as variables in sequence, thereby obtaining the HSV value of the yellow color component in the target area image.

[0101] S105: Acquire first comparison data.

[0102] Converting the acquired color data of the first image data from the HSV format to the RGB format;

[0103] A standard image corresponding to the detection area is extracted from the database, and the color data of the first image data is compared with the color data of the standard image to obtain first comparison data.

[0104] The conversion formula between the HSV format and the RGB format is:

[0105]

[0106]

[0107] p=V*(1-S);

[0108] q = V*(1-f*S);

[0109] t=V*(1-(1-f)*S);

[0110] (r, g, b) = (V, t, p), Hi = 0;

[0111] (r, g, b) = (q, V, p), Hi = 1;

[0112] (r, g, b) = (p, V, t), Hi = 2;

[0113] (r, g, b) = (p, q, V), Hi = 3;

[0114] (r, g, b) = (t, p, V), Hi = 4;

[0115] (r, g, b) = (V, p, q), Hi = 5;

[0116] Among them, H, V and S are the H value, V value and S value displayed in the HSV color space of the color data of the first image data, Hi, f, p, q, t are calculation parameter symbols, and mod6 is the value obtained by dividing the H value by 60.

[0117] S106: Acquire second image data.

[0118] Specifically, the hand of the person to be detected is photographed under blue light, and palm image information of the person to be detected is obtained and marked as second image data.

[0119] After the detection device obtains the first image data of the person to be detected, the detection device guides the person to be detected to place his hand within the detection area of ​​the detection device based on the voice prompt of the detection device, and preliminarily identifies the outline of the hand of the person to be detected based on the shooting module, and determines that the hand of the person to be detected is located in the shooting area. The shooting component of the detection device illuminates the user's hand with blue light, so that the detection device can shoot the hand of the person to be detected under the condition of blue light, thereby obtaining the second image data of the person to be detected.

[0120] Furthermore, the symptoms of jaundice may manifest as yellowing of the skin color of the human hands. Since the human skin color can easily interfere with image analysis, the second image data appears greenish in a shooting environment based on blue light. By extracting and analyzing the yellow color component in the second image data, the changes in the yellow color component of the image of the person to be tested can be accurately obtained.

[0121] S107: Based on Lab color space analysis.

[0122] The second image data is input into a Lab color space model to extract color components, the second image data is adjusted based on the Lab color space, and color component data in the second image data is extracted.

[0123] Specifically, according to the identity information of the person to be detected, the original color components of the person to be detected are marked in the second image data, and the original color components of the hand position of the second image data of the person to be detected are extracted by querying the hand image data corresponding to the person to be detected in the database, and the hand image data of the person to be detected recorded in the database. According to the blue light illumination parameter setting of the second image data, the original color components of the shooting environment of the second image data are obtained. Based on the color component setting of the shooting subject and the shooting environment of the second image data, the accuracy of the color component extraction of the second image data can be improved, thereby improving the accuracy of the second image data in detecting the physical condition of the person to be detected.

[0124] S108: Acquire second comparison data.

[0125] Based on the color components of the shooting subject and the shooting environment of the second image data, the initial value of the green color component in the second image data is converted. According to the green color component detection data fed back by the second image data in the Lab color space, the initial value of the green color component and the green color component detection data are compared to obtain the second comparison data.

[0126] S109: Calculate the chromaticity change evaluation value.

[0127] The calculation formula of the chromaticity change evaluation value is:

[0128]

[0129] Among them, θ is the yellow chromaticity change evaluation value, μ1 is the weight of the first comparison data, μ2 is the weight of the second comparison data, d1 is the first comparison data, d2 is the second comparison data, and γ0 is the initial yellow chromaticity value of the person to be detected.

[0130] S110: Generate detection results.

[0131] Specifically, the probability of the person to be tested developing jaundice symptoms is analyzed based on the calculated value of θ. When θ≤0.2, the physical condition of the person to be tested is marked as normal; when θ>0.2, the physical condition of the person to be tested is marked as abnormal, and relevant detection information is generated and sent to relevant staff so that the relevant staff can conduct further inspections on the person to be tested.

[0132] An embodiment of the present invention provides a jaundice detection method based on visual recognition. By setting different lighting environments, different body parts of a person to be detected are photographed to obtain first image data and second image data of the person to be detected. The first image data is subjected to color analysis based on the HSV color space model to obtain yellow component data of the facial image of the person to be detected. The second image data is subjected to color analysis based on the Lab color space model to obtain yellow component data of the hand image of the person to be detected. Combined with the changes in the iris color and hand skin color of the person to be detected, the change in the yellow color component of the detection image of the person to be detected can be accurately obtained, thereby improving the reliability of the assessment of jaundice symptoms in the person to be detected.

[0133] Example 3:

[0134] Figure 3 FIG. 1 shows a schematic diagram of the structure of a jaundice detection system based on visual recognition in an embodiment of the present invention. The detection system includes:

[0135] The image acquisition component 10 acquires first image data of the person to be detected under white light, and acquires second image data of the person to be detected under blue light.

[0136] A voice prompt is set in the detection device to guide the person to be detected to cooperate with the detection device to perform image capture detection. The camera component of the detection device is equipped with a contour recognition module, so that the detection device can obtain the shooting contour of the person to be detected and perform a shooting operation on the person to be detected.

[0137] Furthermore, the detection equipment can be an access control device, a morning check-up device, a physical examination device, etc., to meet the needs of different usage scenarios.

[0138] Specifically, when the detection device is in operation, the infrared detection module of the detection device detects whether the person to be detected enters the image shooting area. If so, the person to be detected is guided by voice prompts to face the shooting camera with his face, and the shooting module of the detection device performs preliminary contour detection on the person to be detected, and identifies that the person to be detected is in the correct shooting position and shooting posture. The detection device can illuminate the person to be detected by soft white light, so that the detection device can shoot the face of the person to be detected under white light conditions to obtain the first image data of the person to be detected.

[0139] Specifically, after the detection device obtains the first image data of the person to be detected, the detection device guides the person to be detected to place his hand within the detection area of ​​the detection device based on the voice prompt of the detection device, and preliminarily identifies the hand outline of the person to be detected based on the shooting module, and determines that the hand of the person to be detected is located in the shooting area. The shooting component of the detection device illuminates the user's hand with blue light, so that the detection device can shoot the hand of the person to be detected under the condition of blue light, thereby obtaining the second image data of the person to be detected.

[0140] The HSV model processing component 20 performs image color extraction on the first image data based on the HSV color space to obtain color data of the first image data.

[0141] The first image data is input into the HSV color space for image processing, and the H value, S value, and V value of the first image data are identified and extracted in the HSV color space, thereby obtaining the HSV value of the target area in the first image data.

[0142] The HSV color space is a spatial model for representing the color state of an image. In the HSV color space, each color is represented by hue (H), saturation (S), and brightness (V). That is, the image color parameters in the HSV color space model are hue (H), saturation (S), and brightness (V). The color information, that is, the position of the spectral color, is represented by the hue H parameter, and the value range of H is 0° to 360°.

[0143] Furthermore, in the HSV color space, starting from red and counting in a counterclockwise direction, red is 0°, green is 120°, blue is 240°, yellow is 60°, cyan is 180°, and purple is 300°.

[0144] Specifically, the first image data is input into the HSV color space for image segmentation, a target area in the first image data is extracted to obtain a target area image, and color recognition is performed on the target area image to obtain color data of the first image data.

[0145] Color extraction is performed on the target area image, and a color extraction frame is set in the HSV color space. By setting the S value and V value in the extraction frame to fixed values ​​and the H value to an increasing variable, the recognition range of the color recognition frame is refreshed by adjusting the specific numerical value of the H value inside the color recognition frame. By performing multiple color extractions on the target area image, several color extraction images with the target area outline are obtained, and the one with the highest overlap rate with the target area image outline is extracted from the several color extraction images as the color extraction image that completes the matching, and the corresponding H value is recorded as the H value of the target area image in the HSV color space.

[0146] Furthermore, the S value and the V value are sequentially set as variables through the same operation mode, and the S value and the V value of the target area image are extracted, thereby determining the HSV value of the target area image.

[0147] The first comparison component 30 compares the color data of the first image data with the image chromaticity data of the standard image to obtain first comparison data.

[0148] Specifically, the color data of the acquired first image data is converted from the HSV format to the RGB format, and the RGB pixel values ​​of the color data of the first image data are acquired, so as to compare the color data of the first image data with the standard data in the database.

[0149] A standard image corresponding to the detection area is extracted from the database, and the color data of the first image data is compared with the color data of the standard image to obtain first comparison data.

[0150] The standard data corresponding to the first image data detection area is queried in the database, that is, the shooting detection position of the first image data is retrieved, the corresponding standard color image data is extracted, and the color data of the first image data is compared with the color data of the standard image to obtain the first comparison data.

[0151] Furthermore, based on the facial shooting information of the person to be detected, the iris color data of the person to be detected is obtained, and the iris color data of the person to be detected is compared and analyzed with the standard image data to obtain the difference between the iris color range of the person to be detected and the color range of the standard image data.

[0152] Specifically, the database inside the detection device stores basic image data corresponding to the person to be detected, that is, the basic image data of the person to be detected taken and stored in a healthy state, including facial information, iris images, hand images and other image data of various parts of the body of the person to be detected. The basic images pre-entered into the set of people to be detected are used as standard image data of the database, so that the body image data of the person to be detected can be accurately analyzed and judged.

[0153] Furthermore, an access control system equipped with the above-mentioned detection equipment is installed at the entrance and exit locations of campuses and industrial parks. While performing entry and exit management registration for personnel in the campus or park, it can also capture images and perform image detection and analysis for personnel in the campus or park, thereby improving the reliability of the access control system in health management of personnel in the campus and park.

[0154] The Lab model processing component 40 is used to identify the second image data based on the Lab color space, extract the proportions of the green component and the yellow component, and obtain second contrast data based on the shooting parameter settings.

[0155] The second image data is input into a Lab color space model to extract color components, the second image data is adjusted based on the Lab color space, and color component data in the second image data is extracted.

[0156] Furthermore, the Lab color space is based on an international standard for color measurement established by the Commission International Eclairage (CIE). The Lab color model consists of three elements: lightness (L), color channel (a), and color channel (b). Colors in a range from dark green to gray to bright pink, and colors in b range from bright blue to gray to yellow. Based on the Lab color space, color component data of the second image data can be obtained, thereby obtaining color detection data of the person to be detected.

[0157] Specifically, according to the identity information of the person to be detected, the original color components of the person to be detected are marked in the second image data, and the original color components of the hand position of the second image data of the person to be detected are extracted by querying the hand image data corresponding to the person to be detected in the database, and the hand image data of the person to be detected recorded in the database. According to the blue light illumination parameter setting of the second image data, the original color components of the shooting environment of the second image data are obtained. Based on the color component setting of the shooting subject and the shooting environment of the second image data, the accuracy of the color component extraction of the second image data can be improved, thereby improving the accuracy of the second image data in detecting the physical condition of the person to be detected.

[0158] Furthermore, the initial value of the green color component in the second image data is converted based on the color components of the shooting subject and the shooting environment of the second image data. According to the green color component detection data fed back by the second image data in the Lab color space, the initial value of the green color component and the green color component detection data are compared to obtain the second comparison data.

[0159] Evaluation component 50: obtains an evaluation value of the yellow chromaticity change of the image of the person to be detected based on the first comparison data and the second comparison data, and outputs a detection result based on the evaluation value.

[0160] The calculation formula of the chromaticity change evaluation value is:

[0161]

[0162] Among them, θ is the yellow chromaticity change evaluation value, μ1 is the weight of the first comparison data, μ2 is the weight of the second comparison data, d1 is the first comparison data, d2 is the second comparison data, and γ0 is the initial yellow chromaticity value of the person to be detected.

[0163] Furthermore, the values ​​of μ1 and μ2 are set according to the yellow chromaticity of the initial image of the person to be detected. Taking the HSV color space as an example, the hue value H of the yellow component ranges from 26 to 34, and the hue H value of the middle value yellow color component is 29. When the initial hue H value of the yellow component of the second image data of the person to be detected is less than or equal to 29, μ1 is taken as 0.5 and μ2 is taken as 0.5; if the initial hue H value of the yellow component of the second image data of the person to be detected is greater than 29, μ1 is taken as 0.7 and μ2 is taken as 0.3.

[0164] Furthermore, when θ≤0.2, the physical condition of the person to be tested is marked as normal; when θ>0.2, the physical condition of the person to be tested is marked as abnormal, and relevant detection information is generated and sent to relevant staff so that the relevant staff can conduct further inspections on the person to be tested.

[0165] An embodiment of the present invention provides a jaundice detection system based on visual recognition. The system obtains first image data and second image data of a person to be detected through a detection device, compares the color data of the standard image data of the person to be detected, and analyzes the iris color data and hand skin color data of the person to be detected. The system can accurately obtain the change in the yellow color component of the detection image of the person to be detected, thereby improving the reliability of the assessment of jaundice symptoms in the person to be detected.

[0166] In addition, the embodiments of the present invention are introduced in detail above. Specific examples are used herein to illustrate the principles and implementation methods of the present invention. The description of the above embodiments is only used to help understand the method of the present invention and its core idea. At the same time, for those skilled in the art, according to the idea of ​​the present invention, there will be changes in the specific implementation methods and application scope. In summary, the content of this specification should not be understood as limiting the present invention.

Claims

1. A jaundice detection method based on visual recognition, characterized in that: The detection method comprises: Acquire first image data of the person to be detected under white light, and acquire second image data of the person to be detected under blue light; Performing image color extraction on the first image data based on the HSV color space to obtain color data of the first image data; Comparing the color data of the first image data with the image chromaticity data of the standard image to obtain first comparison data; Recognize the second image data based on the Lab color space to extract the proportion of the green component and the yellow component, and obtain the second comparison data based on the shooting parameter settings; An evaluation value of the yellow chromaticity change of the image of the person to be detected is obtained according to the analysis of the first comparison data and the second comparison data, and a detection result is output based on the evaluation value.

2. The jaundice detection method based on visual recognition according to claim 1, characterized in that: The step of acquiring first image data of the person to be detected under white light and acquiring second image data of the person to be detected under blue light includes: The face of the person to be detected is photographed under white light to obtain iris image data of the person to be detected, which is marked as first image data; The hands of the person to be detected are photographed under the blue light, and the palm image information of the person to be detected is obtained and marked as the second image data.

3. The jaundice detection method based on visual recognition according to claim 1, characterized in that: The performing image color extraction on the first image data based on the HSV color space to obtain color data of the first image data comprises: Input the first image data into the HSV color space for image segmentation, extract the target area in the first image data, and obtain a target area image; Perform color recognition on the target area image to obtain color data of the first image data.

4. The jaundice detection method based on visual recognition according to claim 3, characterized in that: The performing color recognition on the target area image to obtain color data of the first image data comprises: Set fixed value parameters and single variables in HSV space, adjust the value of single variable with preset intervals, and generate color recognition frame; The target area image is subjected to color extraction based on the color recognition frame to obtain the color chromaticity of the target area image.

5. The jaundice detection method based on visual recognition according to claim 4, characterized in that: The step of setting a fixed value parameter and a single variable in the HSV space, adjusting the value of the single variable with a preset interval, and generating a color recognition frame includes: In the HSV space, set the S value and V value as fixed values, set the H value as a variable, and generate an H value color identification box; The H value of the target area image is obtained based on the H value color recognition frame.

6. The jaundice detection method based on visual recognition according to claim 1, characterized in that: The step of comparing the color data of the first image data with the image chromaticity data of the standard image to obtain the first comparison data comprises: Convert the acquired color data of the first image data from the HSV format to the RGB format; A standard image corresponding to the detection area is extracted from the database, and the color data of the first image data is compared with the color data of the standard image to obtain first comparison data.

7. The jaundice detection method based on visual recognition according to claim 1, characterized in that: The extracting of the proportions of the green component and the yellow component from the second image data based on the Lab color space and obtaining the second comparison data based on the shooting parameter settings include: The second image data is input into a Lab color space model to extract color components, and the second image data is adjusted based on the Lab color space to extract color component data from the second image data.

8. The jaundice detection method based on visual recognition according to claim 7, characterized in that: The extracting the proportions of the green component and the yellow component from the second image data based on the Lab color space identification and obtaining the second comparison data based on the shooting parameter settings also includes: Extracting the original color component of the hand position of the second image data of the person to be detected according to the hand image data of the person to be detected recorded in the database; According to the blue light illumination parameter setting of the second image data, the original color component of the shooting environment of the second image data is obtained.

9. The jaundice detection method based on visual recognition according to claim 1, characterized in that: The step of obtaining an evaluation value of the yellow chromaticity change of the image of the person to be detected according to the first comparison data and the second comparison data, and outputting the detection result based on the evaluation value comprises: The calculation formula of the chromaticity change evaluation value is: Among them, θ is the yellow chromaticity change evaluation value, μ1 is the weight of the first comparison data, μ2 is the weight of the second comparison data, d1 is the first comparison data, d2 is the second comparison data, and γ0 is the initial yellow chromaticity value of the person to be detected.

10. A jaundice detection system based on visual recognition, characterized in that: The detection system comprises: Image acquisition component: used to acquire first image data of the person to be detected under white light shooting illumination, and to acquire second image data of the person to be detected under blue light shooting illumination; HSV model processing component: used for performing image color extraction on the first image data based on the HSV color space to obtain color data of the first image data; A first comparison component: used for comparing the color data of the first image data with the image chromaticity data of the standard image to obtain first comparison data; Lab model processing component: used to identify the second image data based on the Lab color space to extract the proportion of the green component and the yellow component, and obtain the second comparison data based on the shooting parameter settings; Evaluation component: used to obtain the yellow chromaticity change evaluation value of the image of the person to be detected based on the first comparison data and the second comparison data, and output the detection result based on the evaluation value.

Images