Skin detection method, device, equipment and product based on long-wave infrared image

Abstract

The invention discloses a skin detection method, device, equipment and product based on a long-wave infrared image, and relates to the technical field of skin detection.The method comprises the steps that the long-wave infrared image of a skin area to be detected is acquired; converting the long-wave infrared image into an HIS space and extracting an I-channel image; on the basis of the long-wave infrared image and the I-channel image, determining a feature image of the to-be-detected area of the skin so as to determine a temperature diffusion distance between adjacent isotherms; measuring the temperature diffusion distance difference of adjacent isotherms by using a Hausdorff distance, and determining a skin lesion range; calculating a gray value centroid point in an isotherm contour point interval of the skin lesion range region to construct a skin temperature centroid diffusion model; and calculating a skin lesion diffusion distance based on a skin temperature centroid diffusion model to complete skin lesion area detection. According to the invention, the diseased tissue and the diseased development trend of the skin can be quickly judged, and the skin detection accuracy is effectively improved.

Description

Technical Field

[0001] The present application relates to the technical field of skin detection, and in particular to a skin detection method, device, equipment and product based on long-wave infrared images. Background Art

[0002] Long-wave infrared imaging technology uses the long-wave infrared radiation emitted by objects in the wavelength range of 8 to 14 microns to detect, image and analyze objects. In the field of skin detection, since the changes in the long-wave infrared radiation of the skin are directly related to the characteristics of the skin's physiology, location, and lesions, the use of long-wave infrared imaging technology can obtain information that cannot be detected by conventional intensity imaging technology, such as the temperature distribution of the skin surface, differences in tissue structure, and changes in blood perfusion, making it gradually become an emerging technical means and method in the field of skin medical detection. Factors such as the distribution of skin tissue at different locations, physiological changes, and hardened lesions will produce temperature-related long-wave infrared radiation changes, causing the long-wave infrared image at that location to change or become deformed. Therefore, there is a strong correlation and sensitivity between the long-wave infrared image of the skin and skin damage or deformity. Therefore, by performing long-wave infrared imaging on the differences in the long-wave infrared radiation of the skin, and then reflecting the subtle changes in skin temperature, the visual display and detection of skin tissue distribution, physiological state differences, and pathological development trends can be effectively achieved. In addition, long-wave infrared imaging technology has the advantages of non-contact, high sensitivity and real-time imaging, making it a safe, reliable, simple, effective and cost-effective tool in applications such as early screening of skin lesions, dynamic monitoring, inflammation and infection detection, and blood circulation assessment. It is of great significance for early auxiliary diagnosis, health assessment and precise treatment of skin lesions.

[0003] However, in practical applications, such as the detection of early lesions of systemic sclerosis of the skin, the temperature of skin lesion tissue is less different from that of adjacent healthy tissue, resulting in unclear temperature characteristics of long-wave infrared images of skin lesion areas. This makes it impossible to accurately determine the actual lesion area and lesion development trend of the skin, thus affecting the accuracy of skin lesion detection. Therefore, the single long-wave infrared intensity imaging method cannot effectively achieve early diagnosis of skin lesions.

[0004] Among the existing related technologies, there is a detection method for detecting skin based on blood perfusion coefficient, "Pennesbioheat equation with heterogeneous blood perfusion: A newer perspective.". This method proposes that the skin's heterogeneous blood perfusion coefficient reflects the skin characteristics based on the skin heat conduction equation, and determines whether the skin at that location has lesions by measuring the blood perfusion rate coefficient per unit skin area. Doppler ultrasound equipment is used to evaluate the blood perfusion heterogeneity of skin tissue and measure the location of the heat exchange temperature change of biological tissue, so as to detect irregular tumors and lesions in skin tissue. However, when there is a blood circulation disorder in the skin tissue or inflammation causes no blood flow, the "material properties" of the skin will change, causing errors in the Doppler ultrasound signal, affecting the determination of the skin blood perfusion coefficient, resulting in the method being unable to accurately evaluate the lesion development trend of the skin area with low blood perfusion or no blood perfusion, resulting in the method being unable to effectively detect the skin lesions in the skin area, and the improvement of the detection of skin lesions is not comprehensive enough. At the same time, the detection equipment of this method is relatively expensive, its mechanical structure is relatively complex, and the operating cost requirements and detection costs are high, which has certain limitations in actual skin detection applications.

[0005] In summary, how to use relatively simple and effective skin detection methods and devices to detect skin lesions while ensuring the accuracy of detection is still a problem that needs to be solved urgently. Summary of the invention

[0006] The purpose of this application is to provide a skin detection method, device, equipment and product based on long-wave infrared images, which can quickly determine the diseased tissue and development trend of the skin, and effectively improve the accuracy of skin detection.

[0007] To achieve the above objectives, this application provides the following solutions:

[0008] In a first aspect, the present application provides a skin detection method based on long-wave infrared images, comprising:

[0009] Acquire a long-wave infrared image of the skin area to be detected;

[0010] Convert the long-wave infrared image from RGB space to HIS space and extract the I channel image;

[0011] Determine the characteristic image of the skin area to be detected based on the long-wave infrared image and the I channel image;

[0012] Determine the temperature diffusion distance of adjacent isotherms based on the characteristic image of the skin area to be detected;

[0013] The Hausdorff distance is used to measure the difference in temperature diffusion distances between adjacent isotherms to determine the extent of skin lesions.

[0014] Calculate the gray value centroid point of the isothermal contour point interval in the skin lesion area, and build a skin temperature centroid diffusion model in combination with the Bayesian probability density function;

[0015] The diffusion distance of skin lesions is calculated based on the skin temperature centroid diffusion model to complete the skin lesion area detection.

[0016] Optionally, converting the long-wave infrared image from the RGB space to the HIS space and extracting the I channel image specifically includes:

[0017] Converting the long-wave infrared image of the skin area to be detected into a gray value space and preprocessing it to obtain a preprocessed gray image; the preprocessing includes filtering, denoising, contrast enhancement and correction processing;

[0018] Normalizing the preprocessed grayscale image to obtain a normalized image;

[0019] The gray value data of the normalized image is extracted, and the normalized image is converted into the RGB space and the HIS space in sequence according to the gray value data, and the I channel data of the I channel image in the HIS space is extracted.

[0020] Optionally, determining the characteristic image of the skin area to be detected based on the long-wave infrared image and the I channel image specifically includes:

[0021] Performing image fusion on the long-wave infrared image and the I channel image to obtain a fused image;

[0022] Calculate and count the statistical histogram data of the fused image;

[0023] Based on the statistical histogram data of the fused image, the peak_index formula is used e =max{e|H(e)>H(e-1)andH(e)>H(e+1),1≤e<Q} Determine the maximum value position peak_index in the statistical histogram data e ; Through the formula start_index g =max{g∣H(g)<H(g-1)andH(g)<H(g+1),1≤g<peak_index e}Determine the start_index of the first minimum value in the interval to the left of the maximum value in the statistical histogram data g ; Through the formula descent_index o=min{o|H(o)<H(o-1)andH(o)<H(o+1),peak_index e <o<Q} Determine the first minimum position of the interval to the right of the maximum value in the statistical histogram data descent_index o ; Wherein, e is the maximum value position index in the statistical histogram data; g is the first minimum value position index in the interval to the left of the maximum value in the statistical histogram data; o is the first minimum value position index in the interval to the right of the maximum value in the statistical histogram data; H(·) is the statistical histogram data corresponding to each location in the statistical histogram of the fused image; Q is the maximum grayscale value in the statistical histogram data of the fused image;

[0024] According to start_index g and descent_index o Determine the grayscale value interval of the fused image, assign the values ​​outside the grayscale value interval to 0, and perform binarization to obtain a binarized image;

[0025] The binary image is fused with the grayscale data of the I channel image by dot multiplication to obtain the characteristic image of the skin area to be detected.

[0026] Optionally, determining the temperature diffusion distance of adjacent isotherms based on the characteristic image of the skin area to be detected specifically includes:

[0027] Based on the characteristic image of the skin area to be detected, the formula is used Determine the temperature diffusion distance d between adjacent isotherms inter (A,B); sup represents the supremum; inf represents the infimum; a∈A={contour_levels k,1 ,contour_levels k,2 ,...,contour_levels k,n}; b∈B={contour_levels k+1,1 ,contour_levels k+1 ,2,...,contour_levels k+1,m}; contour_levels k,n is the nth contour point on the kth isotherm, contour_levels k+1,m is the mth contour point on the k+1th isotherm.

[0028] Optionally, the method of measuring the difference in temperature diffusion distances of adjacent isotherms using the Hausdorff distance to determine the range of skin lesions specifically includes:

[0029] The Hausdorff distance is used to measure the temperature diffusion distance of adjacent isotherms between healthy tissue and diseased tissue, analyze the difference in temperature diffusion distance between adjacent isotherms between healthy tissue and diseased tissue, and set the distance judgment threshold T dis ; When the temperature diffusion distance d of adjacent isotherms inter (A, B) is greater than or equal to the distance judgment threshold T dis When the temperature diffusion distance d between adjacent isotherms is inter (A,B) is less than the distance judgment threshold T dis When it is skin lesion tissue, the scope of skin lesions can be determined.

[0030] Optionally, the step of calculating the gray value centroid point of the isothermal contour point interval in the skin lesion range area and constructing a skin temperature centroid diffusion model in combination with a Bayesian probability density function specifically includes:

[0031] By formula and The gray value centroid point (x centroid(k,i) ,y centroid(k,i) ), where M is the range of contour points; (x t ,y t ) is the position coordinate of the kth isotherm contour with the tth pixel point as the center within the contour point interval range M; I segment(k,i) is the gray value index of the ith pixel point in the kth isotherm contour interval where the skin lesion tissue is located in the skin lesion range area in the gray data of the characteristic image of the skin area to be detected;

[0032] Based on the gray value centroid point (x centroid(k,i) ,y centroid(k,i) ), the Bayesian probability density function is used to determine the maximum possible diffusion direction of the temperature centroid as the temperature diffusion direction vector The temperature diffusion direction vector expression is used as the skin temperature centroid diffusion model; where (x contour_levels(k,i) ,y contour_levels(k,i) ) is the position coordinate of the contour point on the isotherm where the skin lesion tissue is located within the skin lesion range.

[0033] Optionally, calculating the skin lesion diffusion distance based on the skin temperature centroid diffusion model to complete the skin lesion area detection specifically includes:

[0034] Constructing the linear equation of skin lesion diffusion direction based on skin temperature centroid diffusion model Among them, d x Represents the temperature diffusion direction vector The component in the x direction; d y Represents the temperature diffusion direction vector The component in the y direction; (x, y) is the position coordinates of the point on the equation of the straight line in the direction of skin lesion diffusion;

[0035] Based on the linear equation and distance measurement formula of the diffusion direction of skin lesions Calculate the diffusion distance D of skin lesions i,j ; Among them, (x cross(k+1,j) ,y cross(k+1,j) ) is the coordinate of the intersection of the linear equation of the skin lesion diffusion direction and the isotherm of the skin lesion tissue adjacent to the skin lesion area.

[0036] In a second aspect, the present application provides a skin detection device based on long-wave infrared images, comprising: a constant temperature loading platform, a hardware controller, a long-wave infrared camera, a first stepper motor, a second stepper motor, and a computer;

[0037] The computer is electrically connected to the constant temperature loading platform, the hardware controller and the long-wave infrared camera through control lines; the computer is used to control the temperature of the constant temperature loading platform; the hardware controller is electrically connected to the first stepper motor and the second stepper motor through control lines; the first stepper motor and the second stepper motor are orthogonally distributed around the constant temperature loading platform, and are connected to the constant temperature loading platform through threaded rods; the computer controls the operation of the first stepper motor and the second stepper motor through the hardware controller, thereby controlling the movement of the constant temperature loading platform to ensure that the skin area to be detected is in the field of view of the long-wave infrared camera; the long-wave infrared camera is supported by a bracket directly above the constant temperature loading platform; the long-wave infrared camera is used to perform long-wave infrared imaging of the skin area to be detected of the test sample placed on the constant temperature loading platform to obtain a long-wave infrared image; the computer performs skin detection based on the long-wave infrared image.

[0038] In a third aspect, the present application provides a computer device, comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the skin detection method based on long-wave infrared images.

[0039] In a fourth aspect, the present application provides a computer program product, including a computer program, which implements the skin detection method based on long-wave infrared images when executed by a processor.

[0040] According to the specific embodiments provided in this application, this application has the following technical effects:

[0041] The present application provides a skin detection method, device, equipment and product based on long-wave infrared images, which mainly relates to a non-contact lesion image detection method using skin long-wave infrared images for isotherm similarity analysis. In order to solve the problem that the existing related technologies cannot accurately judge and evaluate the skin lesion area and the lesion development trend, the present application uses the Hausdorff distance to measure the temperature diffusion distance difference of adjacent isotherms in the long-wave infrared skin image, calculates the gray value centroid point of the isotherm contour point interval in the lesion range area, and combines the Bayesian probability density function to analyze the maximum possible diffusion direction of the temperature centroid point, and constructs a skin temperature centroid diffusion model. The present application determines the skin lesion range and lesion development trend by measuring the skin lesion diffusion distance, establishes a skin detection method based on long-wave infrared imaging, detects the deterioration of the skin lesion area, and provides new detection methods and technical means for skin lesion prevention and medical intervention and treatment. Therefore, the skin detection method, device, equipment and product based on long-wave infrared images in the present application can quickly judge the skin lesion tissue and lesion development trend, and effectively improve the accuracy of skin detection. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] In order to more clearly illustrate the technical solutions in the embodiments of the present application or related technologies, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0043] Figure 1 Schematic diagram of infrared temperature image of early stage of systemic sclerosis of the skin provided in this application.

[0044] Figure 2 Flowchart of the skin detection method based on long-wave infrared images provided in this application.

[0045] Figure 3 Schematic diagram of the skin detection method based on long-wave infrared images provided in this application.

[0046] Figure 4 Schematic diagram of the long-wave infrared image provided for this application.

[0047] Figure 5 This is a schematic diagram of the I channel image provided for this application.

[0048] Figure 6 A schematic diagram of the statistical histogram data of the fused image provided in this application.

[0049] Figure 7 Schematic diagram of the characteristic image of the skin area to be detected provided in this application.

[0050] Figure 8Schematic diagram of the temperature diffusion distance between adjacent isotherms provided in this application.

[0051] Fig. 9 Schematic diagram of the range of skin lesions provided for this application.

[0052] Fig.10 for Fig. 9 Magnified image of the lesion spread distance at point 1 in the middle.

[0053] Fig.11 This is a skin lesion development trend judgment chart provided for this application.

[0054] Fig.12 This is a structural diagram of the skin detection device based on long-wave infrared images provided in this application. DETAILED DESCRIPTION

[0055] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0056] In order to make the above-mentioned objects, features and advantages of the present application more obvious and easy to understand, the present application is further described in detail below in conjunction with the accompanying drawings and specific implementation methods.

[0057] In practical applications, such as the detection of early lesions of systemic sclerosis of the skin, the schematic diagram of infrared temperature image of early lesions of systemic sclerosis of the skin is as follows: Figure 1 As shown, since the temperature difference between skin lesion tissue and adjacent healthy tissue is small, the temperature characteristics of the long-wave infrared image of the skin lesion area are not obvious, and early diagnosis of skin lesions cannot be effectively achieved.

[0058] Based on the above situation, the skin detection method based on long-wave infrared images provided by the present application mainly proposes to construct a skin temperature centroid diffusion model to calculate the diffusion distance of skin lesions for skin lesion judgment. The skin detection method based on long-wave infrared images of the present application further analyzes factors such as skin temperature information and adjacent isothermal temperature diffusion distance on the basis of long-wave infrared images, measures the temperature characteristics of skin lesion areas and healthy areas, and thus realizes effective diagnosis and detection of skin lesions. By calculating and analyzing the adjacent isothermal temperature diffusion distances in long-wave infrared images and combining the skin temperature diffusion distribution, a skin temperature centroid diffusion model is constructed to diagnose and detect skin lesions, realize early diagnosis of skin systemic sclerosis lesions and predict the development trend of lesion areas, and improve the accuracy of detection of lesions such as skin systemic sclerosis. The method is specifically implemented through three steps: first, after obtaining the long-wave infrared image of the skin area to be detected, extract the characteristic image of the skin area to be detected; second, calculate the adjacent isothermal temperature diffusion distance in the characteristic image of the skin area to be detected and judge whether a lesion occurs; third, calculate the skin lesion diffusion distance based on the skin lesion tissue area to determine the lesion development trend, and complete the skin detection based on the skin long-wave infrared image. The specific process of determining the pathological tissue and development trend by the skin detection method based on long-wave infrared images in this application is as follows:

[0059] like Figure 2 and Figure 3 As shown, the present application discloses a skin detection method based on long-wave infrared images, comprising:

[0060] For the first major step, firstly, the grayscale data of the acquired long-wave infrared image of the skin to be detected area is extracted and the long-wave infrared image is converted from RGB space to HIS space and the I channel image is extracted; then, the long-wave infrared image is filtered and preprocessed and its statistical histogram data is calculated; finally, the skin to be detected area is segmented according to the statistical histogram data of the long-wave infrared image combined with the grayscale data of the I channel image to obtain the characteristic image of the skin to be detected area, as shown in the following steps 1 to 3.

[0061] Step 1: Obtain a long-wave infrared image of the skin area to be detected.

[0062] Specifically, an infrared camera is used to photograph the skin area to be detected to obtain a long-wave infrared image of the skin area to be detected, referred to as a skin long-wave infrared image, such as Figure 4As shown, the pixel value of each pixel point of the long-wave infrared image represents the thermal radiation intensity of the position. After obtaining the long-wave infrared image of the skin area to be detected, the long-wave infrared image of the skin area to be detected is converted to the gray value space and preprocessed, including filtering, denoising, contrast enhancement and correction processing, etc., to obtain a preprocessed gray image, so as to improve the accuracy of subsequent conversion. In addition, the present application also normalizes the preprocessed gray image to obtain a normalized image, and the specific formula is:

[0063]

[0064] Where, I is the pixel value of the preprocessed image; I max and I min are the maximum and minimum pixel values ​​of the preprocessed image respectively; I' is the normalized pixel value.

[0065] Step 2: Convert the long-wave infrared image from RGB space to HIS space and extract the I channel image.

[0066] The gray value data of the normalized image is extracted, and the normalized image is converted to the RGB space and the HIS space in turn according to the gray value data, and the I channel data of the I channel image in the HIS space is extracted. Specifically, the normalized skin long-wave infrared image is converted from the RGB space (long-wave infrared RGB image) to the HIS space to obtain the hue H channel component H 分 , Luminance I channel component I 分 And the saturation S channel component S 分 , the specific conversion formulas for each channel component are as follows:

[0067] Calculate the I channel component:

[0068]

[0069] Among them, R, G, and B are the red, green, and blue components of the skin long-wave infrared RGB image, respectively.

[0070] Calculate the S channel component:

[0071]

[0072] Calculate the H channel component:

[0073]

[0074] When B≤G, H 分 =θ, when B>G, H 分 =360-θ. The I channel image in the skin long-wave infrared image can be extracted by formula (2), such as Figure 5 shown.

[0075] Step 3: Determine the characteristic image of the skin area to be detected based on the long-wave infrared image and the I channel image.

[0076] The long-wave infrared image is image-fused with the I-channel image to obtain a fused image; and statistical histogram data of the fused image is calculated and counted. Specifically, the skin long-wave infrared image is first used as target data to perform image fusion with the I-channel image, and the fused image is filtered to obtain a filtered image F. Then, the statistical histogram data H of the filtered image F is calculated as the statistical histogram data of the fused image, and a Gaussian convolution kernel is applied to the statistical histogram to obtain the statistical histogram data H of the filtered fused image.

[0077]

[0078] Wherein, H(L) is the value of the statistical histogram of the fused image at the Lth position; H(Lj) is the value of the statistical histogram of the fused image at the Ljth position; G[H(j)] represents the value of the Gaussian kernel at position j; represents the sum of the Gaussian kernel in the range [-t', t'], which is used to normalize the Gaussian kernel. represents the radius of the Gaussian kernel, and T is the value of the Gaussian kernel window interval.

[0079] According to the statistical histogram data after the fusion image is filtered, the corresponding maximum value position in the histogram and the minimum value position near the maximum gray value point are found, such as Figure 6 As shown, the pixel positions of the characteristic image of the skin area to be detected are segmented to obtain the characteristic image of the skin area to be detected.

[0080] The index formula for the maximum value position in the statistical histogram data is:

[0081] peak_index e =max{e∣H(e)>H(e-1)andH(e)>H(e+1),1≤e<Q} (6)

[0082] The index formula for the first minimum value position in the interval to the left of the maximum value in the statistical histogram data is:

[0083] start_index g =max{g∣H(g)<H(g-1)andH(g)<H(g+1),1≤g<peak_index e}(7)

[0084] The index formula for the first minimum value position in the right interval of the maximum value in the statistical histogram data is:

[0085] descent_index o =min{o|H(o)<H(o-1)andH(o)<H(o+1),peak_index e <o<Q}(8)

[0086] Among them, peak_index e The position of the maximum value in the statistical histogram data of the fused image; start_index g The first minimum position of the interval to the left of the maximum value in the statistical histogram data of the fused image; descent_index o is the first minimum position of the interval to the right of the maximum value in the statistical histogram data of the fused image; H(·) is the statistical histogram data corresponding to each location in the statistical histogram data of the fused image; e is the index of the maximum value position in the statistical histogram data; g is the index of the first minimum position in the interval to the left of the maximum value in the statistical histogram data; o is the index of the first minimum position in the interval to the right of the maximum value in the statistical histogram data; Q is the maximum grayscale value in the statistical histogram data of the fused image.

[0087] According to start_index g and descent_index o The grayscale value in the range of the histogram is used as the grayscale value interval of the grayscale data of the final skin detection area feature image, and the grayscale values ​​outside the interval are assigned to 0 to obtain the image G filter , then for G filter Binarization is performed to obtain image G bin , the image G bin Perform point multiplication fusion with the grayscale data of the I channel image to obtain the characteristic image I of the skin area to be detected segment ,like Figure 7 shown.

[0088] I segment =G bin .*I 分 (9)

[0089] For the second step, firstly, according to the skin feature area image I to be detected segment Divide the number of isotherms and draw the contour curve of the isotherms to obtain different contour points C on different isotherms k,n , C k,n represents the nth contour point on the kth isotherm; then measure C k+1,n With C k,nThe maximum deviation distance between two contour points on adjacent isotherms is used as the value of the temperature diffusion distance of adjacent isotherms. This method needs to find the minimum distance between two sets of adjacent isotherm contour points and take the maximum value. When skin lesions occur, the "material properties" of the skin tissue will change, resulting in the value of the temperature diffusion distance of adjacent isotherms being inconsistent with that of normal skin. Therefore, by measuring the temperature diffusion distance of adjacent isotherms and setting the distance judgment threshold T dis The diseased tissue is segmented. The second major step is to construct a skin lesion detection model to determine whether a lesion occurs by calculating the image data of the feature area, using the Hausdorff distance to determine the diseased tissue, calculating the diffusion distance of adjacent isotherms, and determining the lesion area based on the diffusion distance, as shown in the following steps 4 to 5.

[0090] Step 4: Determine the temperature diffusion distance of adjacent isotherms based on the characteristic image of the skin area to be detected.

[0091] First, set the number of contour lines of the characteristic image of the skin area to be detected, pixel_step, which is set to 5 here, and calculate the grayscale interval inter of adjacent isotherms based on this value:

[0092] inter=floor(max(I segment )-min(I segment >0)) / pixel_step (10)

[0093] Among them, floor is the rounding down function.

[0094] Calculate the number of isotherms based on the grayscale interval inter contour_levels inter :

[0095] contour_levels inter =min(I segment >0):inter:max(I segment ) (11)

[0096] Finally, the temperature diffusion distance d of adjacent isotherms is calculated inter (A,B), such as Figure 8 As shown:

[0097]

[0098] Where sup represents the upper bound of the distance between sets a and b; inf represents the lower bound of the distance between sets a and b; where set a∈A={contour_levels k,1 ,contour_levels k,2 ,...,contour_levelsk,n}, set b∈B={contour_levels k+1,1 ,contour_levels k+1,2 ,...,contour_levels k+1,m}; contour_levels k,n is the nth contour point on the kth isotherm, contour_levels k+1,m is the mth contour point on the k+1th isotherm, that is, A and B represent the contour point data on adjacent isotherms, a belongs to the contour point data on the isotherm of A, and b belongs to the contour point data on the isotherm of B.

[0099] Step 5: Use the Hausdorff distance to measure the difference in temperature diffusion distances between adjacent isotherms to determine the extent of skin lesions.

[0100] Specifically, the Hausdorff distance is used to measure the temperature diffusion distance of adjacent isotherms of healthy tissue and diseased tissue, analyze the difference in temperature diffusion distance of adjacent isotherms of healthy tissue and diseased tissue, and set the distance judgment threshold T dis :The diffusion distance of adjacent isotherms of diseased skin is measured by Hausdorff distance, and the diffusion distance of healthy skin is measured by the same method, where the diseased skin is the skin with lesions confirmed by pathological detection, so as to obtain the segmentation threshold of diseased skin and healthy skin, that is, the distance judgment threshold T dis When the temperature diffusion distance d of adjacent isotherms inter (A, B) is greater than or equal to the distance judgment threshold T dis When the temperature diffusion distance d of adjacent isotherms is inter (A,B) is less than the distance judgment threshold T dis When it is skin lesion tissue, the scope of skin lesions can be determined, such as Fig. 9 shown.

[0101] For the third step, the grayscale centroid points within the range of different isothermal contour points in the skin lesion area are first calculated, and the direction vector angle θ' between the isothermal contour point and the grayscale centroid point is calculated. Then, the maximum possible diffusion direction vector of the temperature centroid within the contour point range is calculated by combining the Bayesian probability density function formula. Finally, the diffusion direction straight line equation of the temperature centroid is constructed according to the direction vector. The intersection point with the adjacent isotherm contour is calculated by the diffusion direction straight line equation and the distance between the new contour point and the intersection point is calculated as the lesion diffusion distance. The lesion diffusion distance is compared with the distance judgment threshold T dis Compare and segment the location where the lesion exists in the region, and determine the development trend of the lesion. The specific process is shown in steps 6 and 7.

[0102] Step 6: Calculate the grayscale centroid point of the isothermal contour point interval in the skin lesion area, and construct a skin temperature centroid diffusion model in combination with the Bayesian probability density function.

[0103] First, calculate the gray value centroid point (x centroid(k,i) ,y centroid(k,i) ):

[0104]

[0105]

[0106] Where M is the range of contour points; (x t ,y t ) is the position coordinate of the kth isotherm contour with the tth pixel point as the center within the contour point interval range M; I segment(k,i) It is the gray value index of the ith pixel in the kth isotherm contour interval where the skin lesion tissue is located in the skin lesion range area in the gray data of the characteristic image of the skin area to be detected.

[0107] Secondly, calculate the temperature diffusion direction vector

[0108] According to the gray value centroid point (x centroid(k,i) ,y centroid(k,i) ) and the position coordinates (x contour_levels(k,i) ,y contour_levels(k,i) ) form the straight line direction vector:

[0109]

[0110] The Bayesian probability density function is then used to predict the maximum possible diffusion direction of each contour point as the temperature diffusion direction, and the temperature diffusion direction vector expression is used as the skin temperature centroid diffusion model. Under these conditions, the observed data (x i ,y i )’s expected position μ i It is expressed as:

[0111]

[0112] Among them, the observation data (x i ,y i ) is the contour point (x contour_levels(k,i) ,y contour_levels(k,i) ), f represents the transformation function, Indicates in the direction The observed data (x i ,yi )The desired location of the diffusion.

[0113] Assuming that the observed data follows a Gaussian distribution, in a given direction The likelihood of the observed data is expressed as:

[0114]

[0115] Among them, σ i Represents the standard deviation of the diffusion position of the observed data at the i-th pixel; Indicates that in a given direction The diffusion probability of the observed data position is i∈[1,2,…,N], where N represents the total number of contour points on the kth isotherm.

[0116] Assuming the direction vector The prior probability of is a Gaussian distribution with zero mean:

[0117]

[0118] in, Indicates the direction vector The probability density at σ d is the standard deviation of the Gaussian distribution, indicating that the vector The fluctuation range in different diffusion directions.

[0119] The marginal probability of the observed data is the vector of all possible directions The corresponding weighted sum of observations:

[0120]

[0121] Among them, P(data) represents all possible direction vectors given The total probability of . Indicates that the vector in a given direction The probability of occurrence under the condition.

[0122] Find the most likely diffusion direction by maximizing the posterior probability As the temperature diffusion direction:

[0123] Substituting the likelihood function and prior probability into the Bayesian probability density function formula is:

[0124]

[0125] in, It means to find the conditional probability Next, the maximum direction vector Then determine the maximum posterior probability (x i ,yi ) parameter, and substituting it into formula (15) can obtain the skin temperature diffusion direction vector at that location, thus completing the construction of the skin temperature centroid diffusion model.

[0126] Step 7: Calculate the skin lesion diffusion distance based on the skin temperature centroid diffusion model to complete the skin lesion area detection.

[0127] Diffusion direction vector according to temperature With the contour point (x contour_levels(k,i) ,y contour_levels(k,i) )Construct the linear equation of the diffusion direction of skin lesions:

[0128]

[0129] Among them, d x Represents the temperature diffusion direction vector The component in the x direction; d y Represents the temperature diffusion direction vector The component in the y direction; (x, y) is the position coordinates of the point on the equation of the straight line in the direction of diffusion of the skin lesion.

[0130] Contour_levels curves formed by adjacent contours i+1 The coordinates of the intersection of the linear equation of the skin lesion diffusion direction and the isothermal line adjacent to the skin lesion tissue in the skin lesion range (x cross(k+1,j) ,y cross(k+1,j) ), and obtain the diffusion distance D of skin lesions i,j :

[0131]

[0132] Calculate the diffusion distance D of skin lesions i,j ,like Fig.10 As shown, using the method in step 5, the skin lesion is spread a distance D i,j and the distance judgment threshold T dis For comparison, when the skin lesion spreads a distance D i,j Greater than or equal to the distance judgment threshold T dis When it is healthy tissue, when the skin lesion spreads a distance D i,j Less than the distance judgment threshold T dis When it is a diseased tissue, combined with the diseased tissue determined in step 5, according to the position change of the diseased tissue from step 5 to step 7, the diseased tissue development trend is determined, such as Fig.11 As shown, skin detection of skin long-wave infrared images is completed.

[0133] In summary, the skin detection method based on long-wave infrared images of the present application can quickly determine the pathological tissue and development trend of the lesions on the skin, and effectively improve the accuracy of skin detection. In addition, the skin detection method based on long-wave infrared images provided by the present application can be used not only to detect the trend of skin lesions on mice or other animals, but also to detect the lesion area and development trend of human skin.

[0134] like Fig.12 As shown, the present application also provides a skin detection device based on long-wave infrared images, including: a constant temperature loading platform, a hardware controller, a long-wave infrared camera, a first stepper motor, a second stepper motor and a computer. The computer is electrically connected to the constant temperature loading platform, the hardware controller and the long-wave infrared camera through control lines, respectively, for linkage control of each part. The computer is used to control the temperature of the constant temperature loading platform. The hardware controller is electrically connected to the first stepper motor and the second stepper motor through control lines. The first stepper motor and the second stepper motor are orthogonally distributed and placed around the constant temperature loading platform, and are connected to the constant temperature loading platform through threaded rods. The computer controls the operation of the first stepper motor and the second stepper motor through the hardware controller, thereby controlling the movement of the constant temperature loading platform to ensure that the skin to be detected area is in the field of view of the long-wave infrared camera. The long-wave infrared camera is supported by a bracket directly above the constant temperature loading platform, and is used to perform long-wave infrared imaging of the skin to be detected area of ​​the test sample placed on the constant temperature loading platform to obtain a long-wave infrared image. The computer performs skin detection based on the long-wave infrared image.

[0135] When the skin lesion detection device based on long-wave infrared images is used for detection, the skin area to be detected is placed on a constant-temperature loading platform. The computer controls the first stepper motor and the second stepper motor through the hardware controller to move the constant-temperature loading platform to ensure that the skin area to be detected is in the field of view of the long-wave infrared camera. According to the relative position of the constant-temperature loading platform, the focal length of the long-wave infrared camera is adjusted so that a clear long-wave infrared image of the skin can be obtained. Then, the computer adjusts and controls the temperature of the constant-temperature loading platform according to the temperature data of the long-wave infrared image of the skin to be detected and keeps it stable, so as to remove the interference of the background temperature to be measured on the long-wave infrared image of the skin area to be measured. After the temperature of the constant-temperature loading platform is stable, the long-wave infrared image of the skin area to be measured is taken by the long-wave infrared camera. Subsequently, the image is analyzed for skin lesions by the computer using a skin detection method based on long-wave infrared images, and finally scleroderma lesion detection is achieved.

[0136] This application uses the diffusion distance of adjacent isotherms of the skin in the long-wave infrared image to construct a skin temperature centroid diffusion model, calculate the diffusion distance of skin lesions, and detect the development trend of lesions. Using a simple long-wave infrared imaging skin detection method and equipment to perform long-wave infrared imaging and processing on the skin area to be detected, effective and rapid imaging detection of skin lesions can be performed on skin areas including low-temperature skin areas with unclear temperature characteristics, low blood perfusion or no blood perfusion skin areas, achieving simple detection equipment, low detection cost and wide application range of skin lesion range judgment and development trend detection.

[0137] In some embodiments, the present application also provides a computer program product, including a computer program, which implements the skin detection method based on long-wave infrared images when executed by a processor.

[0138] In some embodiments, the present application also provides a computer-readable storage medium having a computer program stored thereon, and when the computer program is executed by a processor, the skin detection method based on long-wave infrared images is implemented.

[0139] In some embodiments, the present application also provides a computer device, including a processor, a memory, an input / output interface (Input / Output, referred to as I / O), a communication interface, and a computer program stored in the memory and executable on the processor, and the processor executes the computer program to implement the skin detection method based on long-wave infrared images.

[0140] The processor, memory and input / output interface are connected through a system bus, and the communication interface is connected to the system bus through the input / output interface. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The database of the computer device is used to store pending transactions. The input / output interface of the computer device is used to exchange information between the processor and an external device. The communication interface of the computer device is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, the skin detection method based on long-wave infrared images is implemented.

[0141] It should be noted that the object information (including but not limited to object device information, object personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the object or fully authorized by all parties, and the collection, use and processing of relevant data need to comply with relevant laws, regulations and standards of relevant countries and regions.

[0142] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to the memory, database or other medium used in the embodiments provided in the present application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. As an illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM). The database involved in each embodiment provided in this application may include at least one of a relational database and a non-relational database. Non-relational databases may include distributed databases based on blockchains, etc., but are not limited thereto. The processor involved in each embodiment provided in this application may be a general-purpose processor, a central processing unit, a graphics processor, a digital signal processor, a programmable logic device, a data processing logic device based on quantum computing, etc., but are not limited thereto.

[0143] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0144] This article uses specific examples to illustrate the principles and implementation methods of this application. The description of the above embodiments is only used to help understand the method and core ideas of this application. At the same time, for those skilled in the art, according to the ideas of this application, 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 this application.

Claims

1. A skin detection method based on long-wave infrared images, characterized in that: include: Acquire a long-wave infrared image of the skin area to be detected; Convert the long-wave infrared image from RGB space to HIS space and extract the I channel image; Determine the characteristic image of the skin area to be detected based on the long-wave infrared image and the I channel image; Determine the temperature diffusion distance of adjacent isotherms based on the characteristic image of the skin area to be detected; The Hausdorff distance is used to measure the difference in temperature diffusion distances between adjacent isotherms to determine the extent of skin lesions. Calculate the gray value centroid point of the isothermal contour point interval in the skin lesion area, and build a skin temperature centroid diffusion model in combination with the Bayesian probability density function; The diffusion distance of skin lesions is calculated based on the skin temperature centroid diffusion model to complete the skin lesion area detection.

2. The skin detection method based on long-wave infrared images according to claim 1, characterized in that: The converting of the long-wave infrared image from the RGB space to the HIS space and extracting the I channel image specifically includes: Converting the long-wave infrared image of the skin area to be detected into a gray value space and preprocessing it to obtain a preprocessed gray image; the preprocessing includes filtering, denoising, contrast enhancement and correction processing; Normalizing the preprocessed grayscale image to obtain a normalized image; The gray value data of the normalized image is extracted, and the normalized image is converted into the RGB space and the HIS space in sequence according to the gray value data, and the I channel data of the I channel image in the HIS space is extracted.

3. The skin detection method based on long-wave infrared images according to claim 2, characterized in that: The step of determining the characteristic image of the skin area to be detected based on the long-wave infrared image and the I channel image specifically includes: Performing image fusion on the long-wave infrared image and the I channel image to obtain a fused image; Calculate and count the statistical histogram data of the fused image; Based on the statistical histogram data of the fused image, the peak_index formula is used e =max{e|H(e)>H(e-1)andH(e)>H(e+1),1≤e<Q} Determine the maximum value position peak_index in the statistical histogram data e ; Through the formula start_index g =max{g∣H(g)<H(g-1)andH(g)<H(g+1),1≤g<peak_index e }Determine the start_index of the first minimum value in the interval to the left of the maximum value in the statistical histogram data g ; Through the formula descent_index o =min{o|H(o)<H(o-1)andH(o)<H(o+1),peak_index e <o<Q} Determine the first minimum position descent_index of the interval to the right of the maximum value in the statistical histogram data o ; Wherein, e is the maximum value position index in the statistical histogram data; g is the first minimum value position index in the interval to the left of the maximum value in the statistical histogram data; o is the first minimum value position index in the interval to the right of the maximum value in the statistical histogram data; H(·) is the statistical histogram data corresponding to each location in the statistical histogram of the fused image; Q is the maximum grayscale value in the statistical histogram data of the fused image; According to start_index g and descent_index o Determine the grayscale value interval of the fused image, assign the values ​​outside the grayscale value interval to 0, and perform binarization to obtain a binarized image; The binary image is fused with the grayscale data of the I channel image by dot multiplication to obtain the characteristic image of the skin area to be detected.

4. The skin detection method based on long-wave infrared images according to claim 3 is characterized in that: The step of determining the temperature diffusion distance of adjacent isotherms based on the characteristic image of the skin area to be detected specifically includes: Based on the characteristic image of the skin area to be detected, the formula is used Determine the temperature diffusion distance d between adjacent isotherms inter (A,B); sup represents the supremum; inf represents the infimum; a∈A={contour_levels k,1 ,contour_levels k,2 ,...,contour_levels k,n }; b∈B={contour_levels k+1,1 ,contour_levels k+1,2 ,...,contour_levels k+1,m }; contour_levels k,n is the nth contour point on the kth isotherm, contour_levels k+1,m is the mth contour point on the k+1th isotherm.

5. The skin detection method based on long-wave infrared images according to claim 4, characterized in that: The method of using the Hausdorff distance to measure the difference in temperature diffusion distances of adjacent isotherms to determine the range of skin lesions specifically includes: The Hausdorff distance is used to measure the temperature diffusion distance of adjacent isotherms between healthy tissue and diseased tissue, analyze the difference in temperature diffusion distance between adjacent isotherms between healthy tissue and diseased tissue, and set the distance judgment threshold T dis ; When the temperature diffusion distance d of adjacent isotherms inter (A, B) is greater than or equal to the distance judgment threshold T dis When the temperature diffusion distance d of adjacent isotherms is inter (A, B) is less than the distance judgment threshold T dis When it is skin lesion tissue, the scope of skin lesions can be determined.

6. The skin detection method based on long-wave infrared images according to claim 5, characterized in that: The method of calculating the gray value centroid point of the isothermal contour point interval in the skin lesion range area and constructing a skin temperature centroid diffusion model in combination with the Bayesian probability density function specifically includes: By formula and The gray value centroid point (x centroid(k,i) ,y centroid(k,i) ), where M is the range of contour points; (x t ,y t ) is the position coordinate of the kth isotherm contour with the tth pixel point as the center within the contour point interval range M; I segment(k,i) is the gray value index of the ith pixel point in the kth isotherm contour interval where the skin lesion tissue is located in the skin lesion range area in the gray data of the characteristic image of the skin area to be detected; Based on the gray value centroid point (x centroid(k,i) ,y centroid(k,i) ), the Bayesian probability density function is used to determine the maximum possible diffusion direction of the temperature centroid as the temperature diffusion direction vector The temperature diffusion direction vector expression is used as the skin temperature centroid diffusion model; where (x contour_levels(k,i) ,y contour_levels(k,i) ) is the position coordinate of the contour point on the isotherm where the skin lesion tissue is located within the skin lesion range.

7. The skin detection method based on long-wave infrared images according to claim 5, characterized in that: The method of calculating the diffusion distance of skin lesions based on the skin temperature centroid diffusion model and completing the skin lesion area detection specifically includes: Constructing the linear equation of skin lesion diffusion direction based on skin temperature centroid diffusion model Among them, d x Represents the temperature diffusion direction vector The component in the x direction; d y Represents the temperature diffusion direction vector The component in the y direction; (x, y) is the position coordinates of the point on the equation of the straight line in the direction of skin lesion diffusion; Based on the linear equation and distance measurement formula of the diffusion direction of skin lesions Calculate the diffusion distance D of skin lesions i,j ; Among them, (x cross(k+1,j) ,y cross(k+1,j) ) is the coordinate of the intersection of the linear equation of the skin lesion diffusion direction and the isotherm of the skin lesion tissue adjacent to the skin lesion area.

8. A skin detection device based on long-wave infrared images, characterized in that: include: Constant temperature loading platform, hardware controller, long wave infrared camera, first stepper motor, second stepper motor and computer; The computer is electrically connected to the constant temperature loading platform, the hardware controller and the long-wave infrared camera through control lines; the computer is used to control the temperature of the constant temperature loading platform; the hardware controller is electrically connected to the first stepper motor and the second stepper motor through control lines; the first stepper motor and the second stepper motor are orthogonally distributed around the constant temperature loading platform and connected to the constant temperature loading platform through threaded rods; The computer controls the operation of the first stepper motor and the second stepper motor through the hardware controller, thereby controlling the movement of the constant temperature loading platform to ensure that the skin area to be detected is in the field of view of the long-wave infrared camera; The long-wave infrared camera is supported by a bracket directly above the constant temperature loading platform; The long-wave infrared camera is used to perform long-wave infrared imaging on the skin to-be-tested area of ​​the test sample placed on the constant-temperature loading platform to obtain a long-wave infrared image; and the computer performs skin detection based on the long-wave infrared image.

9. A computer device comprising: A memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the skin detection method based on long-wave infrared images as described in any one of claims 1 to 7.

10. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the skin detection method based on long-wave infrared images described in any one of claims 1 to 7 is implemented.

Images