Foot detection method, system and equipment based on image recognition

Abstract

The invention belongs to the technical field of foot detection, and particularly relates to a foot detection method, system and equipment based on image recognition. According to the method, the extraction precision of the foot shape contour is remarkably improved, the accuracy of a detection result is ensured, the foot appearance information and the footprint area coefficient are combined, detection is carried out from the two aspects of appearance and functional characteristics, all-around foot evaluation is provided, the generated foot detection report can visually reflect the health condition and the functional characteristics of the foot, and the detection accuracy is improved. The method provides scientific basis for foot nursing, correction insole design and the like, is suitable for foot image data of different forms, has high universality and adaptability, and can be popularized to the fields of medical treatment, sports health and the like.

Description

Technical Field

[0001] The present invention belongs to the technical field of foot detection, and particularly relates to a foot detection method, system and device based on image recognition. Background Art

[0002] As an important part of the human body, the foot not only bears the weight of the body, but also plays a key role in movements such as walking and running. Foot health is directly related to the overall health of the human body. Foot diseases such as flat feet, high arches, plantar fasciitis, etc. may have an adverse impact on human posture, gait and motor ability. Therefore, the detection and evaluation of foot morphology and function are of great significance.

[0003] In recent years, with the development of computer vision and image recognition technologies, image-based foot detection methods have gradually become a research hotspot. Such methods can efficiently and accurately extract foot morphology information and footprint features by collecting foot image data through cameras or scanning devices and combining image processing and analysis technologies. However, there are still deficiencies in the existing technologies in practical applications. For example, during the image acquisition process, affected by factors such as optical devices, shooting angles and environmental light, the foot images may be distorted, resulting in a decrease in the accuracy of the detection results. Only focusing on the appearance morphology or footprint information of the foot fails to achieve a comprehensive evaluation of the foot morphology and function, and shows deficiencies in the processing of foot image data with different morphologies, making it difficult to adapt to diverse foot morphologies and health states. Summary of the Invention

[0004] The purpose of the present invention is to provide a foot detection method, system and device based on image recognition, which can obtain foot morphology image data, extract foot contour information and perform distortion correction, and generate a comprehensive foot detection report by combining the comprehensive analysis of appearance features and footprint features.

[0005] The specific technical solutions adopted by the present invention are as follows:

[0006] A foot detection method based on image recognition, comprising:

[0007] Obtaining foot morphology image data, and obtaining foot morphology contour data according to the foot morphology image data;

[0008] Obtaining the initial pixel coordinate information of the contour corresponding to the foot morphology contour data;

[0009] Obtaining image distortion compensation data, and obtaining the corrected pixel coordinate information of the contour according to the image distortion compensation data and the initial pixel coordinate information of the contour;

[0010] Obtaining foot appearance information according to the corrected pixel coordinate information of the contour, and obtaining foot appearance detection information according to the foot appearance information;

[0011] Obtaining a footprint area coefficient according to the contour-corrected pixel coordinate information, and obtaining footprint detection information according to the footprint area coefficient;

[0012] Obtain a foot inspection report based on the footprint appearance inspection information and the footprint inspection information.

[0013] In a preferred embodiment, the step of obtaining foot morphology image data and obtaining foot morphology contour data according to the foot morphology image data includes:

[0014] Acquiring foot morphology image data;

[0015] Acquiring a corresponding foot morphology image according to the foot morphology image data;

[0016] Acquire a corresponding foot morphology grayscale image according to the foot morphology image;

[0017] The corresponding foot shape contour data is obtained according to the foot shape grayscale image.

[0018] In a preferred embodiment, the step of obtaining initial pixel coordinate information of the contour corresponding to the foot shape contour data includes:

[0019] Acquire corresponding foot contour world coordinate data according to the foot shape contour data;

[0020] Acquire a plurality of corresponding foot shape contour world coordinate points according to the foot contour world coordinate data;

[0021] Obtaining the intrinsic parameter matrix and extrinsic parameter matrix of the image acquisition device of the foot morphology image data;

[0022] Acquire corresponding initial pixel coordinate points of the foot contour according to a plurality of world coordinate points of the foot morphology contour, an internal parameter matrix and an external parameter matrix;

[0023] A plurality of initial pixel coordinate points of the foot contour are aggregated, and the aggregated result is marked as the initial pixel coordinate information of the contour.

[0024] In a preferred embodiment, the step of obtaining image distortion compensation data and obtaining contour correction pixel coordinate information according to the image distortion compensation data and contour initial pixel coordinate information includes:

[0025] Acquire image distortion compensation data of an image acquisition device for obtaining foot morphology image data;

[0026] Acquiring corresponding radial distortion parameters and tangential distortion parameters according to the image distortion compensation data;

[0027] Acquire a plurality of corresponding initial pixel coordinate points of the foot contour according to the initial pixel coordinate information of the contour;

[0028] Obtain the corresponding distance parameter of each initial pixel coordinate point of the foot contour from the origin of the pixel coordinate system;

[0029] Obtaining foot contour correction pixel coordinate points according to each foot contour initial pixel coordinate point, a corresponding distance parameter, a radial distortion parameter, and a tangential distortion parameter;

[0030] A plurality of foot contour correction pixel coordinate points are aggregated, and the aggregated result is marked as contour correction pixel coordinate information.

[0031] In a preferred embodiment, the steps of obtaining foot appearance information according to the contour-corrected pixel coordinate information and obtaining foot appearance detection information according to the foot appearance information include:

[0032] Acquire various foot appearance feature coordinates according to contour-corrected pixel coordinate information;

[0033] Obtaining corresponding foot appearance feature values ​​according to each foot appearance feature coordinate;

[0034] Obtaining a corresponding appearance feature table, wherein the appearance feature table includes a plurality of foot appearance feature interval values ​​and a foot appearance detection index corresponding to each foot appearance feature interval value;

[0035] Obtaining a target foot appearance feature interval value according to the foot appearance feature value;

[0036] Obtaining a corresponding foot appearance detection index from an appearance feature table according to a target foot appearance feature interval value;

[0037] A variety of foot appearance detection indicators are summarized, and the summary results are marked as foot appearance detection information.

[0038] In a preferred embodiment, the steps of obtaining a footprint area coefficient according to the contour-corrected pixel coordinate information and obtaining footprint detection information according to the footprint area coefficient include:

[0039] Acquire footprint area contour coordinate information according to the contour-corrected pixel coordinate information, wherein the footprint area contour coordinate information includes forefoot area contour coordinate information, midfoot area contour coordinate information, and hindfoot area contour coordinate information;

[0040] Obtaining the area contours of the forefoot, midfoot, and hindfoot according to the footprint contour coordinate information;

[0041] Obtain footprint area coefficients based on the contour areas of the forefoot, midfoot, and hindfoot regions;

[0042] Obtaining a corresponding footprint table, wherein the footprint table includes a plurality of footprint area interval coefficients and a footprint type corresponding to each footprint area interval coefficient;

[0043] Obtain the corresponding target footprint area interval coefficient according to the footprint area coefficient;

[0044] Obtain the corresponding footprint type from the footprint table according to the target footprint area interval coefficient, and mark it as footprint detection information.

[0045] In a preferred solution, the steps of obtaining the regional contour areas of the forefoot, midfoot, and hindfoot according to the footprint contour coordinate information include:

[0046] Obtain the regional contour inflection point coordinate information of the forefoot, midfoot, and hindfoot according to the footprint contour coordinate information;

[0047] Obtain the corresponding multiple regional contour inflection point coordinates according to the regional contour inflection point coordinate information;

[0048] Obtain the regional contour areas of the forefoot, midfoot, and hindfoot respectively according to the corresponding multiple regional contour inflection point coordinates.

[0049] The present invention also provides a foot detection system based on image recognition for the above-mentioned foot detection method based on image recognition, including:

[0050] A foot contour module, configured to obtain foot morphology image data and obtain foot morphology contour data according to the foot morphology image data;

[0051] An initial pixel module, configured to obtain the contour initial pixel coordinate information corresponding to the foot morphology contour data;

[0052] A corrected pixel module, configured to obtain image distortion compensation data and obtain contour corrected pixel coordinate information according to the image distortion compensation data and the contour initial pixel coordinate information;

[0053] An appearance detection module, configured to obtain foot appearance information according to the contour corrected pixel coordinate information and obtain foot appearance detection information according to the foot appearance information;

[0054] A footprint detection module, configured to obtain a footprint area coefficient according to the contour corrected pixel coordinate information and obtain footprint detection information according to the footprint area coefficient;

[0055] A foot detection module, configured to obtain a foot detection report according to the footprint appearance detection information and the footprint detection information.

[0056] The present invention also provides a foot detection device based on image recognition for the above-mentioned foot detection method based on image recognition, including:

[0057] A device main body;

[0058] A light-transmitting glass, which is arranged above the device main body and is used to carry the foot;

[0059] A rear camera, which is arranged above one side of the device body and is used for collecting rear image data of the foot;

[0060] A footprint camera, which is arranged inside the device body and under the light-transmitting glass and is used for collecting footprint image data.

[0061] Moreover, a foot detection terminal based on image recognition includes:

[0062] One or more processors;

[0063] A storage device on which one or more programs are stored;

[0064] When the one or more programs are executed by the one or more processors, the one or more processors implement a foot detection method based on image recognition.

[0065] The technical effects achieved by the present invention are as follows:

[0066] In the present invention, the extraction accuracy of the foot shape contour is significantly improved, ensuring the accuracy of the detection result. By combining the foot appearance information and the footprint area coefficient, the detection is carried out from two aspects of the appearance form and functional characteristics, providing an all-round foot evaluation. The generated foot detection report can intuitively reflect the health status and functional characteristics of the foot, providing a scientific basis for foot care, orthotic insole design, etc. It is applicable to foot image data of different shapes, has strong versatility and adaptability, and can be extended to fields such as medical treatment and sports health. Description of the Drawings

[0067] Figure 1 is a flowchart of the method provided by the present invention;

[0068] Figure 2 is a system module diagram provided by the present invention;

[0069] Figure 3 is a schematic diagram of the device provided by the present invention.

[0070] In the drawings, the component names represented by the reference numerals are as follows:

[0071] 1. Device body; 2. Light-transmitting glass; 3. Rear camera; 4. Footprint camera. Detailed Embodiments

[0072] To make the above objects, features, and advantages of the present invention more obvious and understandable, the following detailed description of the specific embodiments of the present invention will be given with reference to the accompanying drawings of the specification.

[0073] In the following description, numerous specific details are set forth to provide a thorough understanding of the present invention. However, the present invention may be practiced in other ways different from those described herein. Persons skilled in the art can make similar extensions without departing from the spirit of the present invention. Therefore, the present invention is not limited by the specific embodiments disclosed below.

[0074] Secondly, as used herein, "an embodiment" or "embodiments" refer to specific features, structures, or characteristics that may be included in at least one implementation of the present invention. The phrase "in a preferred embodiment" that appears in different places in this specification does not necessarily refer to the same embodiment, nor is it an individual or alternative embodiment that is mutually exclusive with other embodiments.

[0075] Thirdly, the present invention is described in detail in conjunction with schematic diagrams. When detailing the embodiments of the present invention, for the sake of illustration, the schematic diagrams are only examples and should not limit the scope of protection of the present invention here.

[0076] Please refer to the attached Figure 1 As shown, a foot detection method based on image recognition is provided, including:

[0077] S1. Obtain foot shape image data, and obtain foot shape contour data according to the foot shape image data;

[0078] S2. Obtain the initial pixel coordinate information of the contour corresponding to the foot shape contour data;

[0079] S3. Obtain image distortion compensation data, and obtain the corrected pixel coordinate information of the contour according to the image distortion compensation data and the initial pixel coordinate information of the contour;

[0080] S4. Obtain the foot appearance information according to the corrected pixel coordinate information of the contour, and obtain the foot appearance detection information according to the foot appearance information;

[0081] S5. Obtain the footprint area coefficient according to the corrected pixel coordinate information of the contour, and obtain the footprint detection information according to the footprint area coefficient;

[0082] S6. Obtain a foot detection report according to the footprint appearance detection information and the footprint detection information.

[0083] In the above steps S1 to S6, morphological image data of the foot is acquired, and contour data of the foot morphology is extracted through image processing techniques. Initial pixel coordinate information is extracted using the contour data, and the specific position and shape characteristics of the foot contour are recorded. During the image acquisition process, errors may be introduced due to factors such as optical devices and angles. These errors are corrected through image distortion compensation data, and the initial pixel coordinates are adjusted according to the distortion compensation data to generate corrected pixel coordinate information, ensuring the accuracy of the foot contour data. Using the corrected contour data, external features of the foot (such as foot shape, structure, etc.) are further extracted. Combining the external features, foot appearance detection information is generated to analyze the integrity and health status of the foot appearance. Based on the corrected contour pixel coordinates, the area coefficient of the footprint is calculated as an indicator of foot pressure distribution and contact area. Footprint detection information is generated according to the footprint area coefficient for evaluating foot function and its performance in contact with the ground. By integrating the foot appearance detection information and the footprint detection information, a foot detection report is generated to provide a comprehensive assessment of the foot morphology, health status, and function, significantly improving the extraction accuracy of the foot morphological contour and ensuring the accuracy of the detection results. By combining the foot appearance information and the footprint area coefficient, detection is carried out from two aspects of appearance morphology and functional characteristics, providing an all-round foot assessment. The generated foot detection report can intuitively reflect the health status and functional characteristics of the foot, providing a scientific basis for foot care, orthotic insole design, etc. It is applicable to foot image data of different morphologies, has strong generality and adaptability, and can be extended to fields such as medical care and sports health.

[0084] In a preferred embodiment, the steps of acquiring foot morphological contour data according to the foot morphological image data include:

[0085] S101. Acquire foot morphological image data;

[0086] S102. Obtain the corresponding foot morphological image according to the foot morphological image data;

[0087] S103. Obtain the corresponding foot morphological grayscale image according to the foot morphological image;

[0088] S104. Obtain the corresponding foot morphological contour data according to the foot morphological grayscale image.

[0089] As in the above steps S101 to S104, the original image data of the foot is collected by a camera or a scanner to ensure that the data contains complete morphological information of the foot. The collected image data is preprocessed (such as denoising, contrast enhancement) to generate a clear foot morphological image, and the foot morphological image is converted into a grayscale image. The distribution of grayscale values ​​reflects the light and dark relationship of the foot morphology. The grayscale image simplifies the image information, retains the main features of the foot morphology, and reduces the computational complexity. An image processing algorithm (such as an edge detection algorithm, a Canny algorithm, or a Sobel algorithm) is applied to extract the foot contour in the grayscale image. Through further morphological operations (such as expansion, corrosion, etc.), the contour lines are optimized to generate clear foot morphological contour data, ensuring that each step from image data collection to contour extraction is optimized, and finally a high-precision foot morphological contour is obtained.

[0090] In a preferred embodiment, the step of obtaining the initial pixel coordinate information of the contour corresponding to the foot shape contour data includes:

[0091] S201, acquiring corresponding foot contour world coordinate data according to foot shape contour data;

[0092] S202, acquiring a plurality of corresponding foot shape contour world coordinate points according to the foot contour world coordinate data;

[0093] S203, obtaining an internal parameter matrix and an external parameter matrix of an image acquisition device for foot morphology image data;

[0094] S204, acquiring corresponding initial pixel coordinate points of the foot contour according to a plurality of world coordinate points of the foot morphology contour, an internal parameter matrix and an external parameter matrix;

[0095] S205 , summing up the multiple initial pixel coordinate points of the foot contour, and marking the summing result as the contour initial pixel coordinate information.

[0096] As in the above steps S201 to S205, the world coordinate data of the foot contour in the three-dimensional space is generated according to the extracted foot morphological contour data. The world coordinate data is a three-dimensional coordinate defined by a global reference system, which can reflect the position and shape of the foot contour in the real space. A plurality of key contour points are selected from the world coordinate data to form a set of world coordinate points of the foot morphological contour. The internal parameter matrix is ​​an internal parameter describing the image acquisition device (such as a camera), including focal length, principal point position, pixel scaling factor, etc. The external parameter matrix is ​​an external parameter describing the device, including the position and direction of the device in the world coordinate system. The corresponding two-dimensional pixel coordinate points are calculated using the world coordinate points of the foot morphological contour, the internal parameter matrix and the external parameter matrix. The calculation formula of the initial pixel coordinate points of the foot contour is: Wherein, u represents the x-axis coordinate point of the initial pixel of the foot contour, v represents the y-axis coordinate point of the initial pixel of the foot contour, K represents the internal parameter matrix, M represents the external parameter matrix, X w represents the x-axis coordinate point of the foot contour in the world, Y w represents the y-axis coordinate point of the foot contour in the world, Z w represents the z-axis coordinate point of the foot contour in the world. Summarize all the calculated pixel coordinate points to form the initial pixel coordinate information of the complete foot contour, which can accurately map the foot contour in three-dimensional space to the image plane, ensure the accuracy of the coordinate information, effectively reduce the mapping error caused by equipment parameter errors or shooting angle changes, and improve the data reliability.

[0097] In a preferred embodiment, the steps of obtaining the image distortion compensation data and obtaining the corrected pixel coordinate information of the contour according to the image distortion compensation data and the initial pixel coordinate information of the contour include:

[0098] S301. Obtain the image distortion compensation data of the image acquisition device for the foot morphology image data;

[0099] S302. Obtain the corresponding radial distortion parameters and tangential distortion parameters according to the image distortion compensation data;

[0100] S303. Obtain the corresponding multiple initial pixel coordinate points of the foot contour according to the initial pixel coordinate information of the contour;

[0101] S304. Obtain the corresponding distance parameter of each initial pixel coordinate point of the foot contour from the origin of the pixel coordinate system;

[0102] S305. Obtain the corrected pixel coordinate points of the foot contour according to each initial pixel coordinate point of the foot contour, the corresponding distance parameter, the radial distortion parameter and the tangential distortion parameter;

[0103] S306. Summarize the multiple corrected pixel coordinate points of the foot contour, and mark the summary result as the corrected pixel coordinate information of the contour.

[0104] In the above steps S301 to S306, distortion compensation data of the image acquisition device is obtained. These data are usually generated by the calibration process of the device and contain the distortion parameters of the camera (such as radial distortion and tangential distortion parameters). According to the image distortion compensation data, the radial distortion parameters and tangential distortion parameters are extracted. Radial distortion is usually caused by the geometric characteristics of the lens and is manifested as an imbalance in the ratio between the central region and the edge region of the image. Tangential distortion is caused by the alignment problem of the camera lens and is manifested as the straight lines in the image becoming curved. According to the previously obtained initial pixel coordinate information of the contour, multiple initial pixel coordinate points of the foot contour are extracted. The distance between each initial pixel coordinate point of the foot contour and the origin of the pixel coordinate system (usually the center point of the image) is calculated. The distance parameter is used to describe the relative position of each pixel point with respect to the center of the image, which is crucial for distortion correction. Using the initial coordinates of each pixel point, the distance parameter, the radial distortion parameters, and the tangential distortion parameters, the corrected pixel coordinates of the foot contour are calculated. The calculation formula for the corrected pixel coordinates of the foot contour is where u j represents the x-axis coordinate point of the corrected pixel of the foot contour, and v j represents the y-axis coordinate point of the corrected pixel of the foot contour. u represents the x-axis coordinate point of the initial pixel of the foot contour, and v represents the y-axis coordinate point of the initial pixel of the foot contour. k 1 , k 2 and k 3 represent the radial distortion parameters, p 1 and p 2 represent the tangential distortion parameters, r represents the distance parameter. The position of each pixel point is corrected to restore the true image form. All the corrected pixel coordinates of the foot contour after distortion compensation are summarized to form complete contour corrected pixel coordinate information, eliminating the image distortion caused by the camera lens distortion (radial and tangential distortion), ensuring the accuracy of the foot contour. The foot contour in the image is closer to the true form, providing more accurate data for subsequent foot detection and analysis. It can be flexibly applied according to the distortion compensation data of different camera devices and is applicable to various types of image acquisition devices, including monocular cameras, stereo cameras, depth cameras, etc.

[0105] In a preferred embodiment, the steps of obtaining the foot appearance information according to the contour corrected pixel coordinate information and obtaining the foot appearance detection information according to the foot appearance information include:

[0106] S401. Obtain various foot appearance feature coordinates according to the contour corrected pixel coordinate information;

[0107] S402. Obtain the corresponding foot appearance feature values according to each foot appearance feature coordinate;

[0108] S403, obtaining a corresponding appearance feature table, wherein the appearance feature table includes a plurality of foot appearance feature interval values ​​and a foot appearance detection index corresponding to each foot appearance feature interval value;

[0109] S404, obtaining a target foot appearance feature interval value according to the foot appearance feature value;

[0110] S405, obtaining a corresponding foot appearance detection index from the appearance feature table according to the target foot appearance feature interval value;

[0111] S406: Summarize the various foot appearance detection indicators and mark the summary results as foot appearance detection information.

[0112] As in the above steps S401 to S406, based on the corrected foot contour pixel coordinate information, a plurality of feature coordinates related to the foot appearance are extracted. These feature coordinates may include coordinate points of key parts such as the edge, big toe, and heel of the foot, representing the main morphological features of the foot appearance. According to each extracted foot appearance feature coordinate, a corresponding appearance feature value is calculated. These feature values ​​may include the length, width, height, big toe angle, heel width, etc. of the foot, reflecting the specific parameters of the foot morphology. An appearance feature table is extracted or generated, including a plurality of foot appearance feature interval values, and a foot appearance detection index corresponding to each interval value. These interval values ​​are preset standard values ​​or generated according to actual data, which help to compare the normal range of the foot appearance feature value. For example, the normal range of a certain feature value may be a certain length interval, and exceeding the interval indicates an abnormality. According to the calculated foot appearance feature value, its corresponding target is searched and determined. Foot appearance feature interval values. These target interval values ​​are obtained by comparing the feature values ​​with the interval values ​​in the appearance feature table to ensure that the foot feature values ​​fall within the normal range. According to the target foot appearance feature interval values, the corresponding foot appearance detection indicators are found from the appearance feature table. These detection indicators are usually based on standardized appearance feature values ​​to determine whether there are abnormalities in the foot, such as whether it meets normal heel valgus, foot width and other standards. All foot appearance detection indicators are summarized to form complete foot appearance detection information. This information provides data support for subsequent foot health assessment, treatment recommendations, insole design, etc. The accuracy of the foot appearance features is ensured through precise foot contour correction and feature coordinate extraction. The foot appearance feature values ​​are compared with the standard interval values ​​to ensure the standardization and consistency of the detection. It can be adjusted according to different needs and standards to adapt to different detection scenarios, such as health assessment, medical diagnosis, footwear design, etc.

[0113] In a preferred embodiment, the steps of obtaining a footprint area coefficient according to the contour-corrected pixel coordinate information and obtaining footprint detection information according to the footprint area coefficient include:

[0114] S501. Obtain the footprint area contour coordinate information according to the profile correction pixel coordinate information, wherein the footprint area contour coordinate information includes the front foot area contour coordinate information, the mid-foot area contour coordinate information, and the rear foot area contour coordinate information;

[0115] S502. Obtain the area contours of the front foot, mid-foot, and rear foot according to the footprint contour coordinate information;

[0116] S503. Obtain the footprint area coefficient according to the area contours of the front foot, mid-foot, and rear foot;

[0117] S504. Obtain the corresponding footprint table, wherein the footprint table includes a plurality of footprint area interval coefficients and the footprint types corresponding to each footprint area interval coefficient;

[0118] S505. Obtain the corresponding target footprint area interval coefficient according to the footprint area coefficient;

[0119] S506. Obtain the corresponding footprint type from the footprint table according to the target footprint area interval coefficient and mark it as the footprint detection information.

[0120] As in the above steps S501 to S506, according to the profile correction pixel coordinate information, the contour coordinate information of the footprint area is extracted. The footprint area is divided into three areas: the forefoot, the midfoot, and the hindfoot. The contour coordinate information of each area is extracted separately. These areas represent different parts of the foot, helping to analyze the distribution of the footprint more precisely. According to the extracted contour coordinate information of the forefoot, midfoot, and hindfoot, the contour area of each area is calculated. The area of each area reflects the size of the footprint of that part. According to the area of the forefoot, midfoot, and hindfoot areas, the footprint area coefficient is calculated. The calculation formula of the footprint area coefficient is D = J / (I + J + K), where D represents the footprint area coefficient, I represents the contour area of the forefoot area, J represents the contour area of the midfoot area, and K represents the contour area of the hindfoot area. Extract or generate a footprint table, which includes multiple footprint area interval coefficients and the corresponding footprint types for each interval coefficient. The footprint table is generated according to actual data or standards, defining the footprint types corresponding to different area intervals. For example, a certain footprint area coefficient interval may correspond to flat feet, and another interval may correspond to a normal foot type. According to the calculated footprint area coefficient, find and determine the corresponding target footprint area interval coefficient. The target interval coefficient reflects the specific type of the footprint and helps to judge the health status of the footprint. According to the target footprint area interval coefficient, find and determine the corresponding footprint type from the footprint table. The footprint type can include normal feet, flat feet, high-arched feet, etc., helping to classify and diagnose the foot health status. Marked as footprint detection information, by dividing the footprint into three areas: the forefoot, the midfoot, and the hindfoot, it is possible to analyze each part of the footprint more precisely, avoiding the rough estimation in the overall area calculation and improving the detection accuracy. Corresponding different footprint area coefficient intervals to specific footprint types ensures the standardization and consistency of the footprint types and helps to judge whether there are abnormalities in the foot. It is not only applicable to foot health assessment but also can be applied to fields such as insole design and sports shoe customization, helping to achieve personalized product design and services.

[0121] In a preferred embodiment, the steps of obtaining the regional contour areas of the forefoot, midfoot, and hindfoot according to the footprint contour coordinate information include:

[0122] S5021. Obtain the regional contour inflection point coordinate information of the forefoot, midfoot, and hindfoot according to the footprint contour coordinate information;

[0123] S5022. Obtain the corresponding multiple regional contour inflection point coordinates according to the regional contour inflection point coordinate information;

[0124] S5023. Obtain the regional contour areas of the forefoot, midfoot, and hindfoot respectively according to the corresponding multiple regional contour inflection point coordinates.

[0125] In the above steps S5021 to S5023, through the footprint contour coordinate information, the inflection point coordinates in the footprint contour are identified and extracted. An inflection point is a point where the direction changes in the contour, usually representing the edge or characteristic part of the foot. According to the multiple inflection point coordinates in the footprint contour, the footprint contour is further divided into multiple regions (forefoot, midfoot, and hindfoot), and the inflection point coordinates of each region are extracted and recorded. These coordinates provide data support for subsequent area calculation. According to the inflection point coordinates of the regional contours of the forefoot, midfoot, and hindfoot, the area of each region is calculated respectively. The calculation formula for the regional contour area is where S represents the area of the regional contour, i represents the number of the inflection point coordinates of multiple regional contours, i = 1, 2, 3... n, H i represents the x-axis coordinate point of the i-th regional contour inflection point, H i+1 represents the x-axis coordinate point of the (i + 1)-th regional contour inflection point, Z i represents the y-axis coordinate point of the i-th regional contour inflection point, Z i+1 represents the y-axis coordinate point of the (i + 1)-th regional contour inflection point. Among them, when i takes the value of n, n + 1 represents 1, which helps to better understand and analyze the characteristics and differences of each part of the foot, can adapt to the changes of different foot types, and provide more personalized detection results. For example, for a pes cavus, it may show that the areas of the forefoot and hindfoot regions are larger, while the area of the midfoot region is smaller. By the steps of automatically extracting inflection points and calculating areas, manual intervention is reduced, which is suitable for large-scale foot detection tasks and can process a large amount of data in a short time.

[0126] Please refer to the attached Figure 2 As shown, the present invention also provides a foot detection system based on image recognition for the above-mentioned foot detection method based on image recognition, including:

[0127] A foot contour module for obtaining foot morphology image data and obtaining foot morphology contour data according to the foot morphology image data;

[0128] An initial pixel module for obtaining the contour initial pixel coordinate information corresponding to the foot morphology contour data;

[0129] A corrected pixel module for obtaining image distortion compensation data and obtaining contour corrected pixel coordinate information according to the image distortion compensation data and the contour initial pixel coordinate information;

[0130] An appearance detection module for obtaining foot appearance information according to the contour corrected pixel coordinate information and obtaining foot appearance detection information according to the foot appearance information;

[0131] A footprint detection module for obtaining a footprint area coefficient according to the contour corrected pixel coordinate information and obtaining footprint detection information according to the footprint area coefficient;

[0132] A foot detection module, configured to obtain a foot detection report according to the foot print appearance detection information and the foot print detection information.

[0133] As described above, the foot contour module is responsible for obtaining the foot morphology image data, extracting the foot morphology contour data according to the image data, and based on the foot morphology contour data, the initial pixel module extracts the initial pixel coordinates of the contour to form a set of initial pixel coordinate data, which represent the basic information of the foot morphology. Through the distortion compensation algorithm, the image distortion caused by reasons such as shooting angle and equipment problems is corrected. The correction pixel module uses the internal and external parameter matrices and distortion parameters of the image to correct the initial pixel coordinates, so as to obtain more accurate contour correction pixel coordinates. The appearance detection module extracts the appearance features of the foot, such as the shape, size, and proportion of the foot, by obtaining the corrected pixel coordinates, and compares them with the preset appearance features to obtain the foot appearance detection results. These detection information can help identify the health status or abnormal conditions of the foot. The foot print detection module calculates the area coefficient of the foot print by obtaining the contour and the foot print area of the foot, and determines the type of the foot print (such as normal foot print, flat foot, high arch foot, etc.) according to the preset foot print area interval coefficient table, so as to obtain the foot print detection information. The foot detection module synthesizes the detection results of the foregoing modules, combines the relevant data of the foot appearance and the foot print, and generates a detailed foot detection report. The report content includes the health status of the foot, possible abnormalities, and relevant suggestions, to help users manage foot health. It can comprehensively detect various aspects of the foot morphology, appearance, and foot print, providing all-round data support for foot health assessment. It can accurately obtain the contour and features of the foot, eliminate errors and distortions during the shooting process, and ensure the accuracy of the detection results. It can provide personalized foot health reports for each user, help users understand their foot conditions, and perform personalized care or treatment according to the detection results. It can adapt to users with different foot types and different age groups, and is applicable to a wide range of application scenarios, such as health check-ups, foot assessment for athletes, and foot care for the elderly.

[0134] Please refer to the attached Figure 3 As shown in the figure, the present invention further provides a foot detection device based on image recognition for the above-mentioned foot detection method based on image recognition, including:

[0135] The device main body 1;

[0136] A light-transmitting glass 2, which is arranged above the device main body 1 and is used for carrying the foot;

[0137] A heel camera 3, which is arranged above one side of the device main body 1 and is used for collecting the heel image data of the foot;

[0138] The footprint camera 4 is arranged inside the device main body 1 and under the light-transmitting glass 2, and is used for collecting footprint image data.

[0139] As described above, an image acquisition device, a circuit system and a processing unit are installed inside the device main body 1 to support the normal operation of the entire detection device. The light-transmitting glass 2 is used to bear the user's foot and at the same time allows light to penetrate so that the image acquisition device can obtain the footprint image. The light-transmitting glass 2 has high strength and high light transmittance, can bear the pressure of the foot, and can ensure that the optical camera clearly captures the footprint image data under the glass. The heel camera 3 is arranged on the side above the device main body 1 and is aligned with the heel area of the foot at an appropriate angle, and can capture the appearance details of the heel of the foot, such as the width, height, shape, etc. of the heel, providing key data for the foot appearance detection. The footprint camera 4 is installed under the light-transmitting glass 2 inside the device main body 1 and is aligned upward with the foot. By capturing the image of the contact area of the foot, the contour information and area distribution of the footprint are obtained, providing basic data for the footprint detection, and can collect multi-dimensional image data of the foot appearance and footprint at the same time, ensuring the comprehensiveness and accuracy of the detection result.

[0140] And, a foot detection terminal based on image recognition includes:

[0141] One or more processors;

[0142] A storage device on which one or more programs are stored;

[0143] When the one or more programs are executed by the one or more processors, the one or more processors implement the foot detection method based on image recognition.

[0144] The above is only the preferred embodiment of the present invention. It should be noted that for those of ordinary skill in the art, without departing from the principle of the present invention, several improvements and refinements can be made, and these improvements and refinements should also be regarded as the protection scope of the present invention. The structures, devices and operation methods not specifically described and explained in the present invention are implemented according to the conventional means in the art without special description and limitation.

Claims

1. A foot detection method based on image recognition, characterized in that: include: Acquire foot morphology image data, and acquire foot morphology contour data according to the foot morphology image data; Obtaining initial pixel coordinate information of the contour corresponding to the foot shape contour data; Acquire image distortion compensation data, and acquire contour correction pixel coordinate information according to the image distortion compensation data and the contour initial pixel coordinate information; Acquire foot appearance information according to the contour-corrected pixel coordinate information, and acquire foot appearance detection information according to the foot appearance information; Obtaining a footprint area coefficient according to the contour-corrected pixel coordinate information, and obtaining footprint detection information according to the footprint area coefficient; Obtain a foot inspection report based on the footprint appearance inspection information and the footprint inspection information.

2. The foot detection method based on image recognition according to claim 1, characterized in that: The steps of obtaining foot morphology image data and obtaining foot morphology contour data according to the foot morphology image data include: Acquiring foot morphology image data; Acquiring a corresponding foot morphology image according to the foot morphology image data; Acquire a corresponding foot morphology grayscale image according to the foot morphology image; The corresponding foot shape contour data is obtained according to the foot shape grayscale image.

3. The foot detection method based on image recognition according to claim 1, characterized in that: The step of obtaining initial pixel coordinate information of the contour corresponding to the foot shape contour data includes: Acquire corresponding foot contour world coordinate data according to the foot shape contour data; Acquire a plurality of corresponding foot shape contour world coordinate points according to the foot contour world coordinate data; Obtaining the intrinsic parameter matrix and extrinsic parameter matrix of the image acquisition device of the foot morphology image data; Acquire corresponding initial pixel coordinate points of the foot contour according to a plurality of world coordinate points of the foot morphology contour, an internal parameter matrix and an external parameter matrix; A plurality of initial pixel coordinate points of the foot contour are aggregated, and the aggregated result is marked as the initial pixel coordinate information of the contour.

4. The foot detection method based on image recognition according to claim 1, characterized in that: The steps of obtaining image distortion compensation data and obtaining contour correction pixel coordinate information according to the image distortion compensation data and the initial contour pixel coordinate information include: Acquire image distortion compensation data of an image acquisition device for obtaining foot morphology image data; Acquiring corresponding radial distortion parameters and tangential distortion parameters according to the image distortion compensation data; Acquire a plurality of corresponding initial pixel coordinate points of the foot contour according to the initial pixel coordinate information of the contour; Obtain the corresponding distance parameter of each initial pixel coordinate point of the foot contour from the origin of the pixel coordinate system; Obtaining foot contour correction pixel coordinate points according to each foot contour initial pixel coordinate point, a corresponding distance parameter, a radial distortion parameter, and a tangential distortion parameter; A plurality of foot contour correction pixel coordinate points are aggregated, and the aggregated result is marked as contour correction pixel coordinate information.

5. The foot detection method based on image recognition according to claim 1, characterized in that: The steps of obtaining foot appearance information according to contour-corrected pixel coordinate information and obtaining foot appearance detection information according to the foot appearance information include: Acquire various foot appearance feature coordinates according to contour-corrected pixel coordinate information; Obtaining corresponding foot appearance feature values ​​according to each foot appearance feature coordinate; Obtaining a corresponding appearance feature table, wherein the appearance feature table includes a plurality of foot appearance feature interval values ​​and a foot appearance detection index corresponding to each foot appearance feature interval value; Obtaining a target foot appearance feature interval value according to the foot appearance feature value; Obtaining a corresponding foot appearance detection index from an appearance feature table according to a target foot appearance feature interval value; A variety of foot appearance detection indicators are summarized, and the summary results are marked as foot appearance detection information.

6. The foot detection method based on image recognition according to claim 1, characterized in that: The steps of obtaining a footprint area coefficient according to the contour-corrected pixel coordinate information and obtaining footprint detection information according to the footprint area coefficient include: Acquire footprint area contour coordinate information according to the contour-corrected pixel coordinate information, wherein the footprint area contour coordinate information includes forefoot area contour coordinate information, midfoot area contour coordinate information, and hindfoot area contour coordinate information; Obtaining the area contours of the forefoot, midfoot, and hindfoot according to the footprint contour coordinate information; Obtain footprint area coefficients based on the contour areas of the forefoot, midfoot, and hindfoot regions; Obtaining a corresponding footprint table, wherein the footprint table includes a plurality of footprint area interval coefficients and a footprint type corresponding to each footprint area interval coefficient; Obtain the corresponding target footprint area interval coefficient according to the footprint area coefficient; The corresponding footprint type is obtained from the footprint table according to the target footprint area interval coefficient and marked as footprint detection information.

7. The foot detection method based on image recognition according to claim 1, characterized in that: The step of obtaining the contour areas of the forefoot, midfoot and hindfoot according to the footprint contour coordinate information includes: Obtaining the coordinate information of the inflection points of the regional contours of the forefoot, midfoot and hindfoot according to the coordinate information of the footprint contour; Acquire the coordinates of the corresponding multiple area contour inflection points according to the area contour inflection point coordinate information; The area contours of the forefoot, the midfoot and the hindfoot are respectively obtained according to the corresponding coordinates of the inflection points of the area contours.

8. A foot detection system based on image recognition, applied to the foot detection method based on image recognition according to any one of claims 1 to 7, characterized in that: include: A foot contour module, used to obtain foot morphology image data, and obtain foot morphology contour data according to the foot morphology image data; An initial pixel module is used to obtain initial pixel coordinate information of the contour corresponding to the foot morphology contour data; A correction pixel module is used to obtain image distortion compensation data, and obtain contour correction pixel coordinate information according to the image distortion compensation data and contour initial pixel coordinate information; An appearance detection module, used to obtain foot appearance information according to the contour-corrected pixel coordinate information, and obtain foot appearance detection information according to the foot appearance information; A footprint detection module, used to obtain a footprint area coefficient according to the contour-corrected pixel coordinate information, and to obtain footprint detection information according to the footprint area coefficient; The foot detection module is used to obtain a foot detection report based on the footprint appearance detection information and the footprint detection information.

9. A foot detection device based on image recognition, applied to the foot detection method based on image recognition according to any one of claims 1 to 7, characterized in that: include: Equipment body; Translucent glass, which is arranged above the main body of the device and is used to support the foot; A heel camera is arranged above one side of the device body and is used to collect heel image data of the foot pair; The footprint camera is arranged inside the device body and below the light-transmitting glass, and is used to collect footprint image data.

10. A foot detection terminal based on image recognition, characterized in that: include: one or more processors; a storage device having one or more programs stored thereon; When one or more programs are executed by one or more processors, the one or more processors implement the foot detection method based on image recognition as described in any one of claims 1 to 7.

Images