Pressure-based footprint calibration and feature extraction method and system, medium and equipment

By image processing and feature extraction of sole pressure data, combined with expansion algorithm and correction technology, the accuracy problems of footprint calibration and feature extraction in large-scale data sets and special stances are solved, achieving high accuracy and high reliability footprint calibration and feature extraction of sexuality.

CN119006607APending Publication Date: 2024-11-22BEIJING INFORMATION SCI & TECH UNIV

Patent Information

Application Number
CN202410887836.1
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-07-03
Publication Date
2024-11-22

AI Technical Summary

Technical Problem

Prior art In plantar pressure collector design and plantar pressure analysis of rare diseases, footprint calibration and feature extraction are difficult, especially in large-scale data sets and special stances.

Method used

Using pressure-based footprint calibration method, the original sole pressure data is converted into grayscale images, median filtering and contour extraction are performed, combined with the expansion algorithm and the foot breaking fusion algorithm, the complete footprint array is obtained and the sole pressure partitioning is performed. and correct, calculate accurate plantar pressure characteristic values.

Benefits of technology

Achieve high accuracy and high reliability footprint calibration. The average IoU value can reach 90% in standing posture, the average IoU value can reach 85% in walking posture, and the relative distance of the center of shape does not exceed 0.100, ensuring the complete footprint calibration. The accuracy of sexual and feature extraction is suitable for plantar pressure acquisition instrument design and rare disease analysis.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119006607A_ABST
    Figure CN119006607A_ABST
Patent Text Reader

Abstract

The invention relates to a footprint calibration and feature extraction method and system based on pressure, a medium and equipment, and the method comprises the steps: converting collected original plantar pressure data into a grayscale image, and carrying out the image preprocessing; carrying out primary rectangular frame calibration on the preprocessed image to obtain footprint coordinate information, and adopting an expansion algorithm and a broken foot fusion algorithm to obtain a complete footprint array according to a foot sole footprint breaking condition during calibration; plantar pressure partition is carried out on a footprint array obtained through primary rectangular frame calibration, footprint area correction is carried out, a footprint rectangular frame is obtained, and new plantar pressure partition coordinates are obtained; and obtaining a plantar pressure characteristic value through the corrected footprint rectangular frame and the new plantar pressure partition coordinates. According to the method, footprint areas and plantar pressure characteristic values can be accurately and reliably obtained; the method can be applied to the fields of image processing and biological feature recognition.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the technical field of image processing and biometric identification, and in particular to a pressure-based footprint calibration and feature extraction method, system, medium and equipment. Background Art

[0002] In recent years, extensive research has been conducted on the feature extraction of footprint calibration and plantar pressure at home and abroad. However, the current research on footprint calibration and plantar pressure feature extraction is mainly divided into two major areas. One part is the traditional method, which is to use various clustering algorithms to calibrate footprints, then calculate the similarity between feature values, and finally evaluate the accuracy through the classifier; the other part uses deep learning methods for footprint recognition and feature extraction, and finally calibrates footprints through similarity calculation methods. Among them, the traditional method is more suitable for small and medium-sized data set samples and footprint calibration in specific working environments; the deep learning method has higher universality and accuracy, and is suitable for large data set samples. In addition, the traditional method still plays a very important role in the design of plantar pressure collection instruments, plantar pressure analysis of rare diseases, and rapid annotation of large-scale data sets.

[0003] Therefore, how to solve the problems of footprint calibration and feature extraction in the design of plantar pressure collectors and plantar pressure analysis of rare diseases has become a technical problem that needs to be solved urgently. Summary of the invention

[0004] In view of the above problems, the purpose of the present invention is to provide a pressure-based footprint calibration and feature extraction method, system, medium and equipment, which can accurately and reliably obtain the footprint area and plantar pressure characteristic values.

[0005] To achieve the above-mentioned purpose, in the first aspect, the technical solution adopted by the present invention is: a pressure-based footprint calibration and feature extraction method, which includes: converting the collected original plantar pressure data into a grayscale image and performing image preprocessing; performing an initial rectangular frame calibration on the preprocessed image to obtain footprint coordinate information, and during the calibration, an expansion algorithm and a broken foot fusion algorithm are used to obtain a complete footprint array in the case where the sole footprint is disconnected; performing plantar pressure partitioning on the footprint array obtained by the initial rectangular frame calibration, and performing footprint area correction to obtain a footprint rectangular frame and new plantar pressure partition coordinates; obtaining a plantar pressure characteristic value through the corrected footprint rectangular frame and the new plantar pressure partition coordinates.

[0006] Furthermore, the preprocessed image is calibrated with a primary rectangular frame to obtain the footprint coordinate information, including:

[0007] The outer contour of the footprint area in the image is obtained by using the contour extraction method to obtain the footprint contour;

[0008] Use the minimum rectangular box to select the footprint outline and record the coordinate information (x, y, w, h) of the rectangular box, where x and y are the coordinates of the upper left corner of the rectangle, and w and h are the width and height of the rectangle.

[0009] Furthermore, when calibrating, the expansion algorithm and the broken foot fusion algorithm are used to deal with the disconnected sole footprint, including:

[0010] Before extracting the outer contour of the footprint area in the image, the footprint area in the image is expanded at least twice;

[0011] Extract the outer contour of the footprint area in the dilated image, calculate the contour area size in turn, and sort the contour areas in descending order;

[0012] Determine the minimum rectangular frame coordinates of the outline in turn and store them in the footprint array, and determine the comparison value w of the footprint length o ;

[0013] Traverse the footprint array and compare the footprint length w in the footprint array with the comparison value w o For comparison, if w>w o , is regarded as a complete footprint and stored in the complete footprint array; if w <w o , is regarded as an incomplete footprint;

[0014] Record the incomplete footprint as "half-foot 1" and traverse the footprint array again. If another half-foot appears, the y-axis coordinates of the two half-foot rectangles intersect and the x-axis coordinate difference is less than the comparison value w of the footprint length. o , then record the other half foot as "half foot 2", and merge "half foot 1" with "half foot 2" to form a complete footprint and store it in the complete footprint array.

[0015] Further, the footprint area correction includes:

[0016] The centroid coordinates of the heel area and the joint centroid coordinates of the metatarsal area 1 and the metatarsal area 2 are calculated from the partitioned footprint image, and the horizontal inclination angle of the straight line where the two centroids are located is calculated;

[0017] Determine the rotation center, use the centroid or mass center of the footprint rectangle as the rotation center, rotate the horizontal rectangle, and adjust the foot width to determine the width of the rotated rectangle. The rotated rectangle is obtained with [centroid (x c ,y c ), width (w′), height (h′), tilt angle (angle)], the size and relative position of each partition and the overall area do not change, and the tilt angle of each partition area is consistent with the overall area.

[0018] Furthermore, when rotating around the centroid, for the footprint rectangle, the centroid coordinates before and after the rotation remain unchanged;

[0019] When rotating around the center of mass, the coordinates of the center of mass after rotation A(x0, y0) remain unchanged, and the coordinates of the centroid before rotation are B(x c ,y c ), the rotation angle is α, and the coordinates of the centroid after rotation are set to C(x c ′,y c ′), then the solution of C is:

[0020] Establish a coordinate system XOY, with the x-axis pointing right and the y-axis pointing downward, and draw A, B, and C;

[0021] Draw a horizontal auxiliary line AN through point A, and establish a polar coordinate system with A as the origin, AN as the axis, and clockwise as the positive direction; let the angle between AB and AN be θ, the angle between AC and AN be θ′, AB=AC=r, and the coordinates of points A, B, and C in the polar coordinate system are A(0,0), B(r,θ), and C(r,θ′) respectively;

[0022] The coordinates of the three points ABC obtained in the polar coordinate system are translated and converted back to the rectangular coordinate system XOY to obtain the coordinates of point C (x c ′,y c ′) is: C(r·cosθ′+x0, r·sinθ′+y0).

[0023] Further, foot width adjustment, including:

[0024] Based on the extracted footprint contour shape, cut it off from the x-axis direction and take the contours in the first and second plantar regions;

[0025] Record the difference between the maximum and minimum values ​​of the y-axis coordinates in the extracted contour, and record the maximum difference as the foot width standard width_standard;

[0026] The rotated foot width h′=(width_standard) / cos(angle) is obtained according to the foot width standard, where angle is the inclination angle.

[0027] Furthermore, the new plantar pressure partition coordinates are obtained by determining the centroid coordinates of each partition after rotation, including:

[0028] Let the rectangle of each partition be [(x ci ,y ci ), w i 、h i , angle], i represents the ith partition, and the centroid coordinates of the overall area are C(x c ′,y c′), the width is w, the height is h, the inclination angle is angle, and the centroid coordinates of each partition area are P i (x ci ,y ci );

[0029] Create a coordinate system XOY, with the x axis to the right and the y axis downward, and draw C and P accordingly. i and the boundaries of each subdistrict;

[0030] With point C as the origin, the center line of the footprint rectangle to the right as the x' axis, and the center line of the footprint rectangle downward as the y' axis, a rectangular coordinate system X'CY' is established. In this coordinate system, the polar coordinates of the centroid of each partition are P i (x ci ,y ci );

[0031] With point C as the origin and the center line of the footprint rectangle to the right as the polar axis r1, the first polar coordinate system CR1 is established, and the polar coordinates of the centroid of each partition in this coordinate system are calculated;

[0032] With point C as the origin, parallel to the positive direction of the x-axis, a second polar coordinate system CR2 is established. The angle between the second polar coordinate system CR2 and the polar axis of the first polar coordinate system CR1 is ɑ=angle; in the CR2 coordinate system, P i Recorded as in Therefore, i Coordinate is P i (r i ,θ i +α);

[0033] P i The coordinates are converted to the rectangular coordinate system XOY, then P i Coordinate is P i (r i ·cos(θ i +α)+x c ′,r i ·sin(θ i +α)+y c ′), and obtain the centroid coordinates of each partition to serve as the new plantar pressure partition coordinates.

[0034] In the second aspect, the technical solution adopted by the present invention is: a pressure-based footprint calibration and feature extraction system, which includes: a preprocessing module, which converts the collected original plantar pressure data into a grayscale image and performs image preprocessing; a primary calibration module, which performs primary rectangular frame calibration on the preprocessed image to obtain footprint coordinate information, and adopts an expansion algorithm and a broken foot fusion algorithm to obtain a complete footprint array when the sole footprint is disconnected during calibration; a correction module, which partitions the footprint array obtained by the primary rectangular frame calibration, and performs footprint area correction to obtain a footprint rectangular frame and new plantar pressure partition coordinates; a eigenvalue calculation module, which obtains the plantar pressure eigenvalue through the corrected footprint rectangular frame and the new plantar pressure partition coordinates.

[0035] In a third aspect, the technical solution adopted by the present invention is: a computer-readable storage medium storing one or more programs, wherein the one or more programs include instructions, and when the instructions are executed by a computing device, the computing device executes any one of the above methods.

[0036] In a fourth aspect, the technical solution adopted by the present invention is: a computing device, comprising: one or more processors, a memory and one or more programs, wherein the one or more programs are stored in the memory and configured to be executed by the one or more processors, and the one or more programs include instructions for executing any of the above methods.

[0037] The present invention adopts the above technical solution, which has the following advantages:

[0038] 1. The present invention uses the original data image of the acquisition board as input, and adopts digital image processing methods such as median filtering, contour extraction, image expansion, and coordinate calculation methods to complete preprocessing, footprint rectangular frame calibration, plantar pressure zoning, footprint area correction, and eigenvalue calculation, and finally obtains clear, accurate, and morphologically and positionally favorable footprint areas and plantar pressure eigenvalues. Among them, in view of the situation where the footprint area is disconnected, a footprint fusion algorithm is introduced to ensure the integrity of the footprint calibration; in view of the defect problem of horizontal rectangular frame calibration under special standing postures, a footprint area correction algorithm is proposed to ensure the accuracy of the footprint characteristics. In addition, the effectiveness and reliability of the present invention are proved through the footprint calibration reliability verification experiment.

[0039] 2. The present invention has broad application prospects in the design of plantar pressure collection instruments, plantar pressure feature analysis of rare diseases, and large-scale annotation of footprint data sets. In addition, it can also be used as the original input of deep learning models related to footprint recognition and gait recognition, which can greatly shorten the training time and improve the recognition rate of the model. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] Figure 1is a flow chart of a pressure-based footprint calibration and feature extraction method in an embodiment of the present invention;

[0041] Figure 2a is a schematic diagram of an original image input in the image preprocessing stage in an embodiment of the present invention;

[0042] Figure 2b is a schematic diagram of median filtering of an original image in the image preprocessing stage in an embodiment of the present invention;

[0043] Figure 2c is a schematic diagram of an image after threshold binarization is performed in the image preprocessing stage in an embodiment of the present invention;

[0044] Figure 3 is a schematic diagram of the disconnection of the sole footprint in an embodiment of the present invention;

[0045] Figure 4a is a schematic diagram of a footprint image after multiple expansions during the rectangular frame calibration process in an embodiment of the present invention;

[0046] Figure 4b This is a schematic diagram of a rectangular frame after calibration in an embodiment of the present invention;

[0047] Figure 5 is a schematic diagram of severed foot fusion in an embodiment of the present invention;

[0048] Figure 6 Schematic diagram of a six-zone solution for the sole of the foot in an embodiment of the present invention;

[0049] Figure 7 is a schematic diagram of a coordinate system XOY for solving the centroid coordinates C after rotation in an embodiment of the present invention;

[0050] Figure 8a is the image before rotation in the rectangular frame rotation stage in the embodiment of the present invention;

[0051] Figure 8b is the image after rotation in the rectangular frame rotation stage in the embodiment of the present invention;

[0052] Figure 9a is a footprint image using the centroid rotation center in an embodiment of the present invention;

[0053] Figure 9b is a footprint image using the center of mass rotation in an embodiment of the present invention;

[0054] Fig.10 is a schematic diagram of the plantar pressure division after rotation correction in an embodiment of the present invention;

[0055] Fig.11 is a schematic diagram of a coordinate system XOY for establishing and solving new plantar pressure zone coordinates in an embodiment of the present invention;

[0056] Fig.12a is an IoU scatter plot of the standing posture experimental data in an embodiment of the present invention;

[0057] Figure 12b is a scatter plot of the centroid relative distance of the standing posture experimental data in an embodiment of the present invention;

[0058] Fig.13a is an IoU scatter plot of walking posture experimental data in an embodiment of the present invention;

[0059] Fig.13b It is a scatter plot of the relative distance of the centroid of the walking posture experimental data in an embodiment of the present invention. DETAILED DESCRIPTION

[0060] With the development of the information age, the study of human biometrics has attracted widespread attention in important fields such as security, criminal investigation, medical treatment and sports rehabilitation. Gait is a unique behavioral characteristic of human walking, which contains a large amount of important data such as physical condition and identity information. The plantar pressure distribution characteristics, as one of the main ways to study gait, not only help doctors accurately assess the health of patients' feet, but also provide an important basis for rehabilitation treatment and sports training. At the same time, plantar pressure characteristics, as an important indicator reflecting individual muscle force habits and gait patterns, are also of great significance for the study of human biometric recognition technology. Therefore, it is of great practical significance to study the footprint calibration and feature extraction methods for plantar pressure.

[0061] In order to solve the problem of footprint calibration and feature extraction in the design of plantar pressure collector and plantar pressure analysis of rare diseases, the present invention takes the original data of the thin-film plantar pressure collection board as the research object, combines the digital image processing algorithm and the time-space coordinate calculation method, and proposes a pressure-based footprint calibration and feature extraction method, system, medium and equipment with high accuracy and good calibration effect. The present invention includes: first, converting the pressure data received by the collection board into a grayscale image, and performing median filtering and threshold binarization on it to remove noise interference; secondly, obtaining the outer contour of each footprint through a contour extraction algorithm, and calibrating it using a minimum rectangular frame; wherein, for the case where the sole footprint is disconnected, an expansion algorithm and a broken foot fusion algorithm are introduced; at the same time, in order to obtain more accurate plantar pressure characteristics, the plantar pressure of the initially calibrated footprint is partitioned; thirdly, a footprint area correction algorithm is proposed to obtain a more accurate and tighter footprint rectangular frame; finally, according to the corrected footprint rectangular frame and its partition coordinates, the plantar pressure characteristic value can be calculated. The reliability verification experiment of footprint calibration shows that the average value of IoU (Intersection over Union) in standing posture can reach more than 90%, and the average value of IoU in walking posture can reach more than 85%; the accuracy of footprint calibration of both is 100%; the relative distance of centroids does not exceed 0.100, and is even lower in standing posture. This proves that the present invention has high reliability.

[0062] In order to make the purpose, technical solution and advantages of the embodiment of the present invention clearer, the technical solution of the embodiment of the present invention will be clearly and completely described below in conjunction with the drawings of the embodiment of the present invention. Obviously, the described embodiment is a part of the embodiment of the present invention, not all of the embodiments. Based on the described embodiment of the present invention, all other embodiments obtained by ordinary technicians in this field belong to the scope of protection of the present invention.

[0063] It should be noted that the terms used herein are only for describing specific embodiments and are not intended to limit exemplary embodiments according to the present invention. As used herein, unless the context clearly indicates otherwise, the singular form is also intended to include the plural form. In addition, it should be understood that when the terms "comprising" and / or "including" are used in this specification, it indicates the presence of features, steps, operations, devices, components and / or combinations thereof.

[0064] In one embodiment of the present invention, a pressure-based footprint calibration and feature extraction method is provided. In this embodiment, plantar pressure feature extraction is to analyze the plantar pressure characteristics of a subject at a certain moment when the subject is standing still or walking, which is mainly divided into image preprocessing, rectangular frame calibration, plantar pressure zoning, footprint area correction and plantar pressure feature value calculation. Specifically, Figure 1As shown, the method comprises the following steps:

[0065] 1) Convert the collected original plantar pressure data into a grayscale image and perform image preprocessing;

[0066] 2) Performing initial rectangular frame calibration on the preprocessed image to obtain footprint coordinate information. During calibration, if the footprints on the soles of the feet are disconnected, an expansion algorithm and a broken foot fusion algorithm are used to obtain a complete footprint array;

[0067] 3) partitioning the footprint array obtained by the initial rectangular frame calibration into plantar pressure zones, and performing footprint area correction to obtain a footprint rectangular frame and new plantar pressure zone coordinates;

[0068] 4) The plantar pressure characteristic value is obtained through the corrected footprint rectangle and the new plantar pressure partition coordinates.

[0069] In the above step 1), the image preprocessing stage is used to remove noise interference in the image; it includes the following steps:

[0070] 1.1) For the original image (such as Figure 2a Median filtering can effectively remove large-scale noise points, such as Figure 2b shown.

[0071] 1.2) Binarize the filtered image with an adaptive threshold to obtain a clear and interference-free plantar pressure image, such as Figure 2c shown.

[0072] In the above step 2), the rectangular frame calibration stage is used to preliminarily determine the footprint coordinate information and perform the minimum rectangular frame calibration on it. Specifically, the initial rectangular frame calibration of the preprocessed image to obtain the footprint coordinate information includes the following steps:

[0073] 2.1.1) Using the contour extraction method to obtain the outer contour of the footprint area in the image, the footprint contour is obtained;

[0074] 2.1.2) Use the smallest rectangular box to select the footprint outline and record the coordinate information of the rectangular box (x, y, w, h), where x and y are the coordinates of the upper left corner of the rectangle, and w and h are the width and height of the rectangle.

[0075] However, when the human body is standing, the force is mainly concentrated on the heel, metatarsal bones and toes of the sole of the foot, and the arch area has very limited load-bearing capacity, so there will be a situation where the footprint contour of the same sole is disconnected, such as Figure 3 If the image is directly subjected to contour extraction and minimum rectangular frame calibration, the area of ​​the same sole will be separated. Therefore, the present invention adopts an expansion algorithm and a broken foot fusion algorithm to process the situation where the sole footprint is disconnected.

[0076] In the above step 2), when the sole footprint is disconnected during calibration, an expansion algorithm and a broken foot fusion algorithm are used, which includes the following steps:

[0077] 2.2.1) Before extracting the outer contour of the footprint area in the image, the footprint area in the image is expanded at least twice to ensure that the disconnected areas in the image can be connected together, such as Figure 4a As shown;

[0078] The footprint area is then calibrated with a rectangular box on the dilated image, which can solve the problem of slight image disconnection.

[0079] exist Figure 4a In the figure, it can be clearly seen that the footprints in the right foot area are slightly disconnected and can be connected by multiple expansions, but the entire arch area of ​​the left foot is severely disconnected and cannot be connected by expansion. Therefore, further processing is performed through the broken foot fusion algorithm in the following steps.

[0080] 2.2.2) The image is expanded multiple times to solve the problem of slight disconnection, and then the outer contour of the footprint area in the expanded image is extracted, the size of the contour area is calculated in turn, and the contour area is sorted in order from large to small;

[0081] 2.2.3) Determine the minimum rectangular frame coordinates of the contour in turn and store them in the footprint array, and determine the comparison value w of the footprint length o ;

[0082] Among them, w o It is 70% of the actual foot length, which is determined based on the subject's foot length.

[0083] 2.2.4) Traverse the footprint array and compare the footprint length w in the footprint array with the comparison value w o For comparison, if w>w o , is regarded as a complete footprint and stored in the complete footprint array; if w <w o , is considered as an incomplete footprint, and the other half needs to be found;

[0084] 2.2.5) For incomplete footprints, such as Figure 5 As shown, the incomplete footprint is recorded as "half foot 1", and the footprint array is traversed again. If another half foot appears, the y-axis coordinates of the two half-foot rectangles intersect and the x-axis coordinate difference is less than the comparison value w of the footprint length. o , then record the other half foot as "half foot 2", and merge "half foot 1" with "half foot 2" to form a complete footprint and store it in the complete footprint array.

[0085] Through the above-mentioned broken foot fusion algorithm of the present invention, the complete footprint array can be accurately obtained and the rectangular frame calibration work can be completed. Figure 4b shown.

[0086] In the above step 3), the footprint array obtained by the initial rectangular frame calibration is partitioned into plantar pressure zones, specifically:

[0087] The plantar pressure zoning stage is mainly used to analyze the microscopic information of a single footprint. It is necessary to divide the entire sole of the foot into multiple regions in order to extract features from each region.

[0088] In medical research, the force of the foot is usually divided into eight zones according to the structure of the human foot: the first toe, the second toe, the third to fifth toes, the first metatarsal, the second metatarsal, the third to fifth metatarsal, the arch and the heel bone. There are many mainstream plantar pressure zoning schemes, such as the six-zone scheme, the eight-zone scheme, the four-part ten-zone scheme, etc.

[0089] In this embodiment, a six-zone scheme is adopted to divide the sole of the foot into toe zone 1, toe zone 2, plantar zone 1, plantar zone 2, arch zone and heel zone. Figure 6 The location details of each area are as follows:

[0090] Toe zone 1: The medial area of ​​the front 1 / 6 of the plantar area, corresponding to the first toe area in medicine.

[0091] Toe zone 2: The outer area of ​​the front 1 / 6 of the plantar area, corresponding to the second toe and the third to fifth toe area in medicine.

[0092] Plantar area 1: The medial area of ​​the front 1 / 6 to 1 / 2 of the plantar area, corresponding to the first metatarsal area in medicine.

[0093] Metatarsal zone II: The medial area of ​​the front 1 / 6 to 1 / 2 of the plantar area, corresponding to the second metatarsal bone and the third to fifth metatarsal bones in medicine.

[0094] Arch area: the front 1 / 2 to 1 / 3 area of ​​the sole of the foot, corresponding to the arch area in medicine.

[0095] Heel area: The posterior 1 / 3 area of ​​the sole of the foot, corresponding to the root bone area in medicine.

[0096] According to the above coordinate partitioning scheme, further calculations can be made based on the complete footprint rectangle to obtain the coordinate position information after partitioning, but the method of calibrating with a horizontal rectangular frame is not universal. When the tester adopts a special standing posture (outward-facing, inward-facing, tilted standing, etc.), the calibration method of the horizontal rectangular frame is no longer accurate, and it is difficult to perform reasonable plantar pressure partitioning. Therefore, footprint area correction is required.

[0097] In the above step 3), the footprint area correction stage is used to dynamically adjust the tilt angle of the footprint rectangle according to the different standing postures of the tester to improve the accuracy of footprint area recognition, thereby improving the accuracy of static feature calculation. The footprint area correction method in this embodiment is based on the footprint image after the initial partition (such as Figure 6 (as shown) calculate the footprint inclination angle, select the rotation center, rotate and adjust the footprint rectangular frame, draw a new footprint area partition image, and finally calculate the coordinates of each partition after correction.

[0098] Specifically, the footprint area correction used in this embodiment includes the following steps:

[0099] 3.1) Calculate the coordinates of the center of mass (pressure center) of the heel area and the joint center of mass of metatarsal area 1 and metatarsal area 2 from the partitioned footprint image, and calculate the horizontal inclination angle of the straight line where the two center of mass are located;

[0100] 3.2) Determine the rotation center, take the centroid or mass center of the footprint rectangle as the rotation center, rotate the horizontal rectangle, and adjust the width of the foot to determine the width of the rotated rectangle, and obtain the rotated rectangle (x, y, w, h), with [centroid (x c ,y c ), width (w′), height (h′), tilt angle (angle)], the size and relative position of each partition and the overall area do not change, and the tilt angle of each partition area is consistent with the overall area.

[0101] In this embodiment, when the center of mass is used as the rotation center, the center of mass refers to the center of mass of the rectangular frame, and is also the center of mass of the footprint, because the footprint is completely contained in the rectangular frame. The calculation of the center of mass is based on the calculation of the centroid and the actual pressure value of each point is added. Therefore, only the footprint landing area in the rectangular frame has pressure, and the area without the footprint landing point does not produce pressure, so the center of mass of the rectangular frame is the center of mass of the footprint. The center of mass is actually the pressure weighted center point.

[0102] In the above step 3.2), when rotating with the centroid as the center, for the footprint rectangle, the centroid coordinates before and after the rotation remain unchanged, both are (x c ,y c ), where x c =x+w / 2,y c =y+h / 2.

[0103] When rotating around the center of mass, the coordinates of the center of mass after rotation A(x0, y0) remain unchanged, and the coordinates of the centroid before rotation are B(x c ,y c ), the rotation angle is α, and the coordinates of the centroid after rotation are set to C(x c ′,y c ′), then the solution of C includes the following steps:

[0104] 3.2.1.1) Establish the coordinate system XOY, with the x-axis pointing to the right and the y-axis pointing downward, and draw A, B, and C. The mathematical model is as follows Figure 7 As shown;

[0105] 3.2.1.2) Draw a horizontal auxiliary line AN through point A, and establish a polar coordinate system with A as the origin, AN as the axis, and clockwise as the positive direction; let the angle between AB and AN be θ, the angle between AC and AN be θ′, AB=AC=r, and the coordinates of points A, B, and C in the polar coordinate system are A(0,0), B(r,θ), and C(r,θ′) respectively; where:

[0106]

[0107]

[0108] θ′=θ+α (3)

[0109] 3.2.1.3) Translate the coordinates of the three points ABC obtained in the polar coordinate system and convert them back to the rectangular coordinate system XOY to obtain the coordinates of point C (x c ′,y c ')for:

[0110] C(r·cosθ′+x0,r·sinθ′+y0). (4)

[0111] Substituting formula (1) to formula (3) into formula (4), the coordinates of point C are obtained as follows:

[0112]

[0113] At this time, the horizontal rectangular frame is rotated with the centroid and the center of mass as the rotation center, and the foot width is adjusted, and then the rotation scheme is determined according to the actual visual effect. The foot width adjustment process is used to determine the width of the rotated rectangle, because the initial rectangular frame calibration scheme will cause the rectangle width to be much larger than the actual foot width under different standing postures, such as Figure 8a , Figure 8b As shown, the width of the rotated rectangle needs to be adjusted.

[0114] In the above step 3.2), the foot width adjustment includes the following steps:

[0115] 3.2.2.1) Based on the extracted footprint contour shape, cut it off from the x-axis direction and take the contours in the plantar region 1 and the plantar region 2;

[0116] 3.2.2.2) Record the difference between the maximum and minimum values ​​of the y-axis coordinates of the extracted contour, and record the maximum difference as the foot width standard width_standard;

[0117] 3.2.2.3) According to the foot width standard, the rotated foot width h′=(width_standard) / cos(angle), where angle is the inclination angle.

[0118] contrast Figure 9a , Figure 9b It can be found that Figure 9b (Image rotated around the center of mass) The distance from the inside of the footprint to the bottom border of the rectangular frame is shorter, and the footprint is more centered within the rectangular frame as a whole; Figure 9a (The image rotated with the centroid as the center) has a larger gap at the lower boundary and a smaller gap at the upper boundary, which is not centered enough. Therefore, in this embodiment, the preferred rotation center is the centroid.

[0119] The pressure distribution of the plantar after rotation correction is as follows Fig.10 The last step in the footprint zone correction phase is to calculate the coordinates of the six plantar pressure zones after rotation.

[0120] In the above step 3), the rotated rectangle is [centroid (x c ′,y c ′), width (w′), height (h′), and inclination angle (angle)]. Moreover, before and after the rotation, the size and relative position of the six partitions and the overall area do not change, and the inclination angle of each area is consistent with the overall area. Therefore, it is only necessary to determine the centroid coordinates of the six partitions after rotation to determine the new plantar pressure partition area.

[0121] Therefore, in this embodiment, the new plantar pressure zone coordinates are obtained by determining the centroid coordinates of each zone after rotation, including the following steps:

[0122] 3.3) Let the rectangle of each partition be [(x ci ,y ci ), w i 、h i , angle], i represents the serial number of the ith partition, and the centroid coordinates of the entire area are C(x c ′,y c ′), the width is w, the height is h, the inclination angle is angle, and the centroid coordinates of each partition area are P i (x ci ,y ci ); In this embodiment, the left foot is taken as the research object for illustration.

[0123] 3.4) Establish the coordinate system XOY, with the x-axis pointing to the right and the y-axis pointing downward, and draw C, P accordingly. i and the boundaries of each partition, the mathematical model is as follows Fig.11 As shown;

[0124] 3.5) With point C as the origin, the center line of the footprint rectangle to the right as the x' axis, and the center line of the footprint rectangle downward as the y' axis, a rectangular coordinate system X'CY' is established. In this coordinate system, the polar coordinates of the centroid of each partition are P i (x ci ,y ci );

[0125] In this embodiment, the polar coordinates of the centroid of each partition are P i (x ci ,y ci ) are:

[0126]

[0127] 3.6) With point C as the origin and the center line of the footprint rectangle to the right as the polar axis r1, establish the first polar coordinate system CR1 and calculate the polar coordinates of the centroid of each partition in this coordinate system:

[0128] P i (r i ,θ i ) (7)

[0129] 3.7) With point C as the origin, parallel to the positive direction of the x-axis (horizontal direction), establish a second polar coordinate system CR2. The angle between the second polar coordinate system CR2 and the polar axis of the first polar coordinate system CR1 is α=angle (α is the rotation angle angle from the horizontal direction to the inclined state); in the CR2 coordinate system, P i Recorded as in Therefore, i The coordinates are:

[0130] P i (r i ,θ i +α) (8)

[0131] 3.8) P i The coordinates are converted to the rectangular coordinate system XOY, then P i The coordinates are:

[0132] P i (r i ·cos(θ i +α)+x c ′,r i ·sin(θ i +α)+y c ′), (9)

[0133] Substituting each coordinate in the X′CY′ coordinate system into the above formula, the centroid coordinates of each partition can be calculated to serve as the new plantar pressure partition coordinates.

[0134] If the right foot is used as the research object, the coordinate algebraic expressions of the six points are the same as those of the left foot. However, since the first and third partitions correspond to the first phalanx and metatarsal areas, respectively, and should be located on the inside of both feet, the coordinates of P1 and P2, and P3 and P4 should be exchanged. At this point, the footprint area correction work is completely completed, and after obtaining the new plantar pressure partition coordinates, the plantar pressure characteristic value calculation can be performed.

[0135] In the above step 4), the calculation of the plantar pressure characteristic value needs to be combined with the obtained footprint contour coordinate information and the coordinate information of each partition. The static characteristic value is shown in Table 1 (where C 面 , C 压 etc. are all constants), including but not limited to - force area, average pressure, peak pressure, COP (center of pressure) coordinates, foot length, foot width and other footprint characteristics, as well as plantar zoning characteristics such as the proportion of force area in each zone and the proportion of pressure in each zone.

[0136] Table 1 Eigenvalue calculation formula

[0137]

[0138] In one embodiment of the present invention, a pressure-based footprint calibration and feature extraction system is provided, which includes:

[0139] A preprocessing module converts the collected raw plantar pressure data into a grayscale image and performs image preprocessing;

[0140] The initial calibration module performs the initial rectangular frame calibration on the preprocessed image to obtain the footprint coordinate information. During the calibration, the expansion algorithm and the broken foot fusion algorithm are used to obtain the complete footprint array in case of disconnected sole footprints;

[0141] The correction module partitions the plantar pressure of the footprint array obtained by the initial rectangular frame calibration, and performs footprint area correction to obtain the footprint rectangular frame and obtain new plantar pressure partition coordinates;

[0142] The characteristic value calculation module obtains the characteristic value of plantar pressure through the corrected footprint rectangular frame and the new plantar pressure partition coordinates.

[0143] In the above embodiment, the pre-processed image is initially calibrated with a rectangular frame to obtain footprint coordinate information, including:

[0144] The outer contour of the footprint area in the image is obtained by using the contour extraction method to obtain the footprint contour;

[0145] Use the minimum rectangular box to select the footprint outline and record the coordinate information (x, y, w, h) of the rectangular box, where x and y are the coordinates of the upper left corner of the rectangle, and w and h are the width and height of the rectangle.

[0146] In the above embodiment, when calibrating the disconnected sole footprint, an expansion algorithm and a broken foot fusion algorithm are used, including:

[0147] Before extracting the outer contour of the footprint area in the image, the footprint area in the image is expanded at least twice;

[0148] Extract the outer contour of the footprint area in the dilated image, calculate the contour area size in turn, and sort the contour areas in descending order;

[0149] Determine the minimum rectangular frame coordinates of the outline in turn and store them in the footprint array, and determine the comparison value w of the footprint length o ;

[0150] Traverse the footprint array and compare the footprint length w in the footprint array with the comparison value w o For comparison, if w>w o , is regarded as a complete footprint and stored in the complete footprint array; if w <w o , is regarded as an incomplete footprint;

[0151] Record the incomplete footprint as "half-foot 1" and traverse the footprint array again. If another half-foot appears, the y-axis coordinates of the two half-foot rectangles intersect and the x-axis coordinate difference is less than the comparison value w of the footprint length. o , then record the other half foot as "half foot 2", and merge "half foot 1" with "half foot 2" to form a complete footprint and store it in the complete footprint array.

[0152] In the above embodiment, the footprint area correction includes:

[0153] The centroid coordinates of the heel area and the joint centroid coordinates of the metatarsal area 1 and the metatarsal area 2 are calculated from the partitioned footprint image, and the horizontal inclination angle of the straight line where the two centroids are located is calculated;

[0154] Determine the rotation center, use the centroid or mass center of the footprint rectangle as the rotation center, rotate the horizontal rectangle, and adjust the foot width to determine the width of the rotated rectangle. The rotated rectangle is obtained with [centroid (x c ,y c ), width (w′), height (h′), tilt angle (angle)], the size and relative position of each partition and the overall area do not change, and the tilt angle of each partition area is consistent with the overall area.

[0155] In this embodiment, when rotating with the centroid as the center, for the footprint rectangular frame, the centroid coordinates before and after the rotation remain unchanged;

[0156] When rotating around the center of mass, the coordinates of the center of mass after rotation A(x0, y0) remain unchanged, and the coordinates of the centroid before rotation are B(x c ,y c ), the rotation angle is α, and the coordinates of the centroid after rotation are set to C(x c ′,y c ′), then the solution of C is:

[0157] Establish a coordinate system XOY, with the x-axis pointing right and the y-axis pointing downward, and draw A, B, and C;

[0158] Draw a horizontal auxiliary line AN through point A, and establish a polar coordinate system with A as the origin, AN as the axis, and clockwise as the positive direction; let the angle between AB and AN be θ, the angle between AC and AN be θ′, AB=AC=r, and the coordinates of points A, B, and C in the polar coordinate system are A(0,0), B(r,θ), and C(r,θ′) respectively;

[0159] Translate the coordinates of points A, B, and C in the polar coordinate system and convert them back to the rectangular coordinate system XOY to obtain the coordinates of point C (x c ′,y c ′) is: C(r·cosθ′+x0, r·sinθ′+y0).

[0160] In this embodiment, the foot width adjustment includes:

[0161] Based on the extracted footprint contour shape, cut it off from the x-axis direction and take the contours in the first and second plantar regions;

[0162] Record the difference between the maximum and minimum values ​​of the y-axis coordinates in the extracted contour, and record the maximum difference as the foot width standard width_standard;

[0163] The rotated foot width h′=(width_standard) / cos(angle) is obtained according to the foot width standard, where angle is the inclination angle.

[0164] In the above embodiment, the new plantar pressure zone coordinates are obtained by determining the centroid coordinates of each zone after rotation, including:

[0165] Let the rectangle of each partition be [(x ci ,y ci ), w i 、h i , angle], i represents the ith partition, and the centroid coordinates of the overall area are C(x c ′,y c ′), the width is w, the height is h, the inclination angle is angle, and the centroid coordinates of each partition area are P i (x ci ,yci );

[0166] Create a coordinate system XOY, with the x axis to the right and the y axis downward, and draw C and P accordingly. i and the boundaries of each subdistrict;

[0167] With point C as the origin, the center line of the footprint rectangle to the right as the x' axis, and the center line of the footprint rectangle downward as the y' axis, a rectangular coordinate system X'CY' is established. In this coordinate system, the polar coordinates of the centroid of each partition are P i (x ci ,y ci );

[0168] With point C as the origin and the center line of the footprint rectangle to the right as the polar axis r1, the first polar coordinate system CR1 is established, and the polar coordinates of the centroid of each partition in this coordinate system are calculated;

[0169] With point C as the origin, parallel to the positive direction of the x-axis, a second polar coordinate system CR2 is established. The angle between the second polar coordinate system CR2 and the polar axis of the first polar coordinate system CR1 is ɑ=angle; in the CR2 coordinate system, P i Recorded as in Therefore, i Coordinate is P i (r i ,θ i +α);

[0170] P i The coordinates are converted to the rectangular coordinate system XOY, then P i Coordinate is P i (r i ·cos(θ i +α)+x c ′,r i ·sin(θ i +α)+y c ′), and obtain the centroid coordinates of each partition to serve as the new plantar pressure partition coordinates.

[0171] The system provided in this embodiment is used to execute the above-mentioned method embodiments. Please refer to the above-mentioned embodiments for specific processes and detailed contents, which will not be repeated here.

[0172] Embodiment, the present invention is further explained through specific experiments and analysis of their results. In this embodiment, a plantar pressure collection board is used as the main research platform, which is composed of multiple independent plantar pressure collection units, which can be dynamically spliced ​​and combined according to the actual research scenario. The basic parameters of each independent plantar pressure collection unit are shown in Table 2.

[0173] Table 2 Basic parameters of plantar pressure acquisition unit

[0174]

[0175] (1) Evaluation indicators

[0176] ① Accuracy of footprint calibration: Accuracy is an important indicator for evaluating target detection algorithms. Since the experimental platform in this embodiment sends experimental data in a loop in the form of multiple frames of images, the accuracy of footprint calibration can be calculated in combination with the image sequence information. When standing statically, there should be two footprints in the same frame of image, so it is only necessary to record the image starting sequence number when the two-foot footprints appear and the ending sequence number when the unilateral footprint disappears, and combine it with the number of measured footprints to calculate the accuracy of footprint calibration; in the dynamic walking stage, due to the single support phase in the gait process, 1 to 2 footprints can appear in one frame of image, and there are incomplete footprints, so it is necessary to record the number of complete footprints to be measured one by one, and combine it with the number of measured footprints to calculate the accuracy of footprint calibration.

[0177] ②IoU (Intersection over Union): IoU is the ratio of the intersection area and the union area between the calibrated rectangular box and the real footprint. In order to ensure the accuracy of footprint calibration and plantar pressure partitioning features, it is necessary to ensure that IoU is as high as possible to reduce the impact of blank areas.

[0178] ③ Relative distance of centroid: Calculate the Euclidean distance between the centroid of the footprint and the centroid of the rectangular frame. To eliminate the influence of different footprint sizes, divide it by the footprint width to obtain the relative distance of the centroid. In order to make the footprint sufficiently centered in the rectangular frame, the relative distance of the centroid needs to be as low as possible to ensure the accuracy of the plantar pressure zoning method mentioned above.

[0179] (2) Verification of rotation center

[0180] The main method of the experiment is: for the same set of pressure footprint data, rotation correction is performed with the center of mass or the center of shape as the center, and the experimental results are compared to prove the correctness of determining the center of mass as the rotation center in this embodiment.

[0181] A subject (age 24 years, height 170 cm, weight 85 kg) was recruited to stand on the plantar pressure collection board for 30 seconds. A total of 150-180 complete footprints were collected, and the footprints were calibrated. The rectangular rotation of the center of mass or center of shape was performed, and the IoU and relative distance of the center of shape of each footprint were recorded. The scatter plot was observed, and the average value was recorded. The experimental results were analyzed, as shown in Table 3.

[0182] Table 3 Average values ​​of experimental data at different rotation centers

[0183]

[0184] The experimental results show that the experimental group with the center of mass as the rotation center has a higher IoU and a smaller relative distance to the centroid. Because the data sources of the two groups of experiments are exactly the same, the higher IoU proves that the rectangular box calibrated by the rotation method with the center of mass as the center is tighter, there is less blank area in the footprint rectangular box, and the calculation result of the eigenvalue is more accurate; the smaller relative distance to the centroid proves that the footprint is more centered in the rectangular box. Figure 9a , the conclusion of the visual effect analysis of Figure 9 is the same.

[0185] Since the determined plantar pressure zoning scheme is divided according to the standard footprint position, the plantar pressure zoning results obtained using the scheme with the center of mass as the rotation center are more accurate, and the characteristic values ​​of each partition area are also more accurate. Using this scheme for footprint calibration and plantar characteristic value calculation will also produce better results.

[0186] (3) Verification of the reliability of footprint calibration

[0187] The main method of the experiment is to collect multiple groups of footprints in different states, perform rectangle calibration and correction on them, compare and analyze the evaluation indicators of each group of footprint rectangles, and verify the reliability of the footprint calibration algorithm.

[0188] A subject (age 24 years, height 170 cm, weight 85 kg) was recruited to stand or walk on the plantar pressure collection board in two different postures for 30 seconds. A total of ten groups of footprint images from the subject were collected (150-180 complete footprints in each group in the standing posture, and 100-120 complete footprints in each group in the walking posture). Rectangular rotation footprints were calibrated for each of them, and the IoU and centroid relative distance of each footprint in each group of experiments were recorded. The scatter plot was observed and the experimental results were analyzed.

[0189] Table 4 Footprint calibration accuracy of standing posture

[0190]

[0191] Table 5 Average experimental data of standing posture

[0192]

[0193] In the standing posture, the accuracy calculation of each group of experiments is shown in Table 4. The scatter diagram of IoU and centroid relative distance is shown in Table 4. Fig.12a , Figure 12b The experimental results of each group in standing posture are shown in Table 5.

[0194] Table 6 Footprint calibration accuracy of walking posture

[0195]

[0196] Table 7 Average experimental data of walking posture

[0197]

[0198] In the walking posture, the accuracy calculation of each group of experiments is as shown in Table 4. The scatter diagram of IoU and centroid relative distance is as follows: Fig.12a , Figure 12b The experimental results of each group in standing posture are shown in Table 5.

[0199] As shown in Table 6, the scatter plots of IoU and centroid relative distance are shown in Table 6. Fig.13a , Fig.13b The experimental results of each group in standing posture are shown in Table 7.

[0200] The experimental results show that the average IOU value can reach more than 90% in standing posture and more than 85% in walking posture; the accuracy of footprint calibration for both is 100%; the relative centroid distance does not exceed 0.100, and it is lower in standing posture. The 100% calibration accuracy and high IOU value prove that the footprint calibration algorithm can calibrate an accurate and compact footprint rectangle frame whether standing or walking. The low relative centroid distance proves that the footprint is centered enough in the rectangle frame, ensuring the correctness of the plantar pressure zoning and improving the accuracy of the pressure-based foot feature value extraction results.

[0201] In summary, this experiment proves that the footprint calibration algorithm has high reliability, and this algorithm has profound significance for the study of plantar pressure characteristics.

[0202] In one embodiment of the present invention, a computer-readable storage medium storing one or more programs is also provided. The one or more programs include instructions. When the instructions are executed by a computing device, the computing device executes any one of the methods in the above embodiments.

[0203] In one embodiment of the present invention, a computing device is also provided, which includes: one or more processors, a memory and one or more programs, wherein the one or more programs are stored in the memory and configured to be executed by the one or more processors, and the one or more programs include instructions for executing any method in the above embodiments.

[0204] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A pressure-based footprint calibration and feature extraction method, characterized in that: include: Convert the collected raw plantar pressure data into grayscale images and perform image preprocessing; The pre-processed image is initially calibrated with a rectangular frame to obtain the footprint coordinate information. During the calibration, the expansion algorithm and the broken foot fusion algorithm are used to obtain a complete footprint array in case of disconnected sole footprints. The footprint array obtained by the initial rectangular frame calibration is partitioned into plantar pressure zones, and the footprint area is corrected to obtain a footprint rectangular frame and new plantar pressure zone coordinates; The plantar pressure characteristic value is obtained through the corrected footprint rectangle and the new plantar pressure partition coordinates.

2. The pressure-based footprint calibration and feature extraction method according to claim 1, characterized in that: Perform the initial rectangular frame calibration on the preprocessed image to obtain the footprint coordinate information, including: The outer contour of the footprint area in the image is obtained by using the contour extraction method to obtain the footprint contour; Use the minimum rectangular box to select the footprint outline and record the coordinate information (x, y, w, h) of the rectangular box, where x and y are the coordinates of the upper left corner of the rectangle, and w and h are the width and height of the rectangle.

3. The pressure-based footprint calibration and feature extraction method according to claim 1, characterized in that: When calibrating, the expansion algorithm and the broken foot fusion algorithm are used to deal with the disconnected sole footprint, including: Before extracting the outer contour of the footprint area in the image, the footprint area in the image is expanded at least twice; Extract the outer contour of the footprint area in the dilated image, calculate the contour area size in turn, and sort the contour areas in descending order; Determine the minimum rectangular frame coordinates of the outline in turn and store them in the footprint array, and determine the comparison value w of the footprint length o ; Traverse the footprint array and compare the footprint length w in the footprint array with the comparison value w o For comparison, if w>w o , is regarded as a complete footprint and stored in the complete footprint array; if w <w o , is regarded as an incomplete footprint; Record the incomplete footprint as "half-foot 1" and traverse the footprint array again. If another half-foot appears, the y-axis coordinates of the two half-foot rectangles intersect and the x-axis coordinate difference is less than the comparison value w of the footprint length. o , then record the other half foot as "half foot 2", and merge "half foot 1" with "half foot 2" to form a complete footprint and store it in the complete footprint array.

4. The pressure-based footprint calibration and feature extraction method according to claim 1, characterized in that: Footprint area correction, including: The centroid coordinates of the heel area and the joint centroid coordinates of the metatarsal area 1 and the metatarsal area 2 are calculated from the partitioned footprint image, and the horizontal inclination angle of the straight line where the two centroids are located is calculated; Determine the rotation center, use the centroid or mass center of the footprint rectangle as the rotation center, rotate the horizontal rectangle, and adjust the foot width to determine the width of the rotated rectangle. The rotated rectangle is obtained with [centroid (x c ,y c ), width (w′), height (h′), tilt angle (angle)], the size and relative position of each partition and the overall area do not change, and the tilt angle of each partition area is consistent with the overall area.

5. The pressure-based footprint calibration and feature extraction method according to claim 4, characterized in that: When rotating with the centroid as the center, for the footprint rectangle, the centroid coordinates before and after the rotation remain unchanged; When rotating around the center of mass, the coordinates of the center of mass after rotation A(x0, y0) remain unchanged, and the coordinates of the centroid before rotation are B(x c ,y c ), the rotation angle is α, and the coordinates of the centroid after rotation are set to C(x c ′,y c ′), then the solution of C is: Establish a coordinate system XOY, with the x-axis pointing right and the y-axis pointing downward, and draw A, B, and C; Draw a horizontal auxiliary line AN through point A, and establish a polar coordinate system with A as the origin, AN as the axis, and clockwise as the positive direction; let the angle between AB and AN be θ, the angle between AC and AN be θ′, AB=AC=r, and the coordinates of points A, B, and C in the polar coordinate system are A(0,0), B(r,θ), and C(r,θ′) respectively; The coordinates of the three points AB and C obtained in the polar coordinate system are translated and converted back to the rectangular coordinate system XOY to obtain the coordinates of point C (x c ′,y c ′) is: C(r·cosθ′+x0, r·sinθ′+y0).

6. The pressure-based footprint calibration and feature extraction method according to claim 4, characterized in that: Foot width adjustment, including: Based on the extracted footprint contour shape, cut it off from the x-axis direction and take the contours in the first and second plantar regions; Record the difference between the maximum and minimum values ​​of the y-axis coordinates in the extracted contour, and record the maximum difference as the foot width standard width_standard; The rotated foot width h′=(width_standard) / cos(angle) is obtained according to the foot width standard, where angle is the inclination angle.

7. The pressure-based footprint calibration and feature extraction method according to claim 1, characterized in that: The new plantar pressure zone coordinates are obtained by determining the centroid coordinates of each zone after rotation, including: Let the rectangle of each partition be [(x ci ,y ci ), w i 、h i , angle], i represents the ith partition, and the centroid coordinates of the overall area are C(x c ′,y c ′), the width is w, the height is h, the inclination angle is angle, and the centroid coordinates of each partition area are P i (x ci ,y ci ); Create a coordinate system XOY, with the x-axis to the right and the y-axis downward, and draw C and P accordingly. i and the boundaries of each subdistrict; With point C as the origin, the center line of the footprint rectangle to the right as the x' axis, and the center line of the footprint rectangle downward as the y' axis, a rectangular coordinate system X'CY' is established. In this coordinate system, the polar coordinates of the centroid of each partition are P i (x ci ,y ci ); With point C as the origin and the center line of the footprint rectangle to the right as the polar axis r1, the first polar coordinate system CR1 is established, and the polar coordinates of the centroid of each partition in this coordinate system are calculated; With point C as the origin, parallel to the positive direction of the x-axis, a second polar coordinate system CR2 is established. The angle between the second polar coordinate system CR2 and the polar axis of the first polar coordinate system CR1 is ɑ=angle; in the CR2 coordinate system, P i Recorded as in Therefore, i Coordinate is P i (r i ,θ i +α); P i The coordinates are converted to the rectangular coordinate system XOY, then P i Coordinate is P i (r i ·cos(θ i +α)+x c ′,r i ·sin(θ i +α)+y c ′), and obtain the centroid coordinates of each partition to serve as the new plantar pressure partition coordinates.

8. A pressure-based footprint calibration and feature extraction system, characterized in that: include: A preprocessing module converts the collected raw plantar pressure data into a grayscale image and performs image preprocessing; The initial calibration module performs the initial rectangular frame calibration on the preprocessed image to obtain the footprint coordinate information. During the calibration, the expansion algorithm and the broken foot fusion algorithm are used to obtain the complete footprint array in case of disconnected sole footprints; The correction module partitions the plantar pressure of the footprint array obtained by the initial rectangular frame calibration, and performs footprint area correction to obtain the footprint rectangular frame and obtain new plantar pressure partition coordinates; The characteristic value calculation module obtains the characteristic value of plantar pressure through the corrected footprint rectangular frame and the new plantar pressure partition coordinates.

9. A computer-readable storage medium storing one or more programs, characterized in that: The one or more programs include instructions, which, when executed by a computing device, cause the computing device to perform any one of the methods of claims 1 to 7.

10. A computing device, characterized in that include: One or more processors, a memory, and one or more programs, wherein the one or more programs are stored in the memory and configured to be executed by the one or more processors, and the one or more programs include instructions for executing any one of the methods described in claims 1 to 7.

Citation Information

Patent Citations

  • Foot type parameter measurement and pressure cloud picture generation method based on sole pressure data

    CN110123330A

  • Gait analysis method based on plantar pressure distribution

    CN112347848A

  • Plantar pressure image processing method, plantar pressure image recognition method and gait analysis system

    CN112766142A

  • Footprint geometrical shape characterization method based on combination of polynomial fitting and right-angle filtering

    CN113139555A

  • Walk pattern processor

    JP1998228540A

Cited By

  • Data processing system and method based on calibration board of foot tester

    CN119791645A

  • A data processing system and method based on a podiometer calibration board

    CN119791645B