Method and device for determining length of human body trunk, electronic equipment and vehicle
By obtaining and analyzing the key points data of the human body in the cabin, and judging and correcting the length of the trunk, the problem of errors or missing trunk length caused by the human head outside the camera screen is solved, and the accuracy of judging the human body's attributes is improved.
Patent Information
- Application Number
- CN202311321838.6
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2023-10-12
- Publication Date
- 2025-04-15
AI Technical Summary
Inside the vehicle cockpit, the human head is often outside the picture collected by the camera, which makes it impossible to accurately determine the outline of the human torso, which in turn leads to errors or missing of the length of the human torso, affecting the accuracy of the judgment of human body attributes.
By obtaining the human body key point data of the target human body in the car image, we can determine whether the human head is missing. If it is met, the mapping value in the human body key point neck and hip function is used to determine the neck and hip length, and correct it according to the preset human body proportions to obtain the torso length.
It realizes that the length of the human body's torso is accurately and reliably determined when the human body's head is outside the camera screen, and improves the accuracy of judging human body attributes.
Smart Images
Figure CN119832046A_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the field of data processing technology, and in particular to a method, device, electronic equipment and vehicle for determining the length of a human body. Background Art
[0002] In the vehicle cabin environment, the human torso length is calculated as the straight-line distance from the top of the head to the hips. The human torso length is very important in practical applications. The human attributes can be judged by the human torso length, such as whether the human body is an adult or a child.
[0003] Currently, inside a vehicle cabin, the vehicle cabin camera can be used to capture images inside the cabin, and points of interest and ranges of interest can be established for the captured images to determine the contour of the human torso and thereby obtain the length of the human torso.
[0004] However, when a human body is in the third row of space in a seven-seater vehicle, due to the angle of the on-board camera and the structural design of the vehicle body, the human head is often outside the picture captured by the camera inside the vehicle, which will cause the established points of interest and range of interest to be unable to accurately determine the outline of the human torso, which will in turn cause the obtained human torso length to be incorrect or even missing, thereby affecting the accuracy of judging the human body's attributes. Summary of the invention
[0005] In view of this, the present application provides a method, device, electronic device and vehicle for determining the length of a human body. The main purpose is to improve the technical problem in the current prior art that if the head of a human body is outside the picture captured by the camera inside the vehicle, the determined length of the human body will be wrong or even missing, which will affect the accuracy of judging the attributes of the human body.
[0006] In a first aspect, the present application provides a method for determining the length of a human body trunk, comprising:
[0007] Acquire key point data of a target human body in the in-vehicle image, wherein the key point data of the human body includes key point data of a human head;
[0008] According to the key point data of the human head, determining whether the target human body meets the head missing condition;
[0009] If it is determined that the target human body meets the head missing condition, the mapping value corresponding to the human body key point data in the human body key point neck-hip function is determined, and the mapping value is determined as the human body neck-hip length corresponding to the target human body;
[0010] The human body neck and hip length is corrected according to a preset human body proportion to obtain a human body trunk length corresponding to the target human body.
[0011] In a second aspect, the present application provides a device for determining the length of a human body trunk, comprising:
[0012] An acquisition module is configured to acquire key point data of a target human body in the in-vehicle image, wherein the key point data of the human body includes key point data of a human head;
[0013] A judgment module is configured to judge whether the target human body meets the head missing condition according to the key point data of the head;
[0014] The acquisition module is further configured to determine the mapping value corresponding to the human body key point data in the human body key point neck and hip function if it is determined that the target human body meets the head missing condition, and determine the mapping value as the human body neck and hip length corresponding to the target human body;
[0015] The correction module is configured to correct the neck and hip length of the human body according to a preset human body proportion to obtain the human body torso length corresponding to the target human body.
[0016] In a third aspect, the present application provides a computer-readable storage medium having a computer program stored thereon, and when the computer program is executed by a processor, the method described in the first aspect is implemented.
[0017] In a fourth aspect, the present application provides an electronic device, comprising a storage medium, a processor, and a computer program stored on the storage medium and executable on the processor, wherein the processor implements the method described in the first aspect when executing the computer program.
[0018] In a fifth aspect, the present application provides a vehicle, comprising: the device as described in the second aspect, or the electronic device as described in the fourth aspect.
[0019] By means of the above technical scheme, the present application provides a method, device, electronic device and vehicle for determining the length of a human torso. First, the human key point data of the target human body in the in-vehicle image is obtained, and the human key point data includes the key point data of the head; then, according to the key point data of the head, whether the target human body meets the condition of missing head is determined; if the target human body is determined to meet the condition of missing head, the mapping value corresponding to the human key point data in the neck-hip function of the human key point is determined, and the mapping value is determined as the neck-hip length of the human body corresponding to the target human body; then the neck-hip length of the human body is corrected according to the preset human body proportion to obtain the trunk length of the human body corresponding to the target human body. Compared with the current prior art, the present application can accurately determine whether there is a missing head of the human body in the in-vehicle image containing the human body by judging whether the corresponding key point data of the head in the in-vehicle image meets the condition of missing head. If the head of the human body in the in-vehicle image is determined to be missing, the neck-hip length of the human body can be repaired according to the human body proportion to obtain an accurate and reliable trunk length of the human body, thereby improving the accuracy of judging human attributes.
[0020] The above description is only an overview of the technical solution of the present application. In order to more clearly understand the technical means of the present application, it can be implemented in accordance with the contents of the specification. In order to make the above and other purposes, features and advantages of the present application more obvious and easy to understand, the specific implementation methods of the present application are listed below. BRIEF DESCRIPTION OF THE DRAWINGS
[0021] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present application and, together with the description, serve to explain the principles of the present application.
[0022] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, for ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative labor.
[0023] Figure 1 A schematic flow chart of a method for determining the length of a human body provided in an embodiment of the present application is shown;
[0024] Figure 2 A schematic flow chart of a method for determining the length of a human body provided in an embodiment of the present application is shown;
[0025] Figure 3 A schematic diagram of an example structure provided by an embodiment of the present application is shown;
[0026] Figure 4 A schematic diagram of an example structure provided by an embodiment of the present application is shown;
[0027] Figure 5 A schematic structural diagram of a device for determining the length of a human body provided in an embodiment of the present application is shown. DETAILED DESCRIPTION
[0028] In order to more clearly understand the above-mentioned purposes, features and advantages of the present application, the scheme of the present application will be further described below. It should be noted that the embodiments of the present application and the features in the embodiments can be combined with each other without conflict.
[0029] In order to improve the technical problem that if the head of a person is outside the image captured by the camera in the car, the determined length of the person's torso will be wrong or even missing, which will affect the accuracy of judging the attributes of the person. This embodiment provides a method for determining the length of a person's torso, such as Figure 1 As shown, the method includes:
[0030] Step 101: Acquire key point data of a target person in an in-vehicle image.
[0031] Among them, the human body key point data includes the human head key point data.
[0032] In the embodiment of the present application, the in-vehicle image may be an image captured by a camera in the vehicle cabin. In the vehicle cabin, the camera may capture a large number of images.
[0033] Optionally, the vehicle cabin is the interior space of the car, a compartment for the driver and passengers.
[0034] In the embodiments of the present application, the key points of the human body refer to the key position points of various parts of the human body. The key points of the human body are crucial for describing human posture and predicting human behavior. The key points of the human body mainly include three points of the left and right arms (wrist, elbow, shoulder), three points of the left and right legs (ankle, knee, hip), hip and buttocks points, and head points (ear, chin, top of the head), etc.
[0035] Optionally, the head key points are four key points of the head part included in the human body key points, namely, the head vertex, the left ear point, the right ear point and the jaw point.
[0036] In an embodiment of the present application, the in-vehicle image may be a large number of images collected by a camera in the vehicle cabin. The target human body is the human body in the three rows of space in the vehicle. The images collected by the camera may include images of the driver, images of passengers, images of the environment in the cabin, images of the external environment in which the vehicle is currently located, etc. In an embodiment of the present application, it is necessary to obtain images containing human bodies in the third row of space in the vehicle and to obtain key point data of the human bodies in the images.
[0037] Step 102: determine whether the target human body meets the head missing condition based on the key point data of the human head.
[0038] Optionally, the head key point data may include coordinate data of the head key point and confidence scores corresponding to the coordinate data.
[0039] In the embodiment of the present application, the head missing condition may be preset conditions such as unreliable head data acquired based on the image, incomplete head information, etc.
[0040] Step 103: If it is determined that the target human body meets the head missing condition, the mapping value corresponding to the human body key point data in the human body key point neck-hip function is determined, and the mapping value is determined as the human body neck-hip length corresponding to the target human body.
[0041] In the embodiment of the present application, the human body key point neck-hip mapping function is a function that maps the coordinates of the human body key points to the neck-hip length.
[0042] Optionally, if it is determined that the target human body in the image meets the head missing condition, the mapping value of the human body key point coordinates in the human body key point neck-hip mapping function is used as the neck-hip length of the target human body.
[0043] Step 104: Correct the neck and hip length of the human body according to the preset human body proportions to obtain the human body torso length corresponding to the target human body.
[0044] In the embodiment of the present application, the human body neck-hip length is the straight-line distance between the human body neck and the human body hip; the human body trunk length is the straight-line distance between the human body head and the human body hip.
[0045] Compared with the current existing technology, this embodiment can better judge whether the corresponding head key point data in the in-vehicle image meets the head missing condition, whether the human head information exists and is reliable in the image containing the human body in the third row space of the vehicle; if the human head information in the image is judged to be non-existent or unreliable, the human neck and hip length can be repaired to obtain an accurate and reliable human torso length, thereby improving the accuracy of judging the human attributes.
[0046] Further, as a refinement and extension of the above embodiment, in order to fully illustrate the specific implementation process of the method of the embodiment of the present disclosure, the embodiment of the present application provides the following Figure 4 The specific method shown includes:
[0047] Step 201: Input the in-vehicle image into a first recognition model for calculation to obtain the human body external bounding box data of the target human body.
[0048] The first recognition model is obtained by training with a first training set, and the first training set includes different sample images and coordinate information of human body bounding boxes corresponding to the respective marks.
[0049] In order to achieve specific functions inside the vehicle cabin, it is usually necessary to train a network for recognizing the human body's external frame and a network for recognizing the human body's key points, and apply them. Figure 3 As shown, the first recognition model can be a deep learning network built for detection and classification scenarios. It is usually composed of a backbone network, a shoulder network, and a head network, which is used to detect whether there is a human body or a human head and return the coordinates of the human body's bounding box and the corresponding confidence score, the coordinates of the head's bounding box and the corresponding confidence score.
[0050] In some embodiments, the first training set also includes different sample images corresponding to the marked human head external frame coordinate information. Accordingly, step 201 may specifically include: inputting the in-vehicle image into the first recognition model for calculation to obtain the human body external frame data and the human head external frame data of the target human body;
[0051] In the embodiment of the present application, the name of the first recognition model may be a human body external frame detection model, a human body external frame recognition model, a human body external frame detection network, a human body external frame recognition network, etc., and the name of the first recognition model is not specifically limited here. The first training set may include different sample images, human body external frame coordinate information corresponding to the sample images, and head external frame coordinate information corresponding to the sample images.
[0052] Optionally, the process of inputting the in-vehicle image into the first recognition model for calculation can be: inputting the image into the first recognition model, and matching and calculating with the human body external frame data marked by the sample images in the first training set and the head external frame data marked by the sample images in the first training set, to obtain the human body external frame data and the head external frame data corresponding to the current image.
[0053] In an embodiment of the present application, the human body external frame data may include coordinate information of the human body external frame and the confidence score corresponding to the coordinate information. Correspondingly, the head external frame data may include coordinate information of the head external frame and the confidence score corresponding to the coordinate information.
[0054] For example, for picture A captured by the vehicle camera, picture A includes a human body a in the third row space of the vehicle. By inputting picture A into the first recognition model for calculation, the human body external frame coordinates and the corresponding confidence score corresponding to human body a, and the head external frame coordinates and the corresponding confidence score can be obtained.
[0055] Step 202: input the human body external bounding box data into the second recognition model for calculation to obtain the human body key point data of the target human body.
[0056] The second recognition model is obtained by training with a second training set, and the second training set includes human body bounding box data of different sample images and coordinate information of human body key points corresponding to the marks.
[0057] like Figure 4 As shown, the second recognition model can be a deep learning network built for the detection scenario. It is usually composed of a backbone network, a shoulder network, and a head network, which is used to detect whether there are key points of the human body and return the coordinates of the key points of the human body and the corresponding confidence scores.
[0058] In the embodiment of the present application, the name of the second recognition model can be a human key point detection model, a human key point recognition model, a human key point detection network, a human key point recognition network, etc., and the name of the second recognition model is not specifically limited here. The second training set can include different sample images and the human key point coordinate information corresponding to the sample images.
[0059] Optionally, the process of inputting the in-vehicle image into the second recognition model for calculation can be to input the image and the external frame data of the human body marked in the image into the second recognition model, and perform matching calculations with the human body key point data marked in the sample image in the first training set to obtain the human body key point data corresponding to the current image.
[0060] In an embodiment of the present application, the human body key point data may include coordinate information of the human body key points and confidence scores corresponding to the coordinate information. Correspondingly, the human head key point data may be the four key points of the human head included in the human body key point data, namely, the head vertex, left ear point, right ear point and jaw point.
[0061] Exemplarily, based on step 201, for picture A captured by the vehicle camera, picture A includes human body a in the third row space of the vehicle, and picture A is input into the second recognition model for calculation, the human body key point coordinates corresponding to human body a and the corresponding confidence score can be obtained, and the head key point coordinates and the corresponding confidence score can be determined from the human body key point coordinates and the corresponding confidence score.
[0062] Step 203: determine whether the target human body meets the head missing condition based on the key point data of the human head.
[0063] In some embodiments, step 203 may specifically include: judging whether the target human body meets the head missing condition based on the head key point data and / or the head external frame data.
[0064] In some embodiments, the head key point data includes the head key point coordinate information of the target body and the corresponding first confidence score, and the head external frame data includes the head external frame coordinate information of the target body and the corresponding second confidence score. Accordingly, step 203 specifically also includes: judging whether the target body meets the head missing condition based on the first confidence score and / or the second confidence score.
[0065] In some embodiments, step 203 specifically further includes: if the first confidence score is less than a first preset threshold, and / or the second confidence score is less than a second preset threshold, determining that the target human body meets the head missing condition.
[0066] In the embodiment of the present application, the first confidence score is the sum of the confidence scores corresponding to the head vertex, left ear point, right ear point and chin point in the key points of the head. Accordingly, the first preset threshold is one of the criteria for determining whether the head is reliable or exists. The calculation method of the first confidence score can be specifically shown in the following formula 1:
[0067] y=a+b+c+d (Formula 1)
[0068] In Formula 1, y represents the first confidence score, and a, b, c, and d represent the confidence scores corresponding to the top of the head, the left ear, the right ear, and the chin, respectively.
[0069] Optionally, the second confidence score is a confidence score corresponding to the coordinates of the circumscribed frame of the head, and accordingly, the second preset threshold is one of the criteria for determining whether the head is reliable or exists.
[0070] Exemplarily, based on step 202, based on the coordinates of the human body key points corresponding to human body a and the corresponding confidence scores, the coordinates of the head key points corresponding to human body a and the corresponding confidence scores are determined, the confidence scores of the head key points are added, and the added result is compared with the preset first threshold. If the added result is less than the preset first threshold, it is judged that the human body a in picture A meets the missing head condition. If the added result is greater than or equal to the preset first threshold, it is judged that the human body a in picture A does not meet the missing head condition.
[0071] Correspondingly, based on step 202, based on the confidence score corresponding to the coordinates of the external frame of the head of human body a, the size of the confidence score corresponding to the coordinates of the external frame of the head and the second threshold is judged. If the confidence score corresponding to the coordinates of the external frame of the head is less than the second threshold, it is judged that the human body a in the picture A meets the missing head condition. If the confidence score corresponding to the coordinates of the external frame of the head is greater than or equal to the second threshold, it is judged that the human body a in the picture A does not meet the missing head condition.
[0072] Step 204: If it is determined that the target human body meets the head missing condition, the mapping value corresponding to the human body key point data in the human body key point neck-hip function is determined, and the mapping value is determined as the human body neck-hip length corresponding to the target human body.
[0073] Exemplarily, based on step 203, if person a in picture A meets the head missing condition, the mapping value of the key point data of person a in the key point neck-hip mapping function is obtained as the neck-hip length of person a.
[0074] Correspondingly, based on step 203, if person a in picture A does not meet the head missing condition, the mapping value of the human body key point data of person a in the human body key point torso mapping function is obtained as the human body torso length of person a.
[0075] The human body key point torso length mapping function may be a function that maps the key point coordinates obtained by the human body key point detection network to the torso length.
[0076] Step 205: Correct the neck and hip length of the human body according to the preset human body proportions to obtain the human body torso length corresponding to the target human body.
[0077] In some embodiments, step 205 may specifically include: multiplying the neck and hip length of the human body by a restoration coefficient corresponding to a preset human body proportion to obtain a human body torso length corresponding to the target human body.
[0078] In the embodiment of the present application, the torso length repair coefficient is a ratio value of mapping the neck and hip length to the torso length. The neck and hip length of the human body is corrected to the torso length of the human body, which can be shown in the following formula 2:
[0079] L2 = L1 * r (Formula 2)
[0080] In formula 2, L1 represents the length of the human neck and hips, L2 represents the length of the human trunk, and r represents the trunk length repair coefficient.
[0081] Exemplarily, based on step 204, if the person a in the image A meets the head missing condition, the neck and hip length of the person a is multiplied by the torso length repair coefficient to obtain the torso length of the person a.
[0082] In some embodiments, after step 205, the method of this embodiment further includes: determining the human attribute information corresponding to the target human body by comparing the human body torso length corresponding to the target human body with the standard value of the torso length.
[0083] In some embodiments, after step 205, the method of this embodiment specifically further includes: if the human body trunk length is less than the standard value of the child's trunk length, determining that the target human body does not belong to an adult body; if the human body trunk length is greater than the standard value of the adult's trunk length, determining that the target human body does not belong to a child's body; if the human body trunk length is greater than or equal to the standard value of the child's trunk length, and the human body trunk length is less than or equal to the standard value of the adult's trunk length, then calculating the probability that the target human body belongs to an adult body according to a linear function based on the human body trunk length, and determining whether the target human body belongs to an adult body based on the probability.
[0084] Optionally, the human body attributes are judged to distinguish whether the human body is an adult or a child.
[0085] In the embodiment of the present application, the torso length standard may include a child torso length standard value and an adult torso length standard value; the adult torso length standard value may determine that when the torso length is greater than the value, the probability that the person corresponding to the torso belongs to an adult is 1, and the probability that the person belongs to a child is 0. The child torso length standard value may determine that when the torso length is less than the value, the probability that the person corresponding to the torso belongs to a child is 1, and the probability that the person belongs to an adult is 0.
[0086] Optionally, if the human torso length is greater than or equal to the standard value of the child torso length, and the human torso length is less than or equal to the standard value of the adult torso length, the probability that the target human body belongs to an adult human body can be determined by Formula 3, and Formula 3 can be specifically as follows:
[0087] p = (L - t) / (s - t) (Formula 3)
[0088] In Formula 3, p represents the probability that the target human body belongs to an adult human body, L represents the torso length of the target human body, t represents the standard value of the torso length of a child, and s represents the standard value of the torso length of an adult.
[0089] Optionally, if the probability that the target human body belongs to an adult human body is p, then the probability that the target human body belongs to an adult human body is 1-p.
[0090] Compared with the current existing technologies, in this embodiment, by using the first preset threshold for judging the comprehensive score of the confidence of human head key points and the second preset threshold for judging the confidence score of the bounding box of the human head, it is possible to judge whether the corresponding human head key point data in the in-vehicle image meets the condition of human head missing, and it is possible to better judge whether the head information of the human body exists and is reliable in the image containing the human body in the third row space of the vehicle, improve the accuracy of the judgment result, and increase the credibility of the judgment result; if it is judged that the head information of the human body in the image does not exist or is unreliable, based on the situation that when the head of the human body is outside the captured picture of the camera, the neck information is still clearly visible, and there is a relatively stable ratio between the distance from the neck to the crotch and the distance from the top of the head to the crotch, the neck-crotch length can be used to estimate or correct the torso length, so as to obtain an accurate and reliable human torso length, and further improve the correct rate of human body attribute judgment.
[0091] Further, in order to fully illustrate the specific implementation process of the method in this embodiment, the following examples are provided in this embodiment, but are not limited thereto.
[0092] For a certain human body in the third row space in picture B, record the confidence score of the corresponding human head as x. Record the overall result obtained by the human body key point detection network as n, and the confidence scores of the top head point, left ear point, right ear point and jaw point obtained are a, b, c, d respectively; record the human body key point neck-crotch length mapping function as F1, and the human body key point torso length mapping function as F2; record the human head confidence score threshold as Th; the comprehensive score threshold of the human head key point confidence as Tk; record the torso length repair coefficient as r; record the standard value of the adult torso length as s, and the standard value of the child torso length as t; record the probability that the human body belongs to an adult as p, and the probability of belonging to a child as 1 - p; calculate the neck-crotch length L1 = F1(n), and the original torso length L2 = F2(n); calculate the comprehensive score of the human head key point confidence: y = a + b + c + d.
[0093] Judge whether the current human head is reliable or exists: when x >= Th and y >= Tk: the current human head exists and is reliable, and directly use L2 as the torso length (L = L2). When x < Th or y < Tk: the current human head does not exist or is unreliable, and use the corrected torso length L3 = L1 * r (L = L3).
[0094] Judge the probability that the human body belongs to an adult according to the current torso length: if L < t, at this time the torso length is less than the standard value of the child torso length and it is impossible to be an adult. p = 0
[0095] If L > s, at this time the torso length is greater than the standard value of the adult torso length and it is impossible to be a child. p = 1
[0096] If L>=t and L<=s, the torso length is between the standard value of child torso length and the standard value of adult torso length, and it cannot be completely determined whether it is an adult or a child. The probability that the person belongs to an adult needs to be calculated according to the current torso length according to a linear function: p=(Lt) / (st).
[0097] Compared with the current existing technology, this embodiment uses a first preset threshold for judging the comprehensive confidence score of the key points of the head and a second preset threshold for judging the confidence score of the external frame of the head to judge whether the corresponding key point data of the head in the in-vehicle image meets the head missing condition. It can better judge whether the head information of the human body in the image containing the human body in the third row of the vehicle is present and reliable, thereby improving the accuracy of the judgment result and increasing the credibility of the judgment result. If the head information of the human body in the image is judged to be non-existent or unreliable, the neck information can be clearly visible when the head of the human body is outside the picture captured by the camera, and the distance from the neck to the hips and the distance from the top of the head to the hips have a relatively stable ratio. The neck and hip length can be used to infer or correct the torso length, thereby obtaining an accurate and reliable torso length of the human body, thereby improving the accuracy of judging the attributes of the human body.
[0098] Further, as Figure 1 and Figure 4 The specific implementation of the method shown in the figure, this embodiment provides a device for determining the length of a human body trunk, such as Figure 5 As shown, the device includes: an acquisition module 31, a judgment module 32, and a correction module 33.
[0099] An acquisition module 31 is configured to acquire key point data of a target human body in the in-vehicle image, wherein the key point data of the human body includes key point data of a human head;
[0100] A judgment module 32 is configured to judge whether the target human body meets the head missing condition according to the key point data of the head;
[0101] The acquisition module 31 is further configured to determine the mapping value corresponding to the human body key point data in the human body key point neck and hip function if it is determined that the target human body meets the head missing condition, and determine the mapping value as the human body neck and hip length corresponding to the target human body;
[0102] The correction module 33 is configured to correct the neck-hip length of the human body according to a preset human body proportion to obtain a human body torso length corresponding to the target human body.
[0103] In a specific application scenario, the acquisition module 31 is specifically configured to input the in-vehicle image into a first recognition model for calculation to obtain the external bounding box data of the target human body, wherein the first recognition model is obtained by training with a first training set, and the first training set includes different sample images and corresponding marked external bounding box coordinate information of the human body; input the external bounding box data of the human body into a second recognition model for calculation to obtain the human key point data of the target human body, wherein the second recognition model is obtained by training with a second training set, and the second training set includes the external bounding box data of different sample images and corresponding marked human key point coordinate information.
[0104] In a specific application scenario, the first training set also includes different sample images corresponding to the marked head external frame coordinate information; accordingly, the acquisition module 31 is specifically configured to input the in-vehicle image into the first recognition model for calculation to obtain the body external frame data and head external frame data of the target body; the judgment module 32 is specifically configured to judge whether the target body meets the head missing condition based on the head key point data and / or the head external frame data.
[0105] In a specific application scenario, the head key point data includes the head key point coordinate information of the target body and the corresponding first confidence score, and the head external frame data includes the head external frame coordinate information of the target body and the corresponding second confidence score; accordingly, the judgment module 32 is also specifically configured to determine that the target body meets the head missing condition if the first confidence score is less than a first preset threshold and / or the second confidence score is less than a second preset threshold.
[0106] In a specific application scenario, the correction module 33 is further configured to determine the human attribute information corresponding to the target human body by comparing the human body torso length corresponding to the target human body with the standard value of the torso length.
[0107] In a specific application scenario, the correction module 33 is further configured to determine that the target human body does not belong to an adult body if the human body trunk length is less than the standard value of the child's trunk length; determine that the target human body does not belong to a child body if the human body trunk length is greater than the standard value of the adult's trunk length; if the human body trunk length is greater than or equal to the standard value of the child's trunk length and the human body trunk length is less than or equal to the standard value of the adult's trunk length, calculate the probability that the target human body belongs to an adult body according to a linear function based on the human body trunk length, and determine whether the target human body belongs to an adult body based on the probability.
[0108] It should be noted that for other corresponding descriptions of the functional units involved in the apparatus for determining the length of a human body trunk provided in this embodiment, reference can be made to Figure 1 and Figure 2 The corresponding description in will not be repeated here.
[0109] Based on the above Figure 1 and Figure 2 The method shown in the figure, accordingly, the embodiment of the present disclosure also provides a computer-readable storage medium on which a computer program is stored, and when the computer program is executed by a processor, the above-mentioned Figure 1 and Figure 2 The method shown.
[0110] Based on this understanding, the technical solution of the present disclosure can be embodied in the form of a software product, which can be stored in a non-volatile storage medium (which can be a CD-ROM, a USB flash drive, a mobile hard disk, etc.), and includes a number of instructions for enabling a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods of various implementation scenarios of the present disclosure.
[0111] Based on the above Figure 1 and Figure 2 The method shown, and Figure 5 In order to achieve the above-mentioned purpose, the embodiment of the present disclosure also provides an electronic device, which can be configured on a computer terminal, etc. The device includes a storage medium and a processor; the storage medium is used to store a computer program; the processor is used to execute the computer program to achieve the above-mentioned Figure 1 and Figure 2 The method shown.
[0112] In some embodiments, the above-mentioned physical device may also include a user interface, a network interface, a camera, a radio frequency (RF) circuit, a sensor, an audio circuit, a WI-FI module, etc. The user interface may include a display, an input unit such as a keyboard, etc., and the optional user interface may also include a USB interface, a card reader interface, etc. The network interface may include a standard wired interface, a wireless interface (such as a WI-FI interface), etc. in some embodiments.
[0113] Those skilled in the art will appreciate that the above-mentioned physical device structure provided in the embodiments of the present disclosure does not constitute a limitation on the physical device, and may include more or fewer components, or a combination of certain components, or different arrangements of components.
[0114] The storage medium may also include an operating system and a network communication module. The operating system is a program that manages the hardware and software resources of the above-mentioned physical device, and supports the operation of the information processing program and other software and / or programs. The network communication module is used to realize the communication between the components inside the storage medium, and the communication with other hardware and software in the information processing physical device.
[0115] Based on the above electronic device, the embodiment of the present disclosure further provides a vehicle, which may specifically include: Figure 5 The device shown in the figure or the electronic device as described above. The vehicle can be a new energy vehicle or a traditional vehicle.
[0116] Through the description of the above disclosed implementation methods, the technical personnel in the field can clearly understand that the present disclosure can be implemented by means of software plus the necessary general hardware platform, or by hardware. By applying the scheme of the embodiment of the present disclosure, compared with the current prior art, the present application uses the first preset threshold for judging the comprehensive score of the confidence of the key points of the head and the second preset threshold for judging the confidence score of the external frame of the head to judge whether the corresponding key point data of the head in the in-vehicle image meets the head missing condition, which can better judge whether the head information of the human body in the image containing the human body in the third row of the vehicle exists and is reliable, improve the accuracy of the judgment result, and increase the credibility of the judgment result; if the human head information in the judgment image does not exist or is unreliable, the neck information of the human head can be clearly visible when the human head is outside the picture captured by the camera, and the distance from the neck to the hips and the distance from the top of the head to the hips have a relatively stable ratio. The neck and hip length is used to infer or correct the torso length, thereby obtaining an accurate and reliable human torso length, thereby improving the accuracy of the judgment of human attributes.
[0117] It should be noted that, in this article, relational terms such as "first" and "second" are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the term "comprising" or any other variant thereof is intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the sentence "comprising a ..." do not exclude the existence of other identical elements in the process, method, article or device including the elements.
[0118] The above description is only a specific embodiment of the present disclosure, so that those skilled in the art can understand or implement the present disclosure. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present disclosure. Therefore, the present disclosure will not be limited to the embodiments described herein, but will conform to the widest scope consistent with the principles and novel features applied herein.
Claims
1. A method for determining the length of a human body trunk, characterized in that: include: Acquire key point data of a target human body in the in-vehicle image, wherein the key point data of the human body includes key point data of a human head; According to the key point data of the human head, determining whether the target human body meets the head missing condition; If it is determined that the target human body meets the head missing condition, the mapping value corresponding to the human body key point data in the human body key point neck-hip function is determined, and the mapping value is determined as the human body neck-hip length corresponding to the target human body; The human body neck and hip length is corrected according to a preset human body proportion to obtain a human body trunk length corresponding to the target human body.
2. The method according to claim 1, characterized in that The step of obtaining key point data of a target human body in the in-vehicle image includes: Input the in-vehicle image into a first recognition model for calculation to obtain the human body bounding box data of the target human body, wherein the first recognition model is obtained by training with a first training set, and the first training set includes different sample images and human body bounding box coordinate information corresponding to the respective marks; The human body bounding box data is input into a second recognition model for calculation to obtain human body key point data of the target human body, wherein the second recognition model is obtained by training with a second training set, and the second training set includes human body bounding box data of different sample images and coordinate information of human body key points corresponding to the marks.
3. The method according to claim 2, characterized in that The first training set also includes coordinate information of the external bounding boxes of human heads corresponding to different sample images; The step of inputting the in-vehicle image into a first recognition model for calculation to obtain the human body external bounding box data of the target human body includes: Inputting the in-vehicle image into a first recognition model for calculation to obtain the human body external frame data and the human head external frame data of the target human body; The step of judging whether the target human body satisfies a head missing condition according to the head key point data includes: Whether the target human body satisfies a head missing condition is determined based on the head key point data and / or the head external frame data.
4. The method according to claim 3, characterized in that The head key point data includes the head key point coordinate information of the target human body and the corresponding first confidence score, and the head external frame data includes the head external frame coordinate information of the target human body and the corresponding second confidence score; The determining, based on the head key point data and / or the head external frame data, whether the target human body satisfies the head missing condition includes: If the first confidence score is less than a first preset threshold, and / or the second confidence score is less than a second preset threshold, it is determined that the target human body meets the head missing condition.
5. The method according to any one of claims 1 to 4, characterized in that After the neck-hip length of the human body is corrected according to the preset human body proportion to obtain the human body torso length corresponding to the target human body, the method further includes: The human body attribute information corresponding to the target human body is determined by comparing the human body trunk length corresponding to the target human body with the standard value of the trunk length.
6. The method according to claim 5, characterized in that The determining of the human attribute information corresponding to the target human body by comparing the human body torso length corresponding to the target human body with the torso length standard value includes: If the torso length of the human body is less than the standard value of the torso length of a child, it is determined that the target human body is not an adult human body; If the torso length of the human body is greater than the standard value of the torso length of an adult, it is determined that the target human body is not a child body; If the human body trunk length is greater than or equal to the standard value of the child's trunk length and the human body trunk length is less than or equal to the standard value of the adult's trunk length, the probability that the target human body belongs to an adult body is calculated according to a linear function based on the human body trunk length, and whether the target human body belongs to an adult body is determined based on the probability.
7. A device for determining the length of a human body trunk, characterized in that: The device comprises: An acquisition module is configured to acquire key point data of a target human body in the in-vehicle image, wherein the key point data of the human body includes key point data of a human head; A judgment module is configured to judge whether the target human body meets the head missing condition according to the key point data of the head; The acquisition module is further configured to determine the mapping value corresponding to the human body key point data in the human body key point neck and hip function if it is determined that the target human body meets the head missing condition, and determine the mapping value as the human body neck and hip length corresponding to the target human body; The correction module is configured to correct the neck and hip length of the human body according to a preset human body proportion to obtain the human body torso length corresponding to the target human body.
8. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the method according to any one of claims 1 to 6 is implemented.
9. An electronic device comprising a storage medium, a processor, and a computer program stored in the storage medium and executable on the processor, characterized in that: When the processor executes the computer program, the method according to any one of claims 1 to 6 is implemented.
10. A vehicle, characterized in that: include: The device as claimed in claim 7 or the electronic device as claimed in claim 9.