Vehicle passenger height detection method and device
The body feature size and percentile of the occupant are obtained through the detection equipment, and the relationship function is used to inversely deduce height, which solves the problem of inaccurate height detection of occupants in the vehicle in the prior art, and improves the accuracy and reliability of the child monitoring system.
Patent Information
- Application Number
- CN202411190741.0
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-28
- Publication Date
- 2024-12-06
AI Technical Summary
The prior art cannot accurately detect the height of occupants in the vehicle, resulting in limited accuracy of the child monitoring system.
By obtaining the detection information of the detection equipment in the car, the body characteristic size of the occupant is determined, and the characteristic percentile is calculated. The occupant's height is reversely pushed by the function of the relationship between the body characteristics and the percentile, thereby achieving more accurate height calculation.
Improve the accuracy and robustness of occupant height detection to ensure the effective start of the child monitoring system.
Smart Images

Figure CN119090936A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of vehicle-related technologies, and in particular to a vehicle occupant height detection method, electronic equipment, storage medium and computer program product. Background Art
[0002] The existing vehicle technology, such as the Occupant Monitor System (OMS), provides monitoring of children in the vehicle. The monitoring is activated by judging the age of the occupant based on the height of the occupant. When the occupant's height is less than a preset height threshold, the occupant is judged as a child, and the monitoring is activated.
[0003] However, in a car, since the human body is in a sitting position, it is actually difficult for the in-car detection equipment to detect the height of the occupant. Therefore, the existing technology can only estimate the height of the occupant through empirical values based on the photographed photos. Summary of the invention
[0004] Based on this, it is necessary to provide a vehicle occupant height detection method, electronic equipment, storage medium and computer program product to address the technical problem that the existing technology cannot accurately determine the height of the occupants in the vehicle.
[0005] The present invention provides a vehicle occupant height detection method, comprising:
[0006] Obtain detection information from in-vehicle detection equipment;
[0007] determining, based on the detection information, body feature dimensions of one or more body features of the vehicle occupant as real-time body feature dimensions;
[0008] calculating a characteristic percentile of each of the real-time body characteristic dimensions;
[0009] Based on one or more of the characteristic percentiles, an occupant height is calculated.
[0010] Further:
[0011] The calculating the percentile of each real-time body feature size includes: for each real-time body feature size, obtaining a corresponding body feature size and percentile relationship function, substituting the real-time body feature size into the body feature size and feature percentile relationship function, and obtaining the feature percentile corresponding to the real-time body feature size;
[0012] The calculating the occupant's height based on one or more of the characteristic percentiles includes: calculating a height percentile according to one or more of the characteristic percentiles, substituting the height percentile into a function of the relationship between the occupant's height and the height percentile corresponding to the occupant's height, and obtaining the occupant's height corresponding to the height percentile.
[0013] Furthermore, it also includes:
[0014] For each body feature, obtaining multiple sets of body feature training values, the body feature training values including training body feature sizes and corresponding training feature percentiles;
[0015] For the height of the occupant, multiple sets of height training values are obtained, wherein the height training values include the training height and the corresponding training height percentiles.
[0016] Substitute multiple sets of body feature training values of each body feature into the body feature size and feature percentile relationship function corresponding to the body feature, and obtain the body feature size and feature percentile relationship function corresponding to the body feature by fitting. The body feature size and feature percentile relationship function corresponding to the body feature is:
[0017] Among them, L t is the physical feature size of physical feature t, P(L t ) is the characteristic body size L t The feature percentage, N t is the first direction proportional transformation coefficient of body feature t, K t is the second direction proportional transformation coefficient of body feature t, A t is the first direction translation coefficient of body feature t, B t is the second direction translation coefficient of body feature t;
[0018] Substitute multiple sets of height training values into the height and height percentile relationship function respectively, and obtain the height and percentile relationship function by fitting. The height and percentile relationship function is:
[0019] Among them, L h is height, P(L h ) is the characteristic percentage of height, N h is the first direction proportional transformation coefficient of height, K h is the second direction proportional transformation coefficient of height, A h is the first direction translation coefficient of height, B h is the second direction translation coefficient of height.
[0020] Furthermore, the number of the characteristic percentiles is greater than or equal to three, and the calculating of the height percentile according to the plurality of the characteristic percentiles includes:
[0021] Taking the confidence of each of the real-time body feature dimensions as the confidence of the corresponding feature percentile;
[0022] The confidences of the plurality of characteristic percentiles are calculated in pairs to obtain relative differences of the plurality of confidences. If all the relative differences are less than or equal to the confidence threshold, the average of the plurality of characteristic percentiles is taken as the height percentile, and the relative difference is the relative difference of the two confidences divided by the maximum value of the two confidences; otherwise
[0023] Based on the confidence of the characteristic percentile, the height percentile is calculated based on multiple of the characteristic percentiles.
[0024] Further, the confidence based on the characteristic percentile, calculating the height percentile according to a plurality of the characteristic percentiles, comprises:
[0025] Sort by confidence level from highest to lowest;
[0026] Calculate the first relative difference between the highest confidence and the second highest confidence. If the first relative difference is less than or equal to the confidence threshold, take the average of the feature percentile corresponding to the highest confidence and the feature percentile corresponding to the second highest confidence as the height percentile; otherwise
[0027] Calculate the second relative difference between the second highest confidence and the third highest confidence. If the second relative difference is less than or equal to the confidence threshold, take the average of the feature percentile corresponding to the second highest confidence and the feature percentile corresponding to the third highest confidence as the height percentile; otherwise
[0028] Calculate the relative difference between two adjacent confidence levels, and take the average of the two characteristic percentiles corresponding to the two confidence levels with the smallest relative difference as the height percentile.
[0029] Further, the detection device is a camera device, and determining the physical feature size of one or more physical features of the vehicle occupant as the real-time physical feature size based on the detection information includes:
[0030] Obtaining coordinates of multiple human skeleton key points of the vehicle occupant and the confidence level of each human skeleton key point from the detection information;
[0031] Calculating the body feature size of one or more body features as the real-time body feature size according to the coordinates of the human skeleton key points;
[0032] The confidence of the real-time body feature size is calculated based on the confidence of the human skeleton key points used to calculate the real-time body feature size.
[0033] Furthermore, the body features include: upper arms, forearms and shoulders, and the body feature dimensions include: upper arm length, forearm length and shoulder width.
[0034] The present invention provides an electronic device, comprising:
[0035] at least one processor; and,
[0036] a memory communicatively connected to at least one of the processors; wherein,
[0037] The memory stores instructions that can be executed by at least one of the processors. The instructions are executed by at least one of the processors to enable the at least one of the processors to perform the vehicle occupant height detection method as described above.
[0038] The present invention provides a storage medium, wherein the storage medium stores computer instructions. When a computer executes the computer instructions, the storage medium is used to execute all steps of the vehicle occupant height detection method as described above.
[0039] The present invention provides a computer program product, comprising a computer program / instruction, which implements the vehicle occupant height detection method as described above when the computer program / instruction is executed by a processor.
[0040] The present invention determines the body characteristic dimensions of the occupants in the vehicle based on the detection information of the in-vehicle detection equipment, and determines the height of the occupants based on the body characteristic dimensions. The height of the occupants can be inferred by selecting body characteristics that are easy to detect, so that the calculation of the height of the occupants is more accurate. BRIEF DESCRIPTION OF THE DRAWINGS
[0041] Figure 1 This is a working flow chart of a vehicle occupant height detection method according to an embodiment of the present invention;
[0042] Figure 2 This is a schematic diagram of the installation of an in-vehicle detection device according to an example of the present invention;
[0043] Figure 3 Determine schematic diagrams for key points of human skeleton;
[0044] Figure 4 Human dimensions for minors;
[0045] Figure 5 This is a flowchart of a vehicle occupant height detection method according to another embodiment of the present invention;
[0046] Figure 6 This is a schematic diagram of calculating the characteristic dimensions of a body according to an example of the present invention;
[0047] Figure 7 It is a schematic diagram of discrete data of multiple physical characteristics and percentiles, height and percentiles;
[0048] Figure 8 This is a schematic diagram of the function fitting of the relationship between height and percentile in one example of the present invention;
[0049] Fig. 9 The figure is a fitting curve of a function of the relationship between height and percentile in one example of the present invention;
[0050] Fig.10 A flowchart of a vehicle occupant height detection method according to a preferred embodiment of the present invention;
[0051] Fig.11 A schematic diagram of the activation and judgment logic of the rear child monitoring function in the best embodiment of the present invention;
[0052] Fig.12 The figure is a schematic diagram of the hardware structure of an electronic device of the present invention. DETAILED DESCRIPTION
[0053] The specific embodiments of the present invention are further described below in conjunction with the accompanying drawings. The same components are represented by the same reference numerals. It should be noted that the words "front", "rear", "left", "right", "upper" and "lower" used in the following description refer to directions in the accompanying drawings, and the words "inner" and "outer" refer to directions toward or away from the geometric center of a specific component, respectively.
[0054] like Figure 1 The figure shows a working flow chart of a vehicle occupant height detection method according to an embodiment of the present invention, comprising:
[0055] Step S101, obtaining detection information of an in-vehicle detection device;
[0056] Step S102, determining the physical feature size of one or more physical features of the vehicle occupant as the real-time physical feature size according to the detection information;
[0057] Step S103, calculating the feature percentile of each of the real-time body feature dimensions;
[0058] Step S104, calculating the height of the occupant based on one or more of the characteristic percentiles.
[0059] Specifically, the present invention can be applied to electronic devices with processing capabilities, such as a controller of a vehicle, such as an electronic control unit (ECU) of a vehicle.
[0060] First, step S101 is executed to obtain detection information of the in-vehicle detection equipment.
[0061] The in-vehicle detection equipment is used to detect passengers in the vehicle, including but not limited to camera devices, infrared devices, etc.
[0062] Then, step S102 is executed to determine the physical feature size of one or more physical features of the vehicle occupant as the real-time physical feature size based on the detection information.
[0063] The body feature is a part of the body. The body feature includes but is not limited to the upper arm, forearm, shoulder, etc. The body feature size is the length of the body feature. The body feature size includes but is not limited to the upper arm length, forearm length and shoulder width.
[0064] You can choose a detection device that can easily detect body features with accurate information.
[0065] like Figure 2 FIG. 1 is an example of the present invention, in which a passenger 3 in a vehicle 1 is photographed by a camera device 2, and then Figure 3 As shown, the key points 31 of the human skeleton of the vehicle occupant can be determined by the existing technology, and based on the connecting line 32 of the key points of the human skeleton, the connecting line 32 is the body feature, and the length of the connecting line 32 is the body feature size.
[0066] Then, step S103 is performed to calculate the characteristic percentile of each real-time body characteristic dimension. The characteristic percentile is the percentile of the real-time body characteristic dimension in a large number of people. Figure 4 The following table shows the human body dimensions of Chinese minors in GB / T 26158-2010. It can be seen that different percentiles correspond to different weights, heights, and different body characteristic dimensions.
[0067] Therefore, the characteristic percentile corresponding to each real-time body characteristic dimension may be calculated first, and then step S104 is performed to calculate the occupant height based on one or more of the characteristic percentiles.
[0068] After the height of the occupant is determined, it can be determined whether to turn on the child monitoring function based on the comparison between the occupant's height and the height threshold. For example, in the OMS system, it is determined whether to turn on the child monitoring function based on the comparison between the occupant's height and the height threshold.
[0069] The present invention determines the body characteristic dimensions of the occupants in the vehicle based on the detection information of the in-vehicle detection equipment, and determines the height of the occupants based on the body characteristic dimensions. The height of the occupants can be inferred by selecting body characteristics that are easy to detect, so that the calculation of the height of the occupants is more accurate.
[0070] like Figure 5 The flowchart of a vehicle occupant height detection method according to another embodiment of the present invention is shown, comprising:
[0071] Step S501, acquiring detection information of an in-vehicle detection device, wherein the detection device is a camera device.
[0072] Step S502, obtaining the coordinates of multiple human skeleton key points of the vehicle occupant and the confidence level of each human skeleton key point from the detection information.
[0073] Step S503, calculating the body feature size of one or more body features as the real-time body feature size according to the coordinates of the key points of the human skeleton.
[0074] Step S504, calculating the confidence of the real-time body feature size according to the confidence of the human skeleton key points used to calculate the real-time body feature size.
[0075] In one embodiment, the body features include: upper arm, forearm and shoulder, and the body feature dimensions include: upper arm length, forearm length and shoulder width.
[0076] Step S505, for each real-time body feature size, obtain the corresponding body feature size and percentile relationship function, substitute the real-time body feature size into the body feature size and feature percentile relationship function, and obtain the feature percentile corresponding to the real-time body feature size.
[0077] Step S506, calculating a height percentile according to one or more of the characteristic percentiles, substituting the height percentile into a function of a relationship between the occupant's height and the height percentile corresponding to the occupant's height, and obtaining the occupant's height corresponding to the height percentile.
[0078] Specifically, firstly, step S501 is executed to obtain detection information of an in-vehicle detection device, where the detection device is a camera device.
[0079] Specifically, Figure 2 As shown, the passenger 3 in the vehicle 1 is photographed by the camera device 2, and then the electronic device executes step S501 to obtain the image taken by the camera device. The camera device is preferably a depth camera, such as a Time of Flight (TOF) camera.
[0080] Then, step S502 is executed to obtain the coordinates of multiple human skeleton key points of the vehicle occupant and the confidence level of each human skeleton key point from the detection information.
[0081] Specifically, for the images captured by the camera device, each frame of the image is analyzed, and the existing image analysis method can be used to obtain the coordinates of multiple human skeleton key points in the image and the confidence level of each human skeleton key point.
[0082] Then, step S503 is executed to calculate the body feature size of one or more body features as the real-time body feature size according to the coordinates of the key points of the human skeleton.
[0083] Specifically, if Figure 6 As shown, the coordinates of the key points of the human skeleton are obtained by the camera device. The body features to be calculated are the shoulder, upper arm and forearm. The calculated body feature dimensions are: shoulder width, upper arm length and forearm length. Therefore, select
[0084] Neck first point: Pn = (Xn, Yn, Zn)
[0085] Left shoulder point Pls = (Xls, Yls, Zls), right shoulder point Prs = (Xrs, Yrs, Zrs);
[0086] The left elbow point Ple=(Xle, Yle, Zle), the right elbow point Pre=(Xre, Yre, Zre);
[0087] The left hand Plw=(Xlw, Ylw, Zlw), the right hand Prw=(Xrw, Yrw, Zrw)
[0088] Calculate the Euclidean distance between two points in the three sets of data respectively:
[0089] Left shoulder width:
[0090] Right shoulder width:
[0091] Left upper arm length:
[0092] Right upper arm length:
[0093] Left forearm length:
[0094] Right forearm length:
[0095] Further processing, take the left and right averages of the upper arm length and forearm length, and sum the left and right shoulder widths to obtain:
[0096] Shoulder width: Ls = (Lln + Lrn)
[0097] Upper arm length: Lu = (Llu + Lru) / 2
[0098] Forearm length: Lf = (Llf + Lrf) / 2
[0099] Then, step S504 is executed to calculate the confidence of the real-time body feature size according to the confidence of the human skeleton key points used to calculate the real-time body feature size.
[0100] Specifically, each human skeleton key point has a corresponding confidence level, and the average value of the confidence levels corresponding to the human skeleton key points used to calculate the real-time body feature size is used as the confidence level of the real-time body feature size.
[0101] For the case where the body features are shoulders, upper arms, and forearms, the calculated body feature dimensions are: shoulder width, upper arm length, and forearm length:
[0102] The confidence of shoulder width is the average of the confidences of three key points of human skeleton: left shoulder point Pls, right shoulder point Prs, and neck first point;
[0103] The confidence of the upper arm length is the average of the confidences of the four key points of the human skeleton: the left shoulder point Pls, the right shoulder point Prs, the left elbow point Ple, and the right elbow point Pre;
[0104] The confidence of the forearm length is the average of the confidences of four key points of the human skeleton: the left elbow point Ple, the right elbow point Pre, the left hand Plw, and the right hand Prw.
[0105] In one embodiment, the body features include: upper arm, forearm and shoulder, and the body feature dimensions include: upper arm length, forearm length and shoulder width.
[0106] Specifically, for the camera device, the sizes of the upper arm, forearm and shoulder can be accurately obtained, so the upper arm, forearm and shoulder are used as body features to calculate their body feature sizes.
[0107] This embodiment uses the upper arm, forearm and shoulder as body features, thereby obtaining accurate body feature dimensions to improve the accuracy of the final occupant height calculation.
[0108] After the real-time body feature size is determined, step S505 is executed to obtain the corresponding body feature size and percentile relationship function for each real-time body feature size, substitute the real-time body feature size into the body feature size and feature percentile relationship function, and obtain the feature percentile corresponding to the real-time body feature size.
[0109] Specifically, the relationship function between the body feature size and the percentile corresponding to each body feature size is determined in advance by fitting. The real-time body feature size is the real-time value of the body feature size of the occupant obtained by the detection device. Substituting the real-time body feature size into the relationship function between the corresponding body feature size and the percentile will obtain the feature percentile corresponding to the real-time body feature size. Each real-time body feature size will be calculated to obtain a corresponding feature percentage.
[0110] After obtaining one or more characteristic percentiles, step S506 is executed to calculate the height percentile according to the one or more characteristic percentiles, and the height percentile is substituted into the occupant height and height percentile relationship function corresponding to the occupant height to obtain the occupant height corresponding to the height percentile.
[0111] Specifically, after obtaining one or more characteristic percentiles, the height percentile is first calculated. When there is only one characteristic percentile, the characteristic percentile can be directly used as the height percentile. When there are multiple characteristic percentiles, the average of the multiple characteristic percentiles can be used as the height percentile, or the confidence of the characteristic percentile, and the height percentile is calculated based on the multiple characteristic percentiles.
[0112] In one embodiment, the number of the characteristic percentiles is greater than or equal to three, and the step of calculating the height percentile based on the plurality of the characteristic percentiles comprises:
[0113] Taking the confidence of each of the real-time body feature dimensions as the confidence of the corresponding feature percentile;
[0114] The confidences of the plurality of characteristic percentiles are calculated in pairs to obtain relative differences of the plurality of confidences. If all the relative differences are less than or equal to the confidence threshold, the average of the plurality of characteristic percentiles is taken as the height percentile, and the relative difference is the relative difference of the two confidences divided by the maximum value of the two confidences; otherwise
[0115] Based on the confidence of the characteristic percentile, the height percentile is calculated based on multiple of the characteristic percentiles.
[0116] Specifically, when the number of feature percentiles is greater than or equal to three, the confidence of each feature percentile is first determined. Then, the confidences of the multiple feature percentiles are calculated in pairs to obtain the relative differences of the multiple confidences. For example, for three confidences, the relative difference between the first confidence and the second confidence, the relative difference between the second confidence and the third confidence, and the relative difference between the first confidence and the third confidence are calculated respectively.
[0117] The relative difference is the relative difference between the two confidence levels divided by the maximum of the two confidence levels. For example, if the first confidence level is Xs, the second confidence level is Xu, and the third confidence level is Xf, then:
[0118] The relative difference between Xs and Xu is: |Xs-Xu| / Max(Xs,Xu)*100%, where Max() is the maximum value function and || is the absolute value.
[0119] The relative difference between Xs and Xf is: |Xs-Xf| / Max(Xs,Xf)*100%;
[0120] The relative difference between Xf and Xu is: |Xf-Xu| / Max(Xf,Xu)*100%.
[0121] If all relative differences are less than or equal to the confidence threshold, the difference between the confidences is not large, and the average of the plurality of characteristic percentiles is taken as the height percentile. The confidence threshold is preferably K.
[0122] If any relative difference is greater than the confidence threshold, the height percentile is calculated based on the confidence of the characteristic percentile and the plurality of characteristic percentiles.
[0123] In this embodiment, for characteristic percentiles greater than or equal to three, two calculations are performed, and the height percentile is determined based on the relative difference of the two calculations. When the relative differences are small, the reliability difference between the characteristic percentiles is not large, so the average of the plurality of characteristic percentiles is used as the height percentile. When the relative difference is large, that is, some characteristic percentiles may be wrong, the height percentile is calculated based on the confidence level to ensure the reliability of the height percentile.
[0124] In one embodiment, the confidence based on the feature percentile, calculating the height percentile based on a plurality of the feature percentiles, comprises:
[0125] Sort by confidence level from highest to lowest;
[0126] Calculate the first relative difference between the highest confidence and the second highest confidence. If the first relative difference is less than or equal to the confidence threshold, take the average of the feature percentile corresponding to the highest confidence and the feature percentile corresponding to the second highest confidence as the height percentile; otherwise
[0127] Calculate the second relative difference between the second highest confidence and the third highest confidence. If the second relative difference is less than or equal to the confidence threshold, take the average of the feature percentile corresponding to the second highest confidence and the feature percentile corresponding to the third highest confidence as the height percentile; otherwise
[0128] Calculate the relative difference between two adjacent confidence levels, and take the average of the two characteristic percentiles corresponding to the two confidence levels with the smallest relative difference as the height percentile.
[0129] Specifically, first select the highest confidence and the second highest confidence (second highest confidence) to calculate the relative difference as the first relative difference. When the first relative difference is small, it can be considered that the values of the two characteristic percentiles with the highest confidence and the second highest confidence are the most reliable. Therefore, the average of the two characteristic percentiles with the highest confidence and the second highest confidence is taken as the height percentile.
[0130] When the first relative difference is greater than the confidence threshold, the relative difference between the second highest confidence and the third highest confidence is calculated as the second relative difference. If the second relative difference is less than or equal to the confidence threshold, the corresponding two feature percentiles are more reliable, and the average of the two feature percentiles with the second highest confidence and the third highest confidence can be directly used as the height percentile.
[0131] When the second relative difference is greater than the confidence threshold, it means that the highest, second highest, and third highest feature percentiles are all unreliable. At this time, sort the confidence from large to small, calculate the relative difference between two adjacent confidences, and take the average of the two feature percentiles corresponding to the two confidences with the smallest relative difference as the height percentile. Since the two feature percentiles corresponding to the two confidences with the smallest relative difference confirm each other, the average of the two feature percentiles is taken as the height percentile.
[0132] As the best embodiment of the present invention, for the case of three characteristic percentiles, the three characteristic percentiles are the characteristic percentiles of upper arm length, forearm length and shoulder width, and the height percentile is determined as follows:
[0133] First, the coordinates of the key points of the human skeleton are obtained through the depth camera, and the following can be calculated:
[0134] Upper arm length: Lu, and its confidence: Xu
[0135] Forearm length: Lf, and its confidence: Xf
[0136] Shoulder width: Ls, and its confidence: Xs
[0137] Then, according to the three values of upper arm length, forearm length and shoulder width, we substitute them into the corresponding function to obtain the characteristic percentiles of the real-time body characteristic dimensions; we get:
[0138] The characteristic percentiles of upper arm length P(Lu), forearm length P(Lf), and shoulder width P(Ls);
[0139] Finally, compare P(Lu), P(Lf), and P(Ls) for analysis:
[0140] If the relative differences of the confidences of the three calculated pairwise are all less than or equal to K, then the average of P(Lu), P(Lf), and P(Ls) is taken as the height percentile;
[0141] If any relative difference of the confidence of the three is greater than K, the confidence judgment is:
[0142] First, select the highest confidence and the second highest confidence to calculate the relative difference:
[0143] If the relative difference between the highest confidence and the second highest confidence is less than or equal to K, the average of the two characteristic percentiles of the highest confidence and the second highest confidence is taken as the height percentile;
[0144] If the relative difference between the highest confidence and the second highest confidence is greater than K, then the second highest confidence is compared with the lowest confidence:
[0145] If the relative difference between the second highest confidence and the lowest confidence is less than or equal to K, take the average of the characteristic percentiles corresponding to the second highest confidence and the lowest confidence as the height percentile;
[0146] If the relative difference between the second highest confidence and the lowest confidence is greater than K, then the relative difference between the first two confidences and the last two confidences is determined:
[0147] If the relative difference between the first two confidence levels is small, the average of the characteristic percentiles corresponding to the highest and second highest confidence levels is selected as the height percentile;
[0148] If the relative difference between the last two confidence levels is small, the average of the characteristic percentiles corresponding to the second highest confidence level and the lowest confidence level is selected as the height percentile.
[0149] Among them, K is the confidence threshold.
[0150] This embodiment provides a method for calculating height percentiles based on confidence. Since confidence is used for comparison, the robustness of the result is improved, so the height percentile obtained is more reliable. Similar to the body characteristic size, the relationship function between the occupant's height and the height percentile is predetermined by fitting. After the height percentile is determined, the height percentile is substituted into the relationship function between the occupant's height and the height percentile corresponding to the occupant's height to obtain the occupant's height corresponding to the height percentile.
[0151] This embodiment acquires an image through a camera device, analyzes the image to obtain one or more real-time body feature dimensions and confidence, and obtains the feature percentile corresponding to the real-time body feature dimension through the relationship function between the body feature dimension and the feature percentile, determines the height percentile based on the feature percentile, and then obtains the height of the occupant corresponding to the height percentile based on the relationship function between the occupant height and the height percentile. Since this embodiment uses the upper arm length, forearm length and shoulder width that are easier to measure by the camera device, the obtained occupant height is more accurate, and the accuracy rate is improved. At the same time, when the occupant is riding, some postures or positions may make it impossible to accurately measure certain body feature dimensions, and this embodiment reverses the occupant height based on multiple body feature dimensions. Therefore, even if some body feature dimensions are incorrect or cannot be measured, the occupant height can still be reversed through other body feature dimensions, thereby improving the robustness of the algorithm.
[0152] In one embodiment, it also includes:
[0153] For each body feature, obtaining multiple sets of body feature training values, the body feature training values including training body feature sizes and corresponding training feature percentiles;
[0154] For the height of the occupant, multiple sets of height training values are obtained, wherein the height training values include the training height and the corresponding training height percentiles.
[0155] Substitute multiple sets of body feature training values of each body feature into the body feature size and feature percentile relationship function corresponding to the body feature, and obtain the body feature size and feature percentile relationship function corresponding to the body feature by fitting. The body feature size and feature percentile relationship function corresponding to the body feature is:
[0156] Among them, L t is the physical feature size of physical feature t, P(L t ) is the characteristic body size L t The feature percentage, N t is the first direction proportional transformation coefficient of body feature t, K t is the second direction proportional transformation coefficient of body feature t, A t is the first direction translation coefficient of body feature t, B t is the second direction translation coefficient of body feature t;
[0157] Substitute multiple sets of height training values into the height and height percentile relationship function respectively, and obtain the height and percentile relationship function by fitting. The height and percentile relationship function is:
[0158] Among them, L h is height, P(L h ) is the characteristic percentage of height, N h is the first direction proportional transformation coefficient of height, K h is the second direction proportional transformation coefficient of height, A h is the first direction translation coefficient of height, B h is the second direction translation coefficient of height.
[0159] Specifically, based on discrete data, the relationship function between the body feature size and the feature percentile, and the relationship function between the height and the height percentile corresponding to each body feature are determined by fitting. The relationship function between the body feature size and the feature percentile, and the relationship function between the height and the height percentile corresponding to each body feature are all fitting functions.
[0160] like Figure 7 The figure shows a schematic diagram of discrete data of multiple physical characteristics and percentiles, height and percentiles, where the unit of the horizontal axis is millimeter and the unit of the vertical axis is percentile, including: height discrete data 71, forearm length discrete data 72, upper arm length discrete data 73, shoulder width discrete data 74. These discrete data are fitted using the "S curve function".
[0161] The relationship function between the physical feature size and the feature percentile corresponding to the physical feature is:
[0162] Among them, L t is the physical feature size of physical feature t, P(L t ) is the characteristic body size L t The feature percentage, N t is the first direction proportional transformation coefficient of body feature t, K t is the second direction proportional transformation coefficient of body feature t, A t is the first direction translation coefficient of body feature t, B t is the second direction translation coefficient of body feature t.
[0163] The relationship function between height and percentile is:
[0164] Among them, L h is height, P(L h ) is the characteristic percentage of height, N h is the first direction proportional transformation coefficient of height, K h is the second direction proportional transformation coefficient of height, A h is the first direction translation coefficient of height, B h is the second direction translation coefficient of height.
[0165] More specifically, the relationship function between upper arm length and characteristic percentile is:
[0166] Among them, L u is the body characteristic size of the upper arm, P(L u ) is the upper arm length L u The feature percentage, N u is the first direction proportional transformation coefficient of the upper arm as the body feature, K u is the second direction proportional transformation coefficient of the upper arm as the body feature, A u is the first direction translation coefficient of the upper arm, B u is the second direction translation coefficient of the body feature being the upper arm.
[0167] The relationship function between forearm length and characteristic percentile is:
[0168] Among them, L f is the body characteristic size of the forearm, P(L f ) is the forearm length L f The feature percentage, N f is the first direction proportional transformation coefficient of the body feature forearm, K f is the second direction proportional transformation coefficient of the body feature forearm, A f is the first direction translation coefficient of the body feature forearm, B f is the second direction translation coefficient of the body feature being the forearm.
[0169] The relationship function between shoulder width and characteristic percentile is:
[0170] Among them, L s is the characteristic size of the body with the shoulder as the characteristic, P(L s ) is the shoulder width L s The feature percentage, N s is the first direction proportional transformation coefficient of the body feature shoulder, K s is the second direction proportional transformation coefficient of the body feature shoulder, A s is the first direction translation coefficient of the body feature shoulder, B s is the second direction translation coefficient of the body feature being the shoulder.
[0171] After the relationship function is determined, multiple sets of body feature training values for each body feature are respectively substituted into the body feature size and feature percentile relationship function corresponding to the body feature, and the body feature size and feature percentile relationship function corresponding to the body feature is obtained by fitting. Multiple sets of height training values are respectively substituted into the height and height percentile relationship function, and the height and percentile relationship function is obtained by fitting.
[0172] like Figure 8 As shown, taking children aged 7-10 years old, the threshold is set to 95% as an example, and the height dimension value is taken as an example. Through fitting calculation, the coefficient N of the height and percentile relationship function is obtained. h , A h , K h and B h .like Fig. 9 Shown is a curve of the fitted height-percentile relationship function, with the unit of the horizontal axis being millimeters and the unit of the vertical axis being percentile, including original discrete data 91 (blue points) and the fitted curve 92 (green curve).
[0173] This embodiment uses an S-curve function to fit the relationship function between body characteristic dimensions and characteristic percentiles and the relationship function between height and percentiles to be consistent with actual discrete data.
[0174] like Fig.10 As shown, as a preferred embodiment of the present invention, a vehicle occupant height detection method includes:
[0175] Step S1001, select the age group and threshold of children: the default is: 7 to 10 years old, 95%;
[0176] Step S1002, the depth camera obtains key points of human skeleton, including:
[0177] Neck first point;
[0178] Left shoulder point Pls, right shoulder point Prs;
[0179] The left elbow point is Ple, the right elbow point is Pre;
[0180] The left hand first Plw, the right hand first Prw;
[0181] Step S1003, calculating the body characteristic dimensions and their confidence levels of the human body:
[0182] Shoulder width: Ls, Xs;
[0183] Upper arm length: Lu, Xu;
[0184] Forearm length: Lf, Xf;
[0185] Step S1004, substitute into the function to obtain the corresponding percentile:
[0186] Shoulder width percentile: P(Ls);
[0187] Upper arm length percentile: P(Lu);
[0188] Forearm length percentile: P(Lf);
[0189] Step S1005, determine:
[0190] |Xs-Xu| / Max(Xs,Xu)*100%≤K
[0191] &
[0192] |Xs-Xf| / Max(Xs,Xf)≤K
[0193] &
[0194] |Xf-Xu| / Max(Xf,Xu)*100%≤K
[0195] whether it is true;
[0196] If true, execute step S1010, otherwise execute step S1006;
[0197] Step S1006, sort Xs, Xu, and Xf, and assign numbers from high to low: X1, X2, X3; and assign numbers L1, L2, L3 to the body feature dimensions Ls, Lu, and Lf corresponding to the confidence levels;
[0198] Step S1007, determine whether |X1-X2| / Max(X1,X2)*100%≤K is true, if so, execute step S1011, otherwise execute step S1008;
[0199] Step S1008, determine whether |X2-X3| / Max(X2,X3)*100%≤K is true, if so, execute step S1012, otherwise execute step S1009;
[0200] Step S1009, determine whether |X1-X2| / Max(X1,X2)*100%≥|X2-X3| / Max(X2,X3)*100% is true, if so, execute step S1011, otherwise execute step S1012;
[0201] Step S1010, calculate the height percentile: P(Lh)=(P(Ls)+P(Lu)+P(Lf)) / 3, and execute step S1013;
[0202] Step S1011, calculate the height percentile: P(Lh)=(P(L1)+P(L2) / 2, and execute step S1013;
[0203] Step S1012, calculate the height percentile: P(Lh)=(P(L2)+P(L3) / 2, and execute step S1013;
[0204] Step S1013, based on P(Lh), substitute into the function to obtain the height dimension: Lh.
[0205] After obtaining the height measurement, the rear child monitoring function activation and judgment logic can be executed.
[0206] like Fig.11 The following is a diagram showing the activation and judgment logic of the rear child monitoring function of the best embodiment of the present invention, including:
[0207] Step S1101, based on P(Lh), substitute into the function to obtain the height dimension: Lh (i.e. Fig.10 Step S1013).
[0208] Step S1102, set the values: age group Yc and percentile calculated height threshold L, the percentile calculated height threshold is the height threshold corresponding to the age group. The age group can be set by the manufacturer or the user, and the height threshold is determined according to the age group by, for example, looking up a table.
[0209] Step S1103, determine whether Lh≤L is true, if so, execute step S1104, otherwise execute step S1105;
[0210] Step S1104: if the passenger's age is ≤ Yc, the child monitoring system is activated;
[0211] Step S1105: if the passenger's age is greater than Yc, the child monitoring system is not started or is turned off.
[0212] The overall calculation and judgment process will be completed within one frame. However, in order to ensure the reliability and stability of the results, the results of the three frames 1 / 3 / 5 can be used for judgment. That is, if within 5 consecutive frames, the execution of step S1103 in the three frames 1 / 3 / 5 are all true, the child monitoring system will be started. If one frame executes step S1103 as false, the child monitoring system will not be started or turned off.
[0213] It should be understood that the order of execution of the steps in the above embodiment does not necessarily mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiment of the present invention.
[0214] like Fig.12The figure shows a hardware structure diagram of an electronic device of the present invention, including:
[0215] at least one processor 1201; and,
[0216] A memory 1202 is communicatively connected to at least one of the processors 1201; wherein:
[0217] The memory 1202 stores instructions that can be executed by at least one of the processors. The instructions are executed by at least one of the processors so that the at least one of the processors can execute the vehicle occupant height detection method as described above.
[0218] Fig.12 A processor 1201 is taken as an example.
[0219] The electronic device may further include: an input device 1203 and a display device 1204 .
[0220] The processor 1201, the memory 1202, the input device 1203 and the display device 1204 may be connected via a bus or other means, and the figure takes the connection via a bus as an example.
[0221] The memory 1202 is a non-volatile computer-readable storage medium that can be used to store non-volatile software programs, non-volatile computer executable programs and modules, such as program instructions / modules corresponding to the vehicle occupant height detection method in the embodiment of the present application, for example, Figure 1 , Figure 5 The processor 1201 executes various functional applications and data processing by running the non-volatile software programs, instructions and modules stored in the memory 1202, that is, the vehicle occupant height detection method in the above embodiment is implemented.
[0222] The memory 1202 may include a program storage area and a data storage area, wherein the program storage area may store an operating system and an application required by at least one function; the data storage area may store data created according to the use of the vehicle occupant height detection method, etc. In addition, the memory 1202 may include a high-speed random access memory, and may also include a non-volatile memory, such as at least one disk storage device, a flash memory device, or other non-volatile solid-state storage device. In some embodiments, the memory 1202 may optionally include a memory remotely arranged relative to the processor 1201, and these remote memories may be connected to a device for executing the vehicle occupant height detection method via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.
[0223] The input device 1203 can receive user clicks and generate signal inputs related to user settings and function control of the vehicle occupant height detection method. The display device 1204 can include display devices such as a display screen.
[0224] The one or more modules are stored in the memory 1202 , and when executed by the one or more processors 1201 , the vehicle occupant height detection method in any of the above method embodiments is executed.
[0225] The present invention is adopted.
[0226] An embodiment of the present invention provides a storage medium, wherein the storage medium stores computer instructions. When a computer executes the computer instructions, the storage medium is used to execute all steps of the vehicle occupant height detection method as described above.
[0227] In the context of the present disclosure, a storage medium may be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, device, or apparatus. The storage medium may be a machine-readable signal medium or a machine-readable storage medium. Alternatively, the storage medium may be a non-temporary computer-readable storage medium, for example, a non-temporary computer-readable storage medium may be a ROM, a random access memory (Random Access Memory, RAM), a compact disc read-only memory (Compact Disc ROM, CD-ROM), a magnetic tape, a floppy disk, and an optical data storage device, etc.
[0228] An embodiment of the present invention provides a computer program product, including a computer program / instruction, which implements the vehicle occupant height detection method as described above when executed by a processor.
[0229] The above-mentioned embodiments only express several implementation methods of the present invention, and the description thereof is relatively specific and detailed, but it cannot be understood as limiting the scope of the patent of the present invention. It should be pointed out that, for ordinary technicians in this field, several variations and improvements can be made without departing from the concept of the present invention, which all belong to the protection scope of the present invention. Therefore, the protection scope of the patent of the present invention shall be subject to the attached claims.
Claims
1. A vehicle occupant height detection method, characterized in that: include: Obtain detection information from in-vehicle detection equipment; determining, based on the detection information, body feature dimensions of one or more body features of the vehicle occupant as real-time body feature dimensions; calculating a characteristic percentile of each of the real-time body characteristic dimensions; Based on one or more of the characteristic percentiles, an occupant height is calculated.
2. The vehicle occupant height detection method according to claim 1, characterized in that: The calculating the percentile of each real-time body feature size includes: for each real-time body feature size, obtaining a corresponding body feature size and percentile relationship function, substituting the real-time body feature size into the body feature size and feature percentile relationship function, and obtaining the feature percentile corresponding to the real-time body feature size; The calculating the occupant's height based on one or more of the characteristic percentiles includes: calculating a height percentile according to one or more of the characteristic percentiles, substituting the height percentile into a function of the relationship between the occupant's height and the height percentile corresponding to the occupant's height, and obtaining the occupant's height corresponding to the height percentile.
3. The vehicle occupant height detection method according to claim 2, characterized in that: Also includes: For each body feature, obtaining multiple sets of body feature training values, the body feature training values including training body feature sizes and corresponding training feature percentiles; For the height of the occupant, multiple sets of height training values are obtained, wherein the height training values include the training height and the corresponding training height percentiles. Substitute multiple sets of body feature training values of each body feature into the body feature size and feature percentile relationship function corresponding to the body feature, and obtain the body feature size and feature percentile relationship function corresponding to the body feature by fitting. The body feature size and feature percentile relationship function corresponding to the body feature is: Among them, L t is the physical feature size of physical feature t, P(L t ) is the characteristic body size L t The feature percentage, N t is the first direction proportional transformation coefficient of body feature t, K t is the second direction proportional transformation coefficient of body feature t, A t is the first direction translation coefficient of body feature t, B t is the second direction translation coefficient of body feature t; Substitute multiple sets of height training values into the height and height percentile relationship function respectively, and obtain the height and percentile relationship function by fitting. The height and percentile relationship function is: Among them, L h is height, P(L h ) is the characteristic percentage of height, N h is the first direction proportional transformation coefficient of height, K h is the second direction proportional transformation coefficient of height, A h is the first direction translation coefficient of height, B h is the second direction translation coefficient of height.
4. The vehicle occupant height detection method according to claim 2, characterized in that: The number of the characteristic percentiles is greater than or equal to three, and the calculating of the height percentile according to the plurality of the characteristic percentiles comprises: Taking the confidence of each of the real-time body feature dimensions as the confidence of the corresponding feature percentile; The confidences of the plurality of characteristic percentiles are calculated in pairs to obtain relative differences of the plurality of confidences. If all the relative differences are less than or equal to the confidence threshold, the average of the plurality of characteristic percentiles is taken as the height percentile, and the relative difference is the relative difference of the two confidences divided by the maximum value of the two confidences; otherwise Based on the confidence of the characteristic percentile, the height percentile is calculated based on multiple of the characteristic percentiles.
5. The vehicle occupant height detection method according to claim 4, characterized in that: The confidence based on the characteristic percentile, calculating the height percentile according to a plurality of the characteristic percentiles, comprises: Sort by confidence level from highest to lowest; Calculate the first relative difference between the highest confidence and the second highest confidence. If the first relative difference is less than or equal to the confidence threshold, take the average of the feature percentile corresponding to the highest confidence and the feature percentile corresponding to the second highest confidence as the height percentile; otherwise Calculate the second relative difference between the second highest confidence and the third highest confidence. If the second relative difference is less than or equal to the confidence threshold, take the average of the feature percentile corresponding to the second highest confidence and the feature percentile corresponding to the third highest confidence as the height percentile; otherwise Calculate the relative difference between two adjacent confidence levels, and take the average of the two characteristic percentiles corresponding to the two confidence levels with the smallest relative difference as the height percentile.
6. The vehicle occupant height detection method according to claim 1, characterized in that: The detection device is a camera device, and determining the physical feature size of one or more physical features of the vehicle occupant as the real-time physical feature size based on the detection information includes: Obtaining coordinates of multiple human skeleton key points of the vehicle occupant and the confidence level of each human skeleton key point from the detection information; Calculating the body feature size of one or more body features as the real-time body feature size according to the coordinates of the human skeleton key points; The confidence of the real-time body feature size is calculated based on the confidence of the human skeleton key points used to calculate the real-time body feature size.
7. The vehicle occupant height detection method according to any one of claims 1 to 6, characterized in that: The body features include: upper arms, forearms and shoulders, and the body feature dimensions include: upper arm length, forearm length and shoulder width.
8. An electronic device, characterized in that: include: at least one processor; as well as, a memory communicatively connected to at least one of the processors; wherein, The memory stores instructions that can be executed by at least one of the processors, and the instructions are executed by at least one of the processors to enable at least one of the processors to execute the vehicle occupant height detection method according to any one of claims 1 to 7.
9. A storage medium, characterized in that: The storage medium stores computer instructions, and when a computer executes the computer instructions, it is used to execute all steps of the vehicle occupant height detection method as described in any one of claims 1 to 7.
10. A computer program product comprising a computer program / instructions, characterized in that When the computer program / instruction is executed by a processor, the vehicle occupant height detection method as claimed in any one of claims 1 to 7 is implemented.
Citation Information
Patent Citations
Method and device for estimating height and weight of passengers
CN111507174A
Height and body size measuring method based on monocular vision and key point recognition
CN111611928A
Physique estimation device and physique estimation method
CN111867466A
Lifting table sitting posture height recommendation method for healthy office
CN114297492A
