Target calibration method and device, electronic equipment and storage medium

Abstract

The invention provides a target calibration method and device, electronic equipment and a storage medium, and relates to the technical field of radar videos, and the method comprises the steps: mapping a current radar target to a video coordinate system corresponding to a video target based on a calibration matrix, and obtaining a current calibration target; the radar target is a target detected by radar, and the video target is a target detected by video monitoring equipment; and correcting the position of the current calibration target under the condition that the position deviation between the current video target matched with the current radar target and the current calibration target is greater than a preset deviation threshold value. The calibration result is corrected on the basis of preliminary calibration, the defect that the calibration precision of one-time calibration is difficult to guarantee is overcome, and the calibration precision is improved.

Description

Technical Field

[0001] The present invention relates to the field of radar video technology, and in particular to a target calibration method, device, electronic equipment and storage medium. Background Art

[0002] At present, radar cameras are used for special services of video radar data fusion, in addition to conventional video monitoring and radar target presentation. For example, capturing and superimposing speed trajectories, etc. The calibration matrix is ​​the core link of video radar data fusion, and its accuracy directly affects the effect of the implementation of special services of data fusion.

[0003] At present, the mainstream video radar calibration method in the industry is one-time calibration, which relies on human experience to a certain extent. The calibration accuracy is difficult to guarantee, resulting in uneven presentation of data fusion services and affecting user use. Summary of the invention

[0004] The present invention provides a target calibration method, device, electronic equipment and storage medium, which are used to solve the defect that the calibration accuracy is difficult to ensure in the prior art and improve the calibration accuracy.

[0005] The present invention provides a target calibration method, comprising:

[0006] Based on the calibration matrix, the current radar target is mapped to the video coordinate system corresponding to the video target to obtain the current calibration target; the radar target is the target detected by the radar, and the video target is the target detected by the video surveillance device;

[0007] When a position deviation between a current video target that matches the current radar target and the current calibration target is greater than a preset deviation threshold, the position of the current calibration target is corrected.

[0008] In some embodiments, the correcting the position of the current calibration target includes:

[0009] Determine the abscissa difference and ordinate difference between the position coordinates of the current video target and the position coordinates of the current calibration target;

[0010] Performing fuzzy PID processing on the horizontal coordinate difference and the vertical coordinate difference respectively to obtain a horizontal coordinate correction amount and a vertical coordinate correction amount;

[0011] The position of the current calibration target is corrected according to the horizontal coordinate correction amount and the vertical coordinate correction amount.

[0012] In some embodiments, the performing fuzzy PID processing on the horizontal coordinate difference and the vertical coordinate respectively to obtain the horizontal coordinate correction amount and the vertical coordinate correction amount includes:

[0013] According to the one-to-one correspondence between the difference range and the fuzzy subset, determine the fuzzy subset corresponding to the horizontal coordinate difference, the fuzzy subset corresponding to the rate of change of the horizontal coordinate difference, the fuzzy subset corresponding to the vertical coordinate difference, and the fuzzy subset corresponding to the rate of change of the vertical coordinate difference;

[0014] Performing fuzzy PID processing based on the fuzzy subset corresponding to the horizontal coordinate difference and the fuzzy subset corresponding to the rate of change of the horizontal coordinate difference to obtain a first fuzzy PID coefficient;

[0015] Based on the first fuzzy PID coefficient and the horizontal coordinate difference, obtaining the horizontal coordinate correction amount;

[0016] Performing fuzzy PID processing based on the fuzzy subset corresponding to the ordinate difference and the fuzzy subset corresponding to the rate of change of the ordinate difference to obtain a second fuzzy PID coefficient;

[0017] The ordinate correction amount is obtained based on the second fuzzy PID coefficient and the ordinate difference.

[0018] In some embodiments, before mapping the current radar target to the video coordinate system corresponding to the video target based on the calibration matrix to obtain the current calibration target, the method further includes:

[0019] Acquire matching radar targets and video targets;

[0020] The position coordinates of the matched radar target and the position coordinates of the video target are stored in a hash table;

[0021] When the radar targets and video targets in the hash table meet the preset calibration matrix obtaining conditions, the calibration matrix is ​​obtained according to the position coordinates of the matching radar targets and video targets in the hash table.

[0022] In some embodiments, each of the hash tables includes N hash buckets, each hash bucket includes Y empty slots, each empty slot is used to store an array, N is a positive integer greater than zero, Y is a positive integer greater than zero, and the product of N and Y is 10000;

[0023] The storing the position coordinates of the matched radar target and the position coordinates of the video target in a hash table includes:

[0024] Determining an empty slot corresponding to the radar target according to the ten-thousandth ratio of the position coordinate value of the radar target;

[0025] The ten-thousandth ratio of the position coordinate value of the video target that matches the radar target is stored in the empty slot corresponding to the radar target.

[0026] In some embodiments, the obtaining of matched radar targets and video targets includes:

[0027] Get the video image and radar image with the closest timestamp;

[0028] When the number of video targets in the video image and the number of radar targets in the radar image are the same, and the distance between the video targets in the video image and the distance between the radar targets in the radar image are both greater than a preset distance, the video targets in the video image are matched with the radar targets in the radar image to obtain the matched radar targets and video targets.

[0029] In some embodiments, matching the video target in the video image with the radar target in the radar image to obtain the matched radar target and video target includes:

[0030] The video targets in the video image and the radar targets in the radar image are respectively sorted in order from top to bottom and from left to right based on the center of the image to obtain a video target sorting result and a radar target sorting result;

[0031] The video targets and radar targets at the same sorting position in the video target sorting result and the radar target sorting result are matched to obtain the matched radar targets and video targets.

[0032] The present invention also provides a target calibration device, comprising:

[0033] A calibration module is used to map the current radar target to the video coordinate system corresponding to the video target based on the calibration matrix to obtain the current calibration target; the radar target is the target detected by the radar, and the video target is the target detected by the video monitoring device;

[0034] The correction module is used to correct the position of the current calibration target when the position deviation between the current video target matching the current radar target and the current calibration target is greater than a preset deviation threshold.

[0035] The present invention also provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, any of the target calibration methods described above is implemented.

[0036] The present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, and when the computer program is executed by a processor, the target calibration method as described above is implemented.

[0037] The present invention also provides a computer program product, comprising a computer program, wherein when the computer program is executed by a processor, the computer program implements any of the target calibration methods described above.

[0038] The target calibration method, device, electronic device and storage medium provided by the present invention perform calibration conversion based on the calibration matrix. When the position deviation between the current video target and the current calibration target is greater than a preset deviation threshold, the position of the current calibration target is corrected, thereby correcting the calibration result based on the preliminary calibration, solving the problem that the calibration accuracy of a one-time calibration is difficult to guarantee and improving the calibration accuracy. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] In order to more clearly illustrate the technical solutions in the present invention or the prior art, a brief introduction will be given below to the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0040] Figure 1 It is one of the flow charts of the target calibration method provided by the present invention;

[0041] Figure 2 This is the second flow chart of the target calibration method provided by the present invention;

[0042] Figure 3 This is the third flow chart of the target calibration method provided by the present invention;

[0043] Figure 4 It is a schematic diagram of a video target in a video image and a radar target in a radar image provided by the present invention;

[0044] Figure 5 It is a structural schematic diagram of the target calibration device provided by the present invention;

[0045] Figure 6 It is a structural schematic diagram of the electronic device provided by the present invention. DETAILED DESCRIPTION

[0046] In order to make the purpose, technical solutions and advantages of the present invention clearer, the technical solutions in the present invention will be clearly and completely described below in conjunction with the drawings in the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, rather than all of the embodiments. It should be noted that the embodiments of the present invention and the features in the embodiments can be combined with each other in the absence of conflict. Based on the embodiments in the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0047] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those commonly understood by those skilled in the art of the present invention. The terms used in the specification of the present invention herein are only for the purpose of describing specific embodiments and are not intended to limit the present invention.

[0048] It should be further noted that, in this article, the terms "comprises", "includes" or any other variations thereof are 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 includes other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, an element defined by the sentence "comprises a ..." does not exclude the presence of other identical elements in the process, method, article or device including the element.

[0049] In the present invention, "at least one" means one or more, and "more than one" means two or more than two. The terms "first", "second", "third", "fourth", etc. (if any) in the present invention are used to distinguish similar objects, rather than to describe a specific order or sequence.

[0050] In the embodiments of the present invention, words such as "exemplary" or "for example" are used to indicate examples, illustrations or explanations. Any embodiment or design described as "exemplary" or "for example" in the embodiments of the present invention should not be interpreted as being more preferred or more advantageous than other embodiments or designs. Specifically, the use of words such as "exemplary" or "for example" is intended to present related concepts in a specific way.

[0051] Please refer to Figure 1 , Figure 1 yes Figure 1 This is one of the flow charts of the target calibration method provided by the present invention. The present invention provides a target calibration method, which includes:

[0052] Step 110, based on the calibration matrix, the current radar target is mapped to the video coordinate system corresponding to the video target to obtain the current calibrated target; the radar target is the target detected by the radar, and the video target is the target detected by the video surveillance device.

[0053] Specifically, a radar target is a target detected by a radar, that is, a radar target is determined in a radar coordinate system. A video target is a target detected by a video surveillance device, that is, a video target is determined in a video coordinate system. The calibration matrix represents the conversion relationship between the radar coordinate system and the video coordinate system.

[0054] According to the calibration matrix, the current radar target is mapped to the video coordinate system to obtain the current calibration target, which is the correspondence of the current radar target in the video coordinate system.

[0055] Step 120, when the position deviation between the current video target that matches the current radar target and the current calibration target is greater than a preset deviation threshold, the position of the current calibration target is corrected.

[0056] Specifically, a current video target that matches the current radar target is determined, a position deviation between the current video target and the current calibration target is calculated, and the position deviation between the current video target and the current calibration target is compared with a preset deviation threshold.

[0057] Optionally, the position deviation between the current video target and the current calibration target may be a Euclidean distance between the current video target and the current calibration target. Optionally, the preset deviation threshold may be preset according to the required calibration accuracy.

[0058] In some embodiments, in order to speed up the correction process, the video coordinate system corresponding to the video target may be divided into multiple quadrants, for example, four quadrants, and the Euclidean distance between the video target and the calibration target in each quadrant is calculated.

[0059] When the deviation between the current video target and the current calibration target is less than or equal to the preset deviation threshold, it indicates that the current calibration target meets the calibration accuracy; when the deviation between the video target and the calibration target is greater than the preset deviation threshold, it indicates that the current calibration target does not meet the calibration accuracy, and the current calibration target is corrected so that the corrected current calibration target meets the calibration accuracy.

[0060] The target calibration method provided by the present invention performs calibration conversion based on the calibration matrix. When the position deviation between the current video target and the current calibration target is greater than a preset deviation threshold, the position of the current calibration target is corrected, thereby correcting the calibration result based on the preliminary calibration, solving the problem that the calibration accuracy of a one-time calibration is difficult to guarantee and improving the calibration accuracy.

[0061] Optionally, this embodiment provides a target calibration method, which is a main description of the above embodiment, and mainly describes the specific process of correcting the position of the current calibration target. The target calibration method provided by this embodiment includes:

[0062] Determine the abscissa difference and ordinate difference between the position coordinates of the current video target and the position coordinates of the current calibration target;

[0063] Perform fuzzy PID processing on the horizontal coordinate difference and the vertical coordinate difference respectively to obtain the horizontal coordinate correction value and the vertical coordinate correction value;

[0064] The position of the current calibration target is corrected according to the horizontal coordinate correction amount and the vertical coordinate correction amount.

[0065] Specifically, the horizontal axis of the video coordinate system represents the pixel width, and the vertical axis of the video coordinate system represents the pixel height. The position coordinates of the video target are the pixel width and pixel height of the video target in the video coordinates. The position coordinates of the current calibration target are the pixel width and pixel height of the current calibration target in the video coordinates.

[0066] According to the abscissa difference and ordinate difference between the position coordinates of the current video target and the position coordinates of the current calibration target, the abscissa difference and ordinate difference between the current video target and the current calibration target, that is, the pixel width difference and the pixel height difference, are determined.

[0067] The horizontal coordinate difference and the vertical coordinate difference are processed by fuzzy PID respectively to obtain the horizontal coordinate correction amount and the vertical coordinate correction amount.

[0068] The horizontal coordinate value of the current calibration target is added with the horizontal coordinate correction amount to obtain the corrected horizontal coordinate value, and the vertical coordinate value of the current calibration target is added with the vertical coordinate correction amount to obtain the corrected vertical coordinate value. The corrected position of the current calibration target in the video coordinate system can be determined based on the corrected vertical coordinate value and the corrected vertical coordinate value.

[0069] For example, the current radar target is a(x,y), the current video target that matches the current radar target a is A(X,Y), and the current calibration target obtained by mapping the current radar target a to the video coordinate system is a′(x′,y′). The expression of the current calibration target a′(x′,y′) is as follows:

[0070] a′(x′,y′)=a(x,y)T

[0071] Where a′(x′,y′) represents the position of the current calibration target a′ in the video coordinate system, x′ represents the abscissa value of the current calibration target a′ in the video coordinate system, y′ represents the abscissa value of the current calibration target a′ in the video coordinate system, a(x,y) represents the position of the current radar target a in the radar coordinate system, x represents the abscissa value of the current radar target a in the radar coordinate system, y represents the abscissa value of the current radar target a in the radar coordinate system, and T is the calibration matrix.

[0072] The horizontal coordinate difference between the current video target A and the current calibration target a′ is e x (t), horizontal coordinate difference e x The expression of (t) is as follows:

[0073] e x (t) = Xx′

[0074] In the formula, e x (t) represents the horizontal coordinate difference between the current video target A and the current calibration target a′, X represents the horizontal coordinate value of the current video target A in the video coordinate system, and x′ represents the horizontal coordinate value of the current calibration target a′ in the video coordinate system.

[0075] The vertical coordinate difference between the current video target A and the current calibration target a′ is e y (t), ordinate difference e y The expression of (t) is as follows:

[0076] e y (t) = Yy′

[0077] In the formula, e y (t) represents the ordinate difference between the current video target A and the current calibration target a′, Y represents the ordinate value of the current video target A in the video coordinate system, and y′ represents the ordinate value of the current calibration target a′ in the video coordinate system.

[0078] The horizontal axis difference e x (t) Perform fuzzy PID processing to obtain the horizontal coordinate correction value U x (t), the vertical coordinate difference e y (t) Perform fuzzy PID processing to obtain the ordinate positive quantity U y (t).

[0079] According to the horizontal coordinate correction value U x (t) and the abscissa value of the current calibration target a′, and obtain the corrected abscissa value of the current calibration target a′. The corrected abscissa value of the current calibration target a′ is expressed as follows:

[0080] S x (t) = U x (t)+x′

[0081] In the formula, S x (t) represents the corrected horizontal coordinate value of the current calibration target a′, U x (t) represents the horizontal coordinate correction value, and x′ represents the horizontal coordinate value of the current calibration target a′ in the video coordinate system.

[0082] According to the ordinate correction value U y (t) and the ordinate value of the current calibration target a′, and obtain the corrected ordinate value of the current calibration target a′. The corrected ordinate value of the current calibration target a′ is expressed as follows:

[0083] S y (t) = U y (t)+y′

[0084] In the formula, S y (t) represents the corrected ordinate value of the current calibration target a′, U y (t) represents the vertical coordinate correction value, and y′ represents the vertical coordinate value of the current calibration target a′ in the video coordinate system.

[0085] After determining the corrected horizontal coordinate value S of the current calibration target a′ in the video coordinate system x (t) and the ordinate value S y After (t), the coordinates of the corrected current calibration target a′ in the video coordinate system can be obtained as (S x (t),S y (t)).

[0086] The target calibration method provided by the present invention determines the horizontal coordinate difference and the vertical coordinate difference between the current video target and the current calibration target, uses fuzzy PID to process the horizontal coordinate difference and the vertical coordinate difference respectively, obtains the horizontal coordinate correction amount and the vertical coordinate correction amount, and corrects the position of the current calibration target according to the horizontal coordinate correction amount and the vertical coordinate correction amount. Since fuzzy PID can improve the determination speed and accuracy of the horizontal coordinate correction amount and the vertical coordinate correction amount, the purpose of quickly improving the mapping accuracy is achieved.

[0087] Optionally, this embodiment provides a target calibration method, which is a main description of the above embodiment, and mainly describes the specific process of performing fuzzy PID processing on the horizontal coordinate difference and the vertical coordinate difference to obtain the horizontal coordinate correction value and the vertical coordinate correction value. The target calibration method provided by this embodiment includes:

[0088] According to the one-to-one correspondence between the difference range and the fuzzy subset, the fuzzy subset corresponding to the horizontal coordinate difference, the fuzzy subset corresponding to the change rate of the horizontal coordinate difference, the fuzzy subset corresponding to the vertical coordinate difference and the fuzzy subset corresponding to the change rate of the vertical coordinate difference are determined;

[0089] Performing fuzzy PID processing based on the fuzzy subset corresponding to the abscissa difference and the fuzzy subset corresponding to the rate of change of the abscissa difference to obtain a first fuzzy PID coefficient;

[0090] Based on the first fuzzy PID coefficient and the abscissa difference, obtaining an abscissa correction amount;

[0091] Performing fuzzy PID processing based on the fuzzy subset corresponding to the ordinate difference and the fuzzy subset corresponding to the rate of change of the ordinate difference to obtain a second fuzzy PID coefficient;

[0092] Based on the second fuzzy PID coefficient and the ordinate difference, a ordinate correction amount is obtained.

[0093] Specifically, for the horizontal coordinate difference e x(t) and the rate of change of the horizontal coordinate difference E cx (t) Perform fuzzy processing and map it into fuzzy subsets of negative large, negative small, zero, positive large, and positive small to determine e x (t) The corresponding fuzzy subset and E cx (t) The corresponding fuzzy subset.

[0094] Exemplarily, Table 1 is a table of correspondence between difference ranges and fuzzy subsets. As shown in Table 1, the fuzzy subset corresponding to the difference e(t)∈[-300,-100) is negative large; the fuzzy subset corresponding to the difference e(t)∈[-100,-50) is negative small; the fuzzy subset corresponding to the difference e(t)∈[-50,50) is zero; the fuzzy subset corresponding to the difference e(t)∈[50,100) is positive small; the fuzzy subset corresponding to the difference e(t)∈[-100,300] is positive large.

[0095] Table 1 Correspondence between difference range and fuzzy subset

[0096]

[0097]

[0098] For example, Table 2 is a corresponding table relationship table between the difference change rate and the fuzzy subset. As shown in Table 2, the difference change rate E c (t)∈(-∞,-0.5) corresponds to a large negative fuzzy subset; the rate of change of the difference E c (t)∈[-0.5,-0.1) corresponds to a small negative fuzzy subset; the difference change rate E c (t)∈[-0.1,0.1) corresponds to zero in the fuzzy subset; the difference change rate E c (t)∈[-0.1,0.5) corresponds to a small positive fuzzy subset; the rate of change of the difference E c The corresponding fuzzy subset of (t)∈[0.5,+∞) is positive.

[0099] Table 2 Correspondence between difference change rate range and fuzzy subsets

[0100] Difference range Fuzzy subset <![CDATA[-∞<E c (t)<-0.5]]> Negative <![CDATA[-0.5≤E c (t)<-0.1]]> Negative small <![CDATA[-0.1≤E c (t)<0.1]]> zero <![CDATA[0.1≤E c (t)<0.5]]> Just small <![CDATA[0.5≤E c (t)<+∞]]> Zhengda

[0101] In determining the horizontal coordinate difference e x (t) The corresponding fuzzy subset and the rate of change of the horizontal coordinate difference E cx (t) The corresponding fuzzy subset is then based on the horizontal coordinate difference e x (t) The corresponding fuzzy subset and the rate of change of the horizontal coordinate difference E cxThe fuzzy subset corresponding to (t) is subjected to fuzzy PID processing, that is, fuzzification processing and clarity processing are performed to obtain the first fuzzy PID coefficient. The fuzzy PID coefficient includes a proportional coefficient, an integral coefficient and a differential coefficient.

[0102] The following is based on the horizontal coordinate difference e x (t) The corresponding fuzzy subset and the rate of change of the horizontal coordinate difference E cx The specific process of fuzzy PID processing on the fuzzy subset corresponding to (t) is described in detail.

[0103] According to the horizontal coordinate difference e x (t) The corresponding fuzzy subset and the rate of change of the horizontal coordinate difference E cx (t) The corresponding fuzzy subset is used to infer the incremental fuzzy values ​​of the proportional, integral, and differential coefficients, respectively, to determine the incremental fuzzy value ΔK of the proportional coefficient p , the incremental fuzzy value of the integral coefficient ΔK i and the incremental fuzzy value of the differential coefficient ΔK d .

[0104] For example, Table 3 shows ΔK p Inference rule table, Table 4 is ΔK i Inference rule table, Table 5 is ΔK d Inference rule table. According to e x (t) The corresponding fuzzy subset and E cx (t) The corresponding fuzzy subset and the inference rules in Tables 3 to 5 can determine the incremental fuzzy value ΔK of the proportional coefficient. p , the incremental fuzzy value of the integral coefficient ΔK i and the incremental fuzzy value of the differential coefficient ΔK d .

[0105] Table 3ΔK p Inference rules table

[0106]

[0107] Table 4ΔK i Inference rules table

[0108]

[0109] Table 5ΔK d Inference rules table

[0110]

[0111]

[0112] Finally, the incremental fuzzy value of the proportional coefficient ΔK p , the incremental fuzzy value of the integral coefficient ΔKi , the incremental fuzzy value of the differential coefficient ΔK d Perform clarity processing respectively to obtain the incremental clarity value ΔK of the proportional coefficient P , the incremental clarity value of the integral coefficient ΔK I and the incremental clarity value of the differential coefficient ΔK D .

[0113] For example, Table 6 is a table showing the correspondence between the incremental blur value and the incremental clarity value of the proportional coefficient, Table 7 is a table showing the correspondence between the incremental blur value and the incremental clarity value of the integral coefficient, and Table 8 is a table showing the correspondence between the incremental blur value and the incremental clarity value of the integral coefficient. According to the correspondence in Tables 6 to 8, and the incremental blur value ΔK of the proportional coefficient p , the incremental fuzzy value of the integral coefficient ΔK i , the incremental fuzzy value of the differential coefficient ΔK d , we can get the incremental clarity value of the proportional coefficient ΔK P , the incremental clarity value of the integral coefficient ΔK I and the incremental clarity value of the differential coefficient ΔK D .

[0114] Table 6 Correspondence between incremental blur value and incremental clarity value of proportional coefficient

[0115] <![CDATA[Incremental fuzzy value ΔK of the proportionality coefficient P > <![CDATA[Increment Clear Value ΔK of Proportionality Coefficient P > Negative -0.8 Negative small -0.5 zero 0 Just small 0.5 Zhengda 0.8

[0116] Table 7 Correspondence between incremental blur value and incremental clarity value of integral coefficient

[0117] <![CDATA[Increment fuzzy value ΔK of integral coefficient i > <![CDATA[Clear value of the increment of the integral coefficient ΔK I > Negative -0.06 Negative small -0.03 zero 0 Just small 0.03 Zhengda 0.06

[0118] Table 8 Correspondence between incremental blur value and incremental clarity value of integral coefficient

[0119] <![CDATA[Increment fuzzy value ΔK of differential coefficient d > <![CDATA[Clear value ΔK of the increment of the differential coefficient D > Negative -0.008 Negative small -0.004 zero 0 Just small 0.004 Zhengda 0.008

[0120] In order to obtain the incremental clarity value of the proportional coefficient ΔK P , the incremental clarity value of the integral coefficient ΔK I and the incremental clarity value of the differential coefficient ΔK D Afterwards, according to the proportional coefficient, integral coefficient and differential coefficient of the previous moment, and the incremental clear value ΔK of the proportional coefficient P , the incremental clarity value of the integral coefficient ΔK I and the incremental clarity value of the differential coefficient ΔK D , respectively determine the proportional coefficient, integral coefficient and differential coefficient at the current moment.

[0121] The expression of the current proportional coefficient is as follows:

[0122] Kp (t) = K p (t-1)+ΔK P

[0123] In the formula, K p (t) represents the current proportionality coefficient, K p (t-1) represents the proportionality coefficient at the previous moment, ΔK P Indicates the incremental sharp value of the scale factor.

[0124] The expression of the integral coefficient at the current moment is as follows:

[0125] K i (t) = K i (t-1)+ΔK I

[0126] In the formula, K i (t) represents the integration coefficient at the current moment, K i (t-1) represents the integration coefficient of the previous moment, ΔK I Indicates the incremental clear value of the integral coefficient.

[0127] The expression of the differential coefficient at the current moment is as follows:

[0128] K d (t) = K d (t-1)+ΔK D

[0129] In the formula, K d (t) represents the differential coefficient at the current moment, K d (t-1) represents the differential coefficient at the previous moment, ΔK D Indicates the incremental clarity value of the differential coefficient.

[0130] After determining the current proportional coefficient, integral coefficient and differential coefficient, the current proportional coefficient, integral coefficient and differential coefficient are substituted into the fuzzy PID processing formula to obtain the horizontal coordinate correction amount.

[0131] The fuzzy PID processing formula is as follows:

[0132]

[0133] Where u(t) represents the correction value of fuzzy PID processing, K p represents the proportionality coefficient, e(t) represents the difference, K i Indicates the integral coefficient, K d represents the differential coefficient.

[0134] Since the fuzzy PID processing process of the ordinate difference is similar to that of the abscissa difference, it will not be described here in order to avoid repetition.

[0135] The target calibration method provided by the present invention further improves the determination speed and accuracy of the horizontal coordinate correction amount and the vertical coordinate correction amount by first determining the fuzzy subsets corresponding to the horizontal coordinate difference and the vertical coordinate difference, then determining the fuzzy PID coefficient according to the fuzzy subsets, and finally performing fuzzy PID processing based on the fuzzy PID coefficient.

[0136] In some embodiments, the current radar target and the current video target that matches the current radar target may be considered as the current target pair. If the abscissa difference exceeds the difference range or the ordinate difference exceeds the difference range, the current target pair is considered to be an abnormal target pair, and the abnormal target pair is discarded to prevent affecting the stability of the system.

[0137] Optionally, this embodiment provides a target calibration method, which is a further improvement of the above embodiment. The main improvement is: based on the calibration matrix, the current radar target is mapped to the video coordinate system corresponding to the video target, and before the current calibration target is obtained, the specific process of obtaining the calibration matrix is ​​obtained. The target calibration method provided in this embodiment includes:

[0138] Acquire matching radar targets and video targets;

[0139] The position coordinates of the matched radar target and the position coordinates of the video target are stored in a hash table;

[0140] When the radar target and the video target in the hash table meet the preset calibration matrix obtaining condition, the calibration matrix is ​​obtained according to the position coordinates of the matching radar target and the position coordinates of the video target in the hash table.

[0141] Specifically, please refer to Figure 2 , Figure 2 This is the second flow chart of the target calibration method provided by the present invention. The specific process of obtaining the calibration matrix includes the following steps:

[0142] Step 210, obtaining matching radar targets and video targets.

[0143] Matching radar targets and video targets can be obtained manually or by automatically matching video images with radar images.

[0144] Step 220: Store the position coordinates of the matched radar target and the position coordinates of the video target in a hash table.

[0145] The storage position in the hash table is determined according to the ten-thousandth ratio of the position coordinates of the radar target. After the storage position is determined, the ten-thousandth ratio of the position coordinates of the video target is stored in the storage position.

[0146] Step 230, determining whether the radar target and the video target in the hash table meet the preset calibration matrix obtaining conditions.

[0147] The preset calibration matrix obtaining condition may be pre-set according to the required calibration accuracy or the required calibration effect. For example, the preset calibration matrix obtaining condition may be that the position of the radar target needs to be in a specific area of ​​the radar image, and the position of the video target needs to be in a specific area of ​​the video image. For example, the preset calibration matrix obtaining condition may also be that the radar target and the video target need to reach a preset target number.

[0148] When the radar targets and video targets in the hash table meet the preset calibration matrix obtaining conditions, execute step 240; when the radar targets and video targets in the hash table do not meet the preset calibration matrix obtaining conditions, execute steps 210 to 220 again, that is, continue to obtain matching radar targets and video targets, and store the position coordinates of the matching radar targets and the position coordinates of the video targets in the hash table again, until the radar targets and video targets in the hash table meet the preset calibration matrix obtaining conditions, and execute step 240.

[0149] Step 240, obtaining a calibration matrix according to the position coordinates of the radar target and the position coordinates of the video target that match in the hash table.

[0150] Specifically, according to the position coordinates of the matching radar target and the position coordinates of the video target in the hash table, the conversion relationship between the radar coordinate system and the video coordinate system is inferred, and the conversion relationship is the calibration matrix that needs to be obtained.

[0151] In some embodiments, the target detected in the radar coordinate system is converted to the video coordinate system through the obtained calibration matrix, and the coordinate value of the calibration target is obtained. The coordinate value of the calibration target is compared with the coordinate value of the target detected in the actual video coordinate system. If the difference between the two is greater than the maximum threshold, it is determined that the obtained calibration matrix is ​​inaccurate and the calibration matrix needs to be re-obtained.

[0152] The target calibration method provided by the present invention is beneficial to improving the accuracy of the calibration matrix by obtaining the calibration matrix when the radar target and the video target meet the preset calibration matrix obtaining conditions, thereby further improving the calibration precision.

[0153] Optionally, this embodiment provides a target calibration method, which is a further description of the above embodiment, mainly describing the specific process of storing the position coordinates of the matching radar target and the position coordinates of the video target in a hash table. The target calibration method provided in this embodiment includes:

[0154] According to the ten-thousandth ratio of the position coordinate value of the radar target, the empty slot corresponding to the radar target is determined;

[0155] The ten-thousandth ratio of the position coordinate value of the video target that matches the radar target is stored in the empty slot corresponding to the radar target.

[0156] Specifically, each hash table includes N hash buckets, each hash bucket contains Y empty slots, each empty slot is used to store an array, N is a positive integer greater than zero, Y is a positive integer greater than zero, and the product of N and Y is 10000. The hash table is designed to store 10000 arrays, which corresponds to the range of 0 to 9999 of the horizontal coordinate of the radar coordinate system. The arrays in each hash bucket are connected in the form of a linked list, and the header of the linked list is stored in the hash table.

[0157] For example, each hash table includes 100 hash buckets, the hash buckets are sorted in the range of 0 to 99, each hash bucket contains 100 empty slots, the empty slots are sorted in the range of 0 to 99, and each empty slot can store an array. In this case, each hash table can store 10,000 arrays.

[0158] Since the description units of radar targets and video targets are different, it is necessary to perform quantization processing on the coordinates of radar targets and video targets.

[0159] After the hash table is designed, the hash table corresponding to the radar target is first determined according to the ten-thousandth ratio of the radar target's horizontal coordinate value; then the hash bucket in the hash table corresponding to the radar target is determined according to the result of taking the quotient of N by the ten-thousandth ratio of the radar target's vertical coordinate value; then the empty slot in the hash bucket corresponding to the radar target is determined according to the result of taking the remainder of Y by the ten-thousandth ratio of the radar target's vertical coordinate value; finally, the ten-thousandth ratio of the position coordinate value of the video target matching the radar target is stored in the empty slot.

[0160] Assume that the coordinates of the video target are (400,1000) and the resolution is 2560*1440. After the 10,000-percent ratio processing, the 10,000-percent ratio of the horizontal coordinate of the video target is 400 / 2560*10000=1563, and the 1000 / 1440*10000=6944. The coordinates of the radar target are (10,40) and the radar detection range is 40*80. After the 10,000-percent ratio conversion, the 10,000-percent ratio of the horizontal coordinate of the radar target is 10 / 40*10000=2500, and the 10,000-percent ratio of the vertical coordinate is 40 / 80*10000=5000.

[0161] The converted radar target coordinates (2500, 5000) and the video target coordinates (1563, 6944) are stored in the hash table. The specific storage process is as follows:

[0162] First, search the 2500th hash table. If it does not exist, create the 2500th hash table. Since 5000 / 100=50, then search the 50th hash bucket in the 2500th hash table. Since 5000%100=0, finally save the coordinate ratio (1563, 6944) of the video target in the 0th empty slot in the 50th hash bucket in the 2500th hash table.

[0163] After storing the position coordinates of the matched video target and the position coordinates of the radar target in the hash table and obtaining the calibration matrix, the radar target can be mapped to the video coordinate system. The mapping strategy can be as follows:

[0164] First, the video target corresponding to the radar target is searched in the hash table. If the corresponding video target is found, the video target is the mapping of the radar target in the video coordinate system; if the corresponding video target is not found in the hash bucket, it is converted through the calibration matrix.

[0165] The target calibration method provided by the present invention uses a hash table to store the position coordinates of the matched video target and the position coordinates of the radar target, thereby effectively improving the speed of radar target mapping to the video coordinate system, simplifying the calibration target pair selection strategy, and improving the mapping speed.

[0166] Optionally, this embodiment provides a target calibration method, which is a further description of the above embodiment, mainly describing: a specific process of obtaining matching radar targets and video targets. The target calibration method provided in this embodiment includes:

[0167] Get the video image and radar image with the closest timestamp;

[0168] When the number of video targets in the video image and the number of radar targets in the radar image are the same, and the distance between the video targets in the video image and the distance between the radar targets in the radar image are both greater than a preset distance, the video targets in the video image are matched with the radar targets in the radar image to obtain the matched radar targets and video targets.

[0169] Specifically, please refer to Figure 3 , Figure 3 This is the third flow chart of the target calibration method provided by the present invention. The specific process of obtaining the matching radar target and video target includes the following steps:

[0170] Step 310: Acquire the video image and radar image with the closest timestamps.

[0171] The selection of video domain targets depends on the detection results of intelligent algorithms. Since video and radar are two systems, the detection speeds are also different. Usually, the radar image target detection speed is slightly faster than the video image target detection speed. Therefore, there is a situation where the timestamp of the video image is different from the timestamp of the radar image.

[0172] To solve this problem, a cache area can be set up to store radar detection results. When a new video image target detection result is generated, the timestamps of all radar images in the cache area are traversed according to the timestamp of the new video image to find the video image and radar image with the closest timestamp.

[0173] In some embodiments, in order to avoid the situation where the timestamp difference between the video image and the radar image with the closest timestamp is still relatively large, a preset time difference can be set. The timestamps of all radar images in the cache area are traversed to find the video image and radar image whose timestamp difference meets the preset time difference. If there are multiple radar images whose timestamp difference with the video image meets the preset time difference, the radar image with the timestamp closest to the timestamp of the video image is selected from the multiple radar images.

[0174] In actual monitoring scenarios, there may be situations where lane switching causes targets to overlap or become close, or targets turn and disappear, etc. Therefore, it is necessary to handle abnormal situations.

[0175] Step 320, determining whether the number of video targets in the video image is the same as the number of radar targets in the radar image.

[0176] In the case where the target disappears, it is determined whether the number of the video targets in the video image is the same as the number of the radar targets in the radar image.

[0177] If the number of video targets in the video image is the same as the number of radar targets in the radar image, it indicates that no target disappearance occurs, which is a normal situation, and step 330 is executed; if the number of video targets in the video image is different from the number of radar targets in the radar image, it indicates that target disappearance occurs, which is an abnormal situation, and step 350 is executed.

[0178] Step 330 , determining whether the distances between the video objects in the video image are all greater than a preset distance.

[0179] Step 340: determine whether the distances between radar targets in the radar image are all greater than a preset distance.

[0180] For the case where the targets are overlapping or close, it is determined whether the distances between the video targets in the video image are greater than the preset distances, and it is determined whether the distances between the radar targets in the radar image are greater than the preset distances.

[0181] If the distances between the video targets in the video image are greater than the preset distances, and the distances between the radar targets in the radar image are greater than the preset distances, it indicates that there is no target overlap or proximity, which is a normal situation, and step 360 is executed.

[0182] If the distance between the video targets in the video image is less than or equal to the preset distance, or the distance between the radar targets in the radar image is less than or equal to the preset distance, it indicates that the targets are overlapped or close, which is an abnormal situation, and step 350 is executed.

[0183] Step 350 , discard the video image and the radar image.

[0184] The video image and radar image with abnormal conditions are discarded, and step 310 is executed to reacquire the video image and radar image with the closest time stamp.

[0185] Step 360: Match the video target in the video image with the radar target in the radar image to obtain matched radar targets and video targets.

[0186] When both the video image and the radar image are normal, the video target in the video image is matched with the radar target in the radar image to obtain the matched radar target and video target. The matching rule may be to match the closest video target with the radar target.

[0187] The target calibration method provided by the present invention obtains a video image and a radar image with the closest timestamps, and matches the video target and the radar target when the video target in the video image and the radar target in the radar image meet preset requirements, which is beneficial to improving the matching accuracy of the video target and the radar target, and further beneficial to the accuracy of the subsequent calibration matrix.

[0188] Optionally, this embodiment provides a target calibration method, which is a further description of the above embodiment, mainly describing: matching a video target in a video image with a radar target in a radar image to obtain a specific process of matching radar targets and video targets. The target calibration method provided in this embodiment includes:

[0189] The video targets in the video image and the radar targets in the radar image are respectively sorted in order from top to bottom and from left to right based on the center of the image to obtain a video target sorting result and a radar target sorting result;

[0190] The video targets and radar targets at the same sorting position in the video target sorting result and the radar target sorting result are matched to obtain the matched radar targets and video targets.

[0191] Specifically, there may be a one-to-many match between video targets and radar targets, that is, there may be a situation where one video target can match multiple radar targets, or one radar target can match multiple video targets.

[0192] To avoid this situation, the video targets in the video image are sorted from top to bottom and from left to right based on the center of the image to obtain the video target sorting result; the radar targets in the radar image are sorted from top to bottom and from left to right based on the center of the image to obtain the radar target sorting result. Then, according to the video target sorting result and the radar target sorting result, the video targets and radar targets with the same sorting position are matched.

[0193] Figure 4 is a schematic diagram of a video target in a video image and a radar target in a radar image provided by the present invention, such as Figure 4 As shown, the video image includes: video target A, video target B, video target C and video target D, and the radar image includes: radar target a, radar target b, radar target c and radar target d.

[0194] Based on the center of the image, the targets are sorted from top to bottom and from left to right. The order of video targets is A->B->C->D, and the order of radar targets is a->b->c->d. Therefore, video target A is matched with radar target a, video target B is matched with radar target b, video target C is matched with radar target c, and video target D is matched with radar target d.

[0195] The target calibration method provided by the present invention further helps to improve the matching accuracy of the video target and the radar target by sorting the video target and the radar target in order from top to bottom and from left to right based on the center of the image, and matching the video target and the radar target according to the sorting result.

[0196] The target calibration device provided by the present invention is described below. The target calibration device described below and the target calibration method described above can be referenced to each other.

[0197] Figure 5 is a schematic diagram of the structure of the target calibration device provided by the present invention, such as Figure 5 As shown, the target calibration device provided by the present invention includes: a calibration module 510 and a correction module 520; wherein:

[0198] The calibration module 510 is used to map the current radar target to the video coordinate system corresponding to the video target based on the calibration matrix to obtain the current calibration target; the radar target is the target detected by the radar, and the video target is the target detected by the video monitoring device;

[0199] The correction module 520 is used to correct the position of the current calibration target when the position deviation between the current video target matched with the current radar target and the current calibration target is greater than a preset deviation threshold.

[0200] In some embodiments, the correction module 520 includes: a determination submodule, a fuzzy PID processing submodule and a correction submodule; wherein:

[0201] A determination submodule, used to determine the abscissa difference and ordinate difference between the position coordinates of the current video target and the position coordinates of the current calibration target;

[0202] A fuzzy PID processing submodule, used for performing fuzzy PID processing on the horizontal coordinate difference and the vertical coordinate difference respectively to obtain a horizontal coordinate correction amount and a vertical coordinate correction amount;

[0203] The correction submodule is used to correct the position of the current calibration target according to the horizontal coordinate correction amount and the vertical coordinate correction amount.

[0204] In some embodiments, the fuzzy PID processing submodule is specifically used to: determine the fuzzy subset corresponding to the horizontal coordinate difference, the fuzzy subset corresponding to the rate of change of the horizontal coordinate difference, the fuzzy subset corresponding to the vertical coordinate difference, and the fuzzy subset corresponding to the rate of change of the vertical coordinate difference according to a one-to-one correspondence between the difference range and the fuzzy subset;

[0205] Performing fuzzy PID processing based on the fuzzy subset corresponding to the horizontal coordinate difference and the fuzzy subset corresponding to the rate of change of the horizontal coordinate difference to obtain a first fuzzy PID coefficient;

[0206] Based on the first fuzzy PID coefficient and the horizontal coordinate difference, obtaining the horizontal coordinate correction amount;

[0207] Performing fuzzy PID processing based on the fuzzy subset corresponding to the ordinate difference and the fuzzy subset corresponding to the rate of change of the ordinate difference to obtain a second fuzzy PID coefficient;

[0208] The ordinate correction amount is obtained based on the second fuzzy PID coefficient and the ordinate difference.

[0209] In some embodiments, the device further comprises: an acquisition module, a storage module and a determination module; wherein:

[0210] An acquisition module, used to acquire matching radar targets and video targets;

[0211] A storage module, used for storing the position coordinates of the matched radar target and the position coordinates of the video target in a hash table;

[0212] The obtaining module is used to obtain the calibration matrix according to the position coordinates of the radar target and the position coordinates of the video target matched in the hash table when the radar target and the video target in the hash table meet the preset calibration matrix obtaining condition.

[0213] In some embodiments, each of the hash tables includes N hash buckets, each hash bucket includes Y empty slots, each empty slot is used to store an array, N is a positive integer greater than zero, Y is a positive integer greater than zero, and the product of N and Y is 10000;

[0214] The storage module is specifically used to: determine the empty slot corresponding to the radar target according to the ten-thousandth ratio of the position coordinate value of the radar target;

[0215] The ten-thousandth ratio of the position coordinate value of the video target that matches the radar target is stored in the empty slot corresponding to the radar target.

[0216] In some embodiments, the acquisition module includes: an acquisition submodule and a matching submodule; wherein:

[0217] The acquisition submodule is used to acquire the video image and radar image with the closest timestamp.

[0218] A matching submodule is used to match the video targets in the video image with the radar targets in the radar image to obtain the matched radar targets and video targets when the number of the video targets in the video image and the number of the radar targets in the radar image are the same and the distances between the video targets in the video image and the distances between the radar targets in the radar image are both greater than a preset distance.

[0219] In some embodiments, the matching submodule includes: a sorting unit and a matching unit; wherein:

[0220] A sorting unit, used to sort the video targets in the video image and the radar targets in the radar image in order from top to bottom and from left to right based on the center of the image, to obtain a video target sorting result and a radar target sorting result;

[0221] The matching unit is used to match the video targets and radar targets at the same sorting position in the video target sorting result and the radar target sorting result to obtain the matched radar targets and video targets.

[0222] It should be noted here that the above-mentioned target calibration device provided by the present invention can implement all the method steps implemented by the above-mentioned method embodiment, and can achieve the same technical effect. The parts and beneficial effects that are the same as the method embodiment in this embodiment will not be described in detail here.

[0223] Figure 6 It is a structural schematic diagram provided by the present invention, such as Figure 6 As shown, the system may include: a processor 610, a communication interface 620, a memory 630 and a communication bus 640, wherein the processor 610, the communication interface 620 and the memory 630 communicate with each other through the communication bus 640. The processor 610 may call the logic instructions in the memory 630 to execute the target calibration method, which includes: based on the calibration matrix, mapping the current radar target to the video coordinate system corresponding to the video target to obtain the current calibration target; the radar target is the target detected by the radar, and the video target is the target detected by the video surveillance device; when the position deviation between the current video target matching the current radar target and the current calibration target is greater than the preset deviation threshold, the position of the current calibration target is corrected.

[0224] In addition, the logic instructions in the above-mentioned memory 630 can be implemented in the form of a software functional unit and can be stored in a computer-readable storage medium when it is sold or used as an independent product. Based on such an understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art or the part of the technical solution, can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk and other media that can store program codes.

[0225] On the other hand, the present invention also provides a computer program product, which includes a computer program. The computer program can be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer can execute the target calibration method provided by the above methods, the method including: based on a calibration matrix, mapping the current radar target to a video coordinate system corresponding to the video target to obtain a current calibrated target; the radar target is a target detected by a radar, and the video target is a target detected by a video surveillance device; when the position deviation between the current video target matching the current radar target and the current calibrated target is greater than a preset deviation threshold, correcting the position of the current calibrated target.

[0226] On the other hand, the present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, is implemented to execute the target calibration method provided by the above-mentioned methods, the method comprising: based on a calibration matrix, mapping the current radar target to a video coordinate system corresponding to the video target to obtain a current calibrated target; the radar target is a target detected by a radar, and the video target is a target detected by a video surveillance device; when the position deviation between the current video target matching the current radar target and the current calibrated target is greater than a preset deviation threshold, correcting the position of the current calibrated target.

[0227] The device embodiments described above are merely illustrative, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. Those of ordinary skill in the art may understand and implement it without creative effort.

[0228] Through the description of the above implementation methods, those skilled in the art can clearly understand that each implementation method can be implemented by means of software plus a necessary general hardware platform, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solution is essentially or the part that contributes to the prior art can be embodied in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a disk, an optical disk, etc., including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.

[0229] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A target calibration method, characterized in that: include: Based on the calibration matrix, the current radar target is mapped to the video coordinate system corresponding to the video target to obtain the current calibration target; The radar target is a target detected by a radar, and the video target is a target detected by a video surveillance device; When a position deviation between a current video target that matches the current radar target and the current calibration target is greater than a preset deviation threshold, the position of the current calibration target is corrected.

2. The target calibration method according to claim 1, characterized in that: The correcting the position of the current calibration target includes: Determine the abscissa difference and ordinate difference between the position coordinates of the current video target and the position coordinates of the current calibration target; Performing fuzzy PID processing on the horizontal coordinate difference and the vertical coordinate difference respectively to obtain a horizontal coordinate correction amount and a vertical coordinate correction amount; The position of the current calibration target is corrected according to the horizontal coordinate correction amount and the vertical coordinate correction amount.

3. The target calibration method according to claim 2, characterized in that: The fuzzy PID processing is performed on the horizontal coordinate difference and the vertical coordinate respectively to obtain the horizontal coordinate correction amount and the vertical coordinate correction amount, including: According to a one-to-one correspondence between the difference range and the fuzzy subset, determine the fuzzy subset corresponding to the horizontal coordinate difference, the fuzzy subset corresponding to the rate of change of the horizontal coordinate difference, the fuzzy subset corresponding to the vertical coordinate difference, and the fuzzy subset corresponding to the rate of change of the vertical coordinate difference; Performing fuzzy PID processing based on the fuzzy subset corresponding to the horizontal coordinate difference and the fuzzy subset corresponding to the rate of change of the horizontal coordinate difference to obtain a first fuzzy PID coefficient; Based on the first fuzzy PID coefficient and the horizontal coordinate difference, obtaining the horizontal coordinate correction amount; Performing fuzzy PID processing based on the fuzzy subset corresponding to the ordinate difference and the fuzzy subset corresponding to the rate of change of the ordinate difference to obtain a second fuzzy PID coefficient; The ordinate correction amount is obtained based on the second fuzzy PID coefficient and the ordinate difference.

4. The target calibration method according to claim 1, characterized in that: Before mapping the current radar target to the video coordinate system corresponding to the video target based on the calibration matrix to obtain the current calibration target, the method further includes: Acquire matching radar targets and video targets; The position coordinates of the matched radar target and the position coordinates of the video target are stored in a hash table; When the radar targets and video targets in the hash table meet the preset calibration matrix obtaining conditions, the calibration matrix is ​​obtained according to the position coordinates of the matching radar targets and video targets in the hash table.

5. The target calibration method according to claim 4, characterized in that: Each of the hash tables includes N hash buckets, each hash bucket includes Y empty slots, each empty slot is used to store an array, N is a positive integer greater than zero, Y is a positive integer greater than zero, and the product of N and Y is 10000; The storing of the position coordinates of the matched radar target and the position coordinates of the video target in a hash table includes: Determining an empty slot corresponding to the radar target according to the ten-thousandth ratio of the position coordinate value of the radar target; The ten-thousandth ratio of the position coordinate value of the video target that matches the radar target is stored in the empty slot corresponding to the radar target.

6. The target calibration method according to claim 4, characterized in that: The obtaining of matched radar targets and video targets includes: Get the video image and radar image with the closest timestamp; When the number of video targets in the video image and the number of radar targets in the radar image are the same, and the distance between the video targets in the video image and the distance between the radar targets in the radar image are both greater than a preset distance, the video targets in the video image are matched with the radar targets in the radar image to obtain the matched radar targets and video targets.

7. The target calibration method according to claim 6, characterized in that: The step of matching the video target in the video image with the radar target in the radar image to obtain the matched radar target and video target includes: The video targets in the video image and the radar targets in the radar image are respectively sorted in order from top to bottom and from left to right based on the center of the image to obtain a video target sorting result and a radar target sorting result; The video targets and radar targets at the same sorting position in the video target sorting result and the radar target sorting result are matched to obtain the matched radar targets and video targets.

8. A target calibration device, characterized in that: include: A calibration module is used to map the current radar target to the video coordinate system corresponding to the video target based on the calibration matrix to obtain the current calibration target; The radar target is a target detected by a radar, and the video target is a target detected by a video surveillance device; The correction module is used to correct the position of the current calibration target when the position deviation between the current video target matching the current radar target and the current calibration target is greater than a preset deviation threshold.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the computer program, the target calibration method according to any one of claims 1 to 7 is implemented.

10. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the target calibration method according to any one of claims 1 to 7 is implemented.

Images