Target detection method, device and vehicle
By dynamically determining the target judgment threshold, combining the perceived data of smart cameras and intelligent radars and vehicle speed, the problem of low target detection accuracy in assisted driving is solved, and detection accuracy is improved.
Patent Information
- Application Number
- CN202210588133.X
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2022-05-27
- Publication Date
- 2024-11-05
- Estimated Expiration
- 2042-05-27
AI Technical Summary
In the existing assisted driving technology, the target detection accuracy is low and it is impossible to reasonably reflect the real environment and the error changes of sensors in different scenarios.
By obtaining the vehicle's perceived data set, including the perception data of the smart camera and intelligent radar and the vehicle speed, the target judgment threshold is dynamically determined, which is used to judge whether the two perceived targets are the same and avoid the limitation of the fixed distance threshold.
It improves the accuracy of target detection, can better reflect the real environment and sensor error changes in different scenarios, and achieves accurate detection of targets.
Smart Images

Figure CN115079155B_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of assisted driving, and in particular to a target detection method, device and vehicle. Background Art
[0002] At present, assisted driving technology is developing rapidly. Assisted driving functions are gradually becoming standard features in mid-to-high-end models at home and abroad. They are mainly equipped with intelligent monocular cameras and intelligent millimeter-wave radars for forward target detection. After target data association, fusion processing, path planning and vehicle control, the vehicle is kept safe. However, existing technologies mostly use fixed distance thresholds for target association judgment, which cannot reasonably reflect the real environment and the error changes of sensors in different scenarios.
[0003] To address the above-mentioned problems, no effective solution has been proposed yet. Summary of the invention
[0004] The embodiments of the present invention provide a target detection method, device and vehicle to at least solve the technical problem of low target detection accuracy in the related art.
[0005] According to one aspect of an embodiment of the present invention, a target detection method is provided, including: acquiring a perception data set of a current vehicle, wherein the perception data set includes: first perception data perceived by a first sensor, second perception data perceived by a second sensor, and a vehicle speed; determining a target judgment threshold corresponding to the current vehicle based on the perception data set, wherein the target judgment threshold is used to determine whether a first target perceived by the first sensor is the same as a second target perceived by the second sensor; comparing the first perception data and the second perception data with the target judgment threshold to obtain detection results of the first sensor and the second sensor.
[0006] Optionally, based on the perception data set, a target judgment threshold corresponding to the current vehicle is determined, including: based on the perception data set, determining the target detection condition corresponding to the current vehicle; obtaining perception data corresponding to the target detection condition from a preset mapping relationship, wherein the preset mapping relationship is used to characterize the data distribution corresponding to different detection conditions; based on the target data distribution, determining the target judgment threshold.
[0007] Optionally, based on the perception data set, the target detection conditions corresponding to the current vehicle are determined, including: based on the first perception data, determining the target category of the first target and the target longitudinal distance between the current vehicle and the first target; based on the target category, vehicle speed and target longitudinal distance, determining the target detection conditions.
[0008] Optionally, the method also includes: obtaining third perception data and fourth perception data obtained by the first sensor and the second sensor respectively sensing the same third target under different detection conditions; obtaining the difference between the third perception data and the fourth perception data to obtain a first difference; determining the data distribution of the first difference under different detection conditions; and generating a preset mapping relationship based on different detection conditions and the data distribution corresponding to the different detection conditions.
[0009] Optionally, based on the target data distribution, determining the target judgment threshold includes: in response to the first detection of the first sensor and the second sensor in the current cycle, determining the target judgment threshold as the sum of the mean and the mean square error of the target data distribution; in response to the first sensor and the second sensor being detected not for the first time in the current cycle, and the first target and the second target being the same in a cycle before the current cycle, determining the target judgment threshold as the sum of the mean and a first product, the first product being the product of the mean square error and a first preset value; in response to the first sensor and the second sensor being detected not for the first time in the current cycle, and the first target and the second target being the same in the two cycles before the current cycle, determining the target judgment threshold as the sum of the mean and the second product, the second product being the product of the mean square error and the second preset value.
[0010] Optionally, the first perception data and the second perception data are compared with the target judgment threshold to obtain detection results of the first sensor and the second sensor, including: obtaining the difference between the first perception data and the second perception data to obtain the second difference; comparing the second difference with the target judgment threshold to obtain the detection result.
[0011] Optionally, the second difference is compared with the target judgment threshold to obtain a detection result, including: in response to the second difference being less than or equal to the target judgment threshold, determining that the detection result is that the first target and the second target are the same; in response to the second difference being greater than the target judgment threshold, determining that the detection result is that the first target and the second target are different.
[0012] Optionally, the target judgment threshold includes at least one of the following: a longitudinal distance threshold, a lateral distance threshold, a longitudinal speed threshold and a lateral speed threshold.
[0013] According to another aspect of an embodiment of the present invention, a target detection device is also provided, including: a first acquisition module, used to acquire a perception data set of a current vehicle, wherein the perception data set includes: first perception data perceived by a first sensor, second perception data perceived by a second sensor, and a vehicle speed; a first determination module, used to determine a target judgment threshold corresponding to the current vehicle based on the perception data set, wherein the target judgment threshold is used to determine whether a first target perceived by the first sensor is the same as a second target perceived by the second sensor; a judgment module, used to compare the first perception data and the second perception data with the target judgment threshold to obtain detection results of the first sensor and the second sensor, and the detection results are used to characterize whether the first sensor and the second sensor are successfully associated.
[0014] According to another aspect of an embodiment of the present invention, a target vehicle is also provided, comprising: one or more processors; a storage device for storing one or more programs; when the one or more programs are executed by one or more processors, the one or more processors execute any one of the above-mentioned target detection methods.
[0015] According to another aspect of an embodiment of the present invention, a non-volatile storage medium is provided. The non-volatile storage medium includes a stored program, wherein when the program is executed, a processor of a device is controlled to execute any one of the above-mentioned target detection methods.
[0016] In an embodiment of the present invention, a method is adopted to obtain a perception data set of the current vehicle, wherein the perception data set includes: first perception data perceived by a first sensor, second perception data perceived by a second sensor, and vehicle speed; based on the perception data set, a target judgment threshold corresponding to the current vehicle is determined, wherein the target judgment threshold is used to determine whether the first target perceived by the first sensor is the same as the second target perceived by the second sensor; the first perception data and the second perception data are compared with the target judgment threshold to obtain the detection results of the first sensor and the second sensor. It is easy to notice that the target judgment threshold is determined based on the perception data set acquired in real time, and is no longer a fixed distance threshold, but a dynamically changing threshold, so that the target judgment threshold can reflect the actual environmental conditions and the error changes of the sensors in different scenarios, thereby achieving the purpose of being able to accurately detect the target, thereby achieving the technical effect of improving the accuracy of target detection, and thus solving the technical problem of low target detection accuracy in related technologies. BRIEF DESCRIPTION OF THE DRAWINGS
[0017] The drawings described herein are used to provide a further understanding of the present invention and constitute a part of this application. The exemplary embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute an improper limitation of the present invention. In the drawings:
[0018] Figure 1 is a flow chart of a target detection method according to an embodiment of the present invention;
[0019] Figure 2 is a flow chart of an optional target detection method according to an embodiment of the present invention;
[0020] Figure 3 is a schematic structural diagram of an optional target detection device according to an embodiment of the present invention. DETAILED DESCRIPTION
[0021] In order to enable those skilled in the art to better understand the scheme of the present invention, the technical scheme in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present invention.
[0022] It should be noted that the terms "first", "second", etc. in the specification and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments of the present invention described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.
[0023] Example 1
[0024] According to an embodiment of the present invention, a method embodiment of target detection is provided. It should be noted that the steps shown in the flowchart of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and although a logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in an order different from that shown here.
[0025] Figure 1 is a flow chart of a target detection method according to an embodiment of the present invention. Figure 1 As shown, the method comprises the following steps:
[0026] Step S102, obtaining a perception data set of the current vehicle, wherein the perception data set includes: first perception data perceived by a first sensor, second perception data perceived by a second sensor, and vehicle speed.
[0027] The above-mentioned first sensor can be any intelligent camera sensor on a vehicle with an assisted driving function. In this embodiment, an intelligent monocular camera sensor is taken as an example for illustration; the second sensor can be any intelligent radar sensor on a vehicle with an assisted driving function. In this embodiment, an intelligent millimeter-wave radar sensor is taken as an example for illustration; the first perception data can be the first target category, the longitudinal distance, lateral distance, longitudinal speed, lateral speed and other data between the current vehicle and the first target detected in real time by the first sensor; the second perception data can be the second target category, the longitudinal distance, lateral distance, longitudinal speed, lateral speed and other data between the current vehicle and the second target detected in real time by the second sensor; the vehicle speed can be the real-time speed of the current vehicle.
[0028] In an optional embodiment, a perception data set of the current vehicle can be first obtained, wherein the perception data set includes: first perception data perceived by a first sensor, such as data such as the longitudinal distance, lateral distance, longitudinal speed, and lateral speed to the target; second perception data perceived by a second sensor, such as data such as the longitudinal distance, lateral distance, longitudinal speed, and lateral speed to the target; and the real-time speed of the current vehicle.
[0029] Step S104, based on the perception data set, determining a target judgment threshold corresponding to the current vehicle, wherein the target judgment threshold is used to determine whether a first target perceived by the first sensor is the same as a second target perceived by the second sensor.
[0030] The above-mentioned target judgment threshold can be set in advance by the user based on a preset mapping relationship, and is used to determine whether the targets detected by the first sensor and the second sensor are the same target. The target judgment threshold includes at least one of the following: a longitudinal distance threshold, a lateral distance threshold, a longitudinal speed threshold and a lateral speed threshold, wherein the longitudinal distance threshold, the lateral distance threshold, the longitudinal speed threshold and the lateral speed threshold are obtained by the mean, the mean square error and the preset value of the Gaussian distribution of the difference between the third perception data and the fourth perception data; the first target can be the target perceived by the first sensor; the second target can be the target perceived by the second sensor.
[0031] The above-mentioned preset mapping relationship can be a data set set in advance by the user that can reflect the correspondence between the target, vehicle speed, longitudinal distance of the first perception data and the longitudinal distance difference, lateral distance difference, longitudinal speed difference and lateral speed difference between the third perception data and the fourth perception data, including but not limited to the target, vehicle speed, longitudinal distance of the third perception data and the longitudinal distance difference, lateral distance difference, longitudinal speed difference and lateral speed difference between the third perception data and the fourth perception data.
[0032] In an optional embodiment, the specific conditions under which the current vehicle is in a preset mapping relationship can be first determined through the target category, the longitudinal distance of the first perception data in the perception data set, and the current vehicle speed, and then the corresponding judgment threshold in the corresponding condition can be obtained to obtain the target judgment threshold.
[0033] Step S106, comparing the first perception data and the second perception data with the target judgment threshold to obtain detection results of the first sensor and the second sensor.
[0034] In an optional embodiment, first, the difference between the first perception data and the second perception data in the perception data set can be obtained, and then the difference is fitted to obtain the mean and mean square error of the Gaussian distribution, and then based on the mean and mean square error of the Gaussian distribution, the real-time longitudinal distance difference, lateral distance difference, longitudinal speed difference and lateral speed difference between the first perception data and the second perception data are obtained, and finally the real-time longitudinal distance difference, lateral distance difference, longitudinal speed difference and lateral speed difference are compared with the target judgment threshold to obtain whether the first target is the same as the second target.
[0035] In an embodiment of the present invention, a method is adopted to obtain a perception data set of the current vehicle, wherein the perception data set includes: first perception data perceived by a first sensor, second perception data perceived by a second sensor, and vehicle speed; based on the perception data set, a target judgment threshold corresponding to the current vehicle is determined, wherein the target judgment threshold is used to determine whether the first target perceived by the first sensor is the same as the second target perceived by the second sensor; the first perception data and the second perception data are compared with the target judgment threshold to obtain the detection results of the first sensor and the second sensor. It is easy to notice that the target judgment threshold is determined based on the perception data set acquired in real time, and is no longer a fixed distance threshold, but a dynamically changing threshold, so that the target judgment threshold can reflect the actual environmental conditions and the error changes of the sensors in different scenarios, thereby achieving the purpose of being able to accurately detect the target, thereby achieving the technical effect of improving the accuracy of target detection, and thus solving the technical problem of low target detection accuracy in related technologies.
[0036] Optionally, based on the perception data set, a target judgment threshold corresponding to the current vehicle is determined, including: based on the perception data set, determining the target detection condition corresponding to the current vehicle; obtaining perception data corresponding to the target detection condition from a preset mapping relationship, wherein the preset mapping relationship is used to characterize the data distribution corresponding to different detection conditions; based on the target data distribution, determining the target judgment threshold.
[0037] The above-mentioned target detection conditions may be conditions in a preset mapping relationship that match the current target of the current vehicle, the vehicle speed, and the longitudinal distance of the first perception data. For example, the preset target may be the same as the current target of the current vehicle, the preset vehicle speed may be the same as the vehicle speed of the current vehicle, and the longitudinal distance in the third perception data may be the same as the longitudinal distance in the first perception data.
[0038] The above-mentioned target data distribution can be any distribution that can reflect the data relationship between the longitudinal distance difference, lateral distance difference, longitudinal speed difference and lateral speed difference between the third perception data and the fourth perception data. Each target detection condition corresponds to a target data distribution, and there is no specific limitation. In this embodiment, in order to reduce the amount of data calculation, Gaussian distribution is taken as an example for illustration. For example, the mean and variance of the longitudinal distance difference Gaussian distribution, the mean and variance of the lateral distance difference Gaussian distribution, the mean and variance of the longitudinal speed difference Gaussian distribution and the mean and variance of the lateral speed difference Gaussian distribution of the third perception data and the fourth perception data can be obtained.
[0039] In an optional embodiment, based on the perception data set, determining the target judgment threshold corresponding to the current vehicle includes, first, determining the matching conditions (i.e., target detection conditions) in a preset mapping relationship based on the target category, vehicle speed, and longitudinal distance in the first perception data in the perception data set, and secondly obtaining the perception data perceived by the third sensor and the fourth sensor corresponding to the target detection conditions in the preset mapping relationship, obtaining the Gaussian distribution mean and mean square error (i.e., target data distribution) of the difference based on the perception data difference, and finally obtaining the target judgment threshold based on the mean, mean square error, and preset value.
[0040] Optionally, based on the perception data set, the target detection conditions corresponding to the current vehicle are determined, including: based on the first perception data, determining the target category of the first target and the target longitudinal distance between the current vehicle and the first target; based on the target category, vehicle speed and target longitudinal distance, determining the target detection conditions.
[0041] In an optional embodiment, the first perception data may include data such as the first target category detected in real time by the first sensor, the longitudinal distance, lateral distance, longitudinal speed, and lateral speed between the current vehicle and the first target. Based on the first perception data, the target category of the first target and the target longitudinal distance between the current vehicle and the first target can be determined. Based on the target category, the real-time speed of the current vehicle, and the real-time longitudinal distance to the target, the conditions that are consistent with the target category, the real-time speed of the current vehicle, and the real-time longitudinal distance to the target in the preset mapping relationship can be determined, that is, the target detection conditions.
[0042] Optionally, the method also includes: obtaining third perception data and fourth perception data obtained by the first sensor and the second sensor respectively sensing the same third target under different detection conditions; obtaining the difference between the third perception data and the fourth perception data to obtain a first difference; determining the data distribution of the first difference under different detection conditions; and generating a preset mapping relationship based on different detection conditions and the data distribution corresponding to the different detection conditions.
[0043] The third target mentioned above may be the same target perceived by the first sensor and the second sensor selected by the user when establishing the preset mapping relationship. There is no specific limitation. In this embodiment, pedestrians (including two-wheelers and bicycles), cars, and trucks are taken as examples for illustration.
[0044] The above-mentioned third perception data may be data perceived by the first sensor of the third target, and may include data such as the third target category detected by the first sensor, and the longitudinal distance, lateral distance, longitudinal speed, and lateral speed between the vehicle and the third target; the fourth perception data may be data perceived by the second sensor of the third target, and may include data such as the third target category detected by the second sensor, and the longitudinal distance, lateral distance, longitudinal speed, and lateral speed between the vehicle and the third target.
[0045] The above-mentioned different detection conditions can be 192 different conditions generated based on different target categories, different vehicle speed ranges and different ranges of longitudinal distances between the vehicle and the third target in the third perception data, among which the vehicle speed range is not specifically limited. In this embodiment, the four categories of 0-40km / h, 40-80km / h, 80-120km / h and greater than 120km / h are taken as examples for explanation. The longitudinal distance range between the vehicle and the third target is not specifically limited. In this embodiment, the 16 ranges of 0-10m, 10-20m, 20-30m, 30-40m, 40-50m, 50-60m, 60-70m, 70-80m, 80-90m, 90-100m, 100-110m, 110-120m, 120-130m, 130-140m, 140-150m and greater than 150m are taken as examples.
[0046] The above-mentioned first difference may be the longitudinal distance difference, lateral distance difference, longitudinal speed difference, and lateral speed difference between the third perception data and the fourth perception data.
[0047] The above-mentioned data distribution can be obtained by fitting the first difference to obtain the mean and variance of the Gaussian distribution of the longitudinal distance difference, the mean and variance of the Gaussian distribution of the lateral distance difference, the mean and variance of the Gaussian distribution of the longitudinal speed difference, and the mean and variance of the Gaussian distribution of the lateral speed difference. Among them, because the target category, vehicle speed and the longitudinal distance between the vehicle and the third target will change, the first difference will also change accordingly. Therefore, there will be a corresponding data distribution under each detection condition.
[0048] In an optional embodiment, first, the third perception data and the fourth perception data obtained by respectively sensing the third target with the first sensor and the second sensor under different detection conditions can be obtained, and then the difference between the third perception data and the fourth perception data can be obtained to obtain the first difference. After fitting the first difference, the Gaussian distribution mean and variance (i.e., data distribution) of the first difference can be obtained. Because each detection condition corresponds to a data distribution, a preset mapping relationship can be generated based on different detection conditions and the data distributions corresponding to the different detection conditions.
[0049] Optionally, based on the target data distribution, determining the target judgment threshold includes: in response to the first detection of the first sensor and the second sensor in the current cycle, determining the target judgment threshold as the sum of the mean and the mean square error of the target data distribution; in response to the first sensor and the second sensor being detected not for the first time in the current cycle, and the first target and the second target being the same in a cycle before the current cycle, determining the target judgment threshold as the sum of the mean and a first product, the first product being the product of the mean square error and a first preset value; in response to the first sensor and the second sensor being detected not for the first time in the current cycle, and the first target and the second target being the same in the two cycles before the current cycle, determining the target judgment threshold as the sum of the mean and the second product, the second product being the product of the mean square error and the second preset value.
[0050] The above-mentioned period can be a period for judgment set in advance by the user. The specific value can be set according to user needs and is not specifically limited in this implementation; the above-mentioned mean can be represented by μ; the above-mentioned mean square error can be σ; the first preset value can be a constant value set by the user to increase the judgment threshold. The specific value is not limited. In this embodiment, 2 is taken as an example for explanation; the first product can be 2σ; the second preset value can be a constant value set by the user to increase the judgment threshold. The specific value is not limited. In this embodiment, 3 is taken as an example for explanation; the second product can be 3σ.
[0051] In an optional embodiment, if the first sensor and the second sensor are detected for the first time in the current cycle, the target judgment threshold is determined to be μ+σ (i.e., the sum of the mean and the mean square error of the target data distribution); if it is not the first time that the first sensor and the second sensor are detected in the current cycle, and the first target and the second target are the same in the previous cycle, the target judgment threshold is determined to be μ+2σ (i.e., the sum of the mean and the first product); if it is not the first time that the first sensor and the second sensor are detected in the current cycle, and the first target and the second target are the same in the previous two cycles, the target judgment threshold is determined to be μ+3σ (i.e., the sum of the mean and the second product).
[0052] Optionally, the first perception data and the second perception data are compared with the target judgment threshold to obtain detection results of the first sensor and the second sensor, including: obtaining the difference between the first perception data and the second perception data to obtain the second difference; comparing the second difference with the target judgment threshold to obtain the detection result.
[0053] The above-mentioned second difference can be obtained by fitting the difference between the first perception data and the second perception data to obtain the Gaussian distribution mean and mean square error of the difference, and then the difference between the longitudinal distance difference, lateral distance difference, lateral speed difference, and the sum of the mean and mean square error of the lateral speed difference, including: longitudinal distance difference XD, lateral distance difference YD, longitudinal speed difference XV, and lateral speed difference YV.
[0054] In an optional embodiment, if the first sensor and the second sensor are detected for the first time in the current cycle, the longitudinal distance difference XD, the lateral distance difference YD, the longitudinal speed difference XV, and the lateral speed difference YV (i.e., the second difference) are compared with the target judgment threshold μ+σ to obtain the detection result.
[0055] In another optional embodiment, if the first sensor and the second sensor are not detected for the first time in the current cycle, and the first target and the second target are the same in the previous cycle, the longitudinal distance difference XD, the lateral distance difference YD, the longitudinal speed difference XV, and the lateral speed difference YV (i.e., the second difference) are compared with the target judgment threshold μ+2σ to obtain the detection result.
[0056] In another optional embodiment, if the current cycle is not the first time that the first sensor and the second sensor are detected, and the first target is the same as the second target in the previous two cycles, the longitudinal distance difference XD, the lateral distance difference YD, the longitudinal speed difference XV, and the lateral speed difference YV (i.e., the second difference) are compared with the target judgment threshold μ+3σ to obtain the detection result.
[0057] Optionally, the second difference is compared with the target judgment threshold to obtain a detection result, including: in response to the second difference being less than or equal to the target judgment threshold, determining that the detection result is that the first target and the second target are the same; in response to the second difference being greater than the target judgment threshold, determining that the detection result is that the first target and the second target are different.
[0058] Optionally, the target judgment threshold includes at least one of the following: a longitudinal distance threshold, a lateral distance threshold, a longitudinal speed threshold and a lateral speed threshold.
[0059] In an optional embodiment, when the longitudinal distance difference XD is less than or equal to the longitudinal distance threshold, and the lateral distance difference YD is less than or equal to the lateral distance threshold, and the longitudinal speed difference XV is less than or equal to the longitudinal speed threshold, and the lateral speed difference YV is less than or equal to the lateral speed threshold, the detection result is determined to be that the first target is the same as the second target.
[0060] In another optional embodiment, when the longitudinal distance difference XD is greater than the longitudinal distance threshold, and the lateral distance difference YD is greater than the lateral distance threshold, and the longitudinal speed difference XV is greater than the longitudinal speed threshold, and the lateral speed difference YV is greater than the lateral speed threshold, the detection result is determined to be that the first target is different from the second target.
[0061] Combine the following Figure 2 This embodiment is further described.
[0062] The purpose of the present invention is to overcome the shortcomings and deficiencies in the prior art and to propose a method for associating data of an assisted driving intelligent monocular camera and an intelligent millimeter-wave radar, which is suitable for assisted driving vehicles on structured roads such as highways and urban expressways.
[0063] The above-mentioned data association method includes two links, namely offline sensor difference statistics and online target association; the offline sensor difference statistics link refers to the statistics of the perception differences between the intelligent monocular camera and the intelligent millimeter-wave radar under different conditions; the different conditions in the offline sensor difference statistics link are composed of target category, vehicle speed range, and longitudinal distance range of the camera detection target.
[0064] In the offline sensor difference statistics link, under each of the above-mentioned different conditions, the perception results of the camera and millimeter wave for the same target are tested and recorded multiple times, and the Gaussian distributions of the longitudinal distance difference, lateral distance difference, longitudinal speed difference, and lateral speed difference of the two sensors under different conditions are fitted respectively, and then the mean and variance of the Gaussian distribution are obtained.
[0065] The above-mentioned online target association link determines whether the perceived target of the intelligent monocular camera and the perceived target of the intelligent millimeter-wave radar are the same target; the judgment method of online target association is to determine whether the longitudinal distance difference, lateral distance difference, longitudinal speed difference, and lateral speed difference of the targets perceived by the two sensors are within the judgment threshold; the judgment method of online target association first determines which condition the current state is in in the offline sensor difference statistics link, and sets the corresponding judgment threshold according to the mean and variance of the Gaussian distribution of the longitudinal distance difference, lateral distance difference, longitudinal speed difference, and lateral speed difference tested and fitted under this condition, and the judgment threshold is dynamically adjusted according to the number of associations.
[0066] The overall steps of the present invention include two links, namely offline sensor difference statistics and online target association.
[0067] The implementation steps of offline sensor difference statistics are as follows:
[0068] First, as shown in Table 1, according to the target category, different vehicle speed ranges, and longitudinal distance ranges of camera detection targets, 192 different conditions are combined. Among them, the camera detection targets are subdivided into three categories: pedestrians (including two-wheeled vehicles and bicycles), cars, and large vehicles; the longitudinal distance range of camera detection targets is subdivided into 16 categories: 0-10m, 10-20m, 20-30m, 30-40m, 40-50m, 50-60m, 60-70m, 70-80m, 80-90m, 90-100m, 100-110m, 110-120m, 120-130m, 130-140m, 140-150m, and above 150m; the vehicle speed range is subdivided into four categories: 0-40km / h, 40-80km / h, 80-120km / h, and above 120km / h.
[0069] Table 1 Schematic diagram of offline sensor difference statistics results
[0070]
[0071] Secondly, the intelligent monocular camera and intelligent millimeter-wave radar collected perception data of the same target under 192 different conditions. The perception data included longitudinal distance, lateral distance, longitudinal speed, and lateral speed, and the number of collection times was 80-100.
[0072] Secondly, the Gaussian distribution of the longitudinal distance difference, lateral distance difference, longitudinal velocity difference, and lateral velocity difference of the two sensors' acquisition results for 80-100 times is fitted.
[0073] Finally, the mean μxd-1-μxd-192 and variance σxd-1-σxd-192 of the Gaussian distribution of the longitudinal distance difference, the mean μyd-1-μyd-192 and variance σxv-1-σxv-192 of the Gaussian distribution of the lateral distance difference, the mean μxv-1-μxv-192 and variance σxv-1-σxv-192 of the Gaussian distribution of the longitudinal velocity difference, and the mean μyv-1-μyv-192 and variance σyv-1-σyv-192 of the Gaussian distribution of the lateral velocity difference under 192 different conditions were obtained.
[0074] The steps for implementing online data association are as follows:
[0075] First, obtain the perception data of the intelligent monocular camera, intelligent millimeter-wave radar and ego vehicle in this cycle.
[0076] Secondly, according to the target category and longitudinal distance in the camera perception data and the vehicle speed in the vehicle perception data, it is determined which condition the current state is in in the offline sensor difference statistics, and then the longitudinal distance difference Gaussian distribution mean μxd and variance σxd, the lateral distance difference Gaussian distribution mean μyd and variance σyd, the longitudinal speed difference Gaussian distribution mean μxv and variance σxv, and the lateral speed difference Gaussian distribution mean μyv and variance σyv tested and fitted under this condition are obtained.
[0077] Secondly, set the judgment threshold for this cycle. There are 4 judgment thresholds in total, namely the longitudinal distance difference threshold, the lateral distance difference threshold, the longitudinal speed difference threshold, and the lateral speed difference threshold. When the first association judgment is made, the longitudinal distance difference threshold is "μxd+σxd", the lateral distance difference threshold is "μyd+σyd", the lateral speed difference threshold is "μxv+σxv", and the lateral speed difference threshold is "μyv+σyv".
[0078] Secondly, calculate the longitudinal distance difference XD, transverse distance difference YD, longitudinal speed difference XV, and transverse speed difference YV between the camera and the millimeter wave sensing target in this cycle. If the longitudinal distance difference XD is less than the longitudinal distance difference threshold "μxd+σxd", and the transverse distance difference YD is less than the transverse distance difference threshold "μyd+σyd", and the longitudinal speed difference YV is less than the transverse speed difference threshold "μxv+σxv", and the transverse speed difference YV is less than the transverse speed difference threshold "μyv+σyv", then the target association in this cycle is successful.
[0079] If it is not the first association judgment, and the perception targets of the two sensors in the previous cycle are successfully associated, the four judgment thresholds are increased to "μ+2σ" respectively.
[0080] If it is not the first association judgment, and the perception targets of the two sensors in the first two cycles are successfully associated, the four judgment thresholds are increased to "μ+3σ" respectively.
[0081] Figure 2 is a flow chart of an optional target detection method according to an embodiment of the present invention. Figure 2 As shown, the method comprises the following steps:
[0082] Step S201: dividing 192 conditions according to target category, vehicle speed range, and longitudinal distance range of the camera to detect the target;
[0083] Step S202: sequentially completing the sensing data collection of the same target by two sensors under 192 different conditions;
[0084] Step S203: fitting the Gaussian distribution of the longitudinal distance difference, the lateral distance difference, the longitudinal speed difference, and the lateral speed difference under each condition respectively;
[0085] Step S204: obtaining the mean μ and variance σ of the Gaussian distribution in step S203, and constructing a preset mapping relationship;
[0086] Step S205: Acquire the first sensing data, the second sensing data and the current vehicle speed of the current vehicle;
[0087] Step S206: determining the target detection condition corresponding to the current state based on the preset mapping relationship;
[0088] Step S207: Determine whether it is the first time to detect the first perception data and the second perception data, if so, proceed to step S208, if not, proceed to step S209;
[0089] Step S208: the target judgment threshold is μ+σ;
[0090] Step S209: Determine whether the first target and the second target of the first two cycles are the same, if so, proceed to step S210, if not, proceed to step S211;
[0091] Step S210: increasing the target judgment threshold to μ+3σ;
[0092] Step S211: increasing the target judgment threshold to μ+2σ;
[0093] Step S212: determine whether the second difference is less than or equal to the target determination threshold, if so, proceed to step S213;
[0094] Step S213: The first target is the same as the second target, and the step ends.
[0095] The purpose of the present invention is to overcome the shortcomings and deficiencies in the prior art and to propose a method for associating data of an assisted driving intelligent monocular camera and an intelligent millimeter-wave radar, which is suitable for assisted driving vehicles on structured roads such as highways and urban expressways.
[0096] Example 2
[0097] According to another aspect of an embodiment of the present invention, a target detection device is also provided, which can execute the target detection method provided in the above embodiment 1. The specific implementation method and preferred application scenario are the same as those of the above embodiment 1 and will not be repeated here.
[0098] Figure 3 is a schematic diagram of the structure of a target detection device according to an embodiment of the present invention. Figure 3 As shown, the device includes: a first acquisition module 30, used to acquire a perception data set of the current vehicle, wherein the perception data set includes: first perception data perceived by the first sensor, second perception data perceived by the second sensor, and vehicle speed; a first determination module 32, used to determine a target judgment threshold corresponding to the current vehicle based on the perception data set, wherein the target judgment threshold is used to determine whether the first target perceived by the first sensor is the same as the second target perceived by the second sensor; a determination module 34, used to compare the first perception data and the second perception data with the target judgment threshold to obtain detection results of the first sensor and the second sensor, and the detection results are used to characterize whether the first sensor is successfully associated with the second sensor.
[0099] Optionally, the determination module includes: a first determination unit, used to determine the target detection condition corresponding to the current vehicle based on the perception data set; an acquisition unit, used to acquire the perception data corresponding to the target detection condition from a preset mapping relationship, wherein the preset mapping relationship is used to characterize the data distribution corresponding to different detection conditions; a second determination unit, used to determine the target judgment threshold based on the target data distribution.
[0100] Optionally, the first determination unit includes: a first determination subunit, used to determine the target category of the first target and the target longitudinal distance between the current vehicle and the first target based on the first perception data; and a second determination subunit, used to determine the target detection condition based on the target category, vehicle speed and target longitudinal distance.
[0101] Optionally, the device also includes: a second acquisition module, used to obtain third perception data and fourth perception data obtained by the first sensor and the second sensor respectively sensing the same third target under different detection conditions; a processing module, used to obtain the difference between the third perception data and the fourth perception data to obtain a first difference; a second determination module, used to determine the data distribution of the first difference under different detection conditions; and a generation module, used to generate a preset mapping relationship based on different detection conditions and the data distribution corresponding to the different detection conditions.
[0102] Optionally, the second determination unit includes: a first determination subunit, used to determine the target judgment threshold as the sum of the mean and the mean square error of the target data distribution in response to the first detection of the first sensor and the second sensor in the current cycle; a second determination subunit, used to determine the target judgment threshold as the sum of the mean and a first product in response to the first sensor and the second sensor being detected not for the first time in the current cycle, and the first target and the second target being the same in a cycle before the current cycle, where the first product is the product of the mean square error and a first preset value; a third determination subunit, used to determine the target judgment threshold as the sum of the mean and the second product in response to the first sensor and the second sensor being detected not for the first time in the current cycle, and the first target and the second target being the same in the two cycles before the current cycle, where the second product is the product of the mean square error and the second preset value.
[0103] Optionally, the determination module includes: an acquisition unit, used to acquire the difference between the first perception data and the second perception data to obtain a second difference; and a detection unit, used to compare the second difference with a target judgment threshold to obtain a detection result.
[0104] Optionally, the detection unit includes: a first determination subunit, used to determine that the detection result is that the first target and the second target are the same in response to the second difference being less than or equal to the target judgment threshold; and a second determination subunit, used to determine that the detection result is that the first target and the second target are different in response to the second difference being greater than the target judgment threshold.
[0105] Optionally, the second determination unit is used to determine a target judgment threshold based on the target data distribution, and the target judgment threshold includes at least one of the following: a longitudinal distance threshold, a lateral distance threshold, a longitudinal speed threshold and a lateral speed threshold.
[0106] Example 3
[0107] According to another aspect of an embodiment of the present invention, a target vehicle is also provided, comprising: one or more processors; a storage device for storing one or more programs; when the one or more programs are executed by one or more processors, the one or more processors execute the target detection method described in the above embodiment.
[0108] Example 4
[0109] According to another aspect of an embodiment of the present invention, a non-volatile storage medium is further provided. The non-volatile storage medium includes a stored program, wherein when the program is running, the processor of the device where the program is located is controlled to execute the target detection method described in the above embodiment.
[0110] The serial numbers of the above embodiments of the present invention are only for description and do not represent the advantages or disadvantages of the embodiments.
[0111] In the above embodiments of the present invention, the description of each embodiment has its own emphasis. For parts that are not described in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.
[0112] In the several embodiments provided in this application, it should be understood that the disclosed technical content can be implemented in other ways. Among them, the device embodiments described above are only schematic. For example, the division of the units can be a logical function division. There may be other division methods in actual implementation. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of units or modules, which can be electrical or other forms.
[0113] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple units. Some or all of the units may be selected according to actual needs to achieve the purpose of the present embodiment.
[0114] In addition, each functional unit in each embodiment of the present invention may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.
[0115] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or 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, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), mobile hard disk, magnetic disk or optical disk and other media that can store program codes.
[0116] The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principle of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.
Claims
1. A target detection method, characterized in that: include: Acquire a perception data set of the current vehicle, wherein the perception data set includes: first perception data perceived by a first sensor, second perception data perceived by a second sensor, and vehicle speed; Based on the perception data set, determining a target judgment threshold corresponding to the current vehicle, wherein the target judgment threshold is used to determine whether a first target perceived by the first sensor is the same as a second target perceived by the second sensor; Comparing the first sensing data and the second sensing data with the target judgment threshold to obtain detection results of the first sensor and the second sensor; Wherein, determining the target judgment threshold corresponding to the current vehicle based on the perception data set includes: determining the target detection condition corresponding to the current vehicle based on the perception data set; acquiring the perception data corresponding to the target detection condition from a preset mapping relationship, wherein the preset mapping relationship is used to characterize the data distribution corresponding to different detection conditions; determining the target judgment threshold based on the data distribution; The method further includes: obtaining third sensing data and fourth sensing data obtained by respectively sensing the same third target by the first sensor and the second sensor under different detection conditions; obtaining a difference between the third sensing data and the fourth sensing data to obtain a first difference; determining a data distribution of the first difference under the different detection conditions; and generating the preset mapping relationship based on the different detection conditions and the data distribution corresponding to the different detection conditions; Based on the data distribution, determining the target judgment threshold includes: in response to the first detection of the first sensor and the second sensor in the current cycle, determining the target judgment threshold as the sum of the mean and the mean square error of the data distribution; in response to the first sensor and the second sensor being detected not for the first time in the current cycle, and the first target and the second target being the same in a cycle before the current cycle, determining the target judgment threshold as the sum of the mean and a first product, the first product being the product of the mean square error and a first preset value; in response to the first sensor and the second sensor being detected not for the first time in the current cycle, and the first target and the second target being the same in two cycles before the current cycle, determining the target judgment threshold as the sum of the mean and the second product, the second product being the product of the mean square error and a second preset value.
2. The method according to claim 1, characterized in that Determining a target detection condition corresponding to the current vehicle based on the perception data set includes: Determining, based on the first perception data, a target category of the first target and a target longitudinal distance between the current vehicle and the first target; The object detection condition is determined based on the object category, the vehicle speed and the object longitudinal distance.
3. The method according to claim 1, characterized in that Comparing the first sensing data and the second sensing data with the target judgment threshold to obtain detection results of the first sensor and the second sensor includes: Obtaining a difference between the first sensed data and the second sensed data to obtain a second difference; The second difference is compared with the target judgment threshold to obtain the detection result.
4. The method according to claim 3, characterized in that Comparing the second difference with the target judgment threshold to obtain the detection result includes: In response to the second difference being less than or equal to the target determination threshold, determining that the detection result is that the first target is the same as the second target; In response to the second difference being greater than the target determination threshold, it is determined that the detection result is that the first target is different from the second target.
5. The method according to any one of claims 1 to 4, characterized in that: The target judgment threshold includes at least one of the following: a longitudinal distance threshold, a lateral distance threshold, a longitudinal speed threshold and a lateral speed threshold.
6. A target detection device, characterized in that: include: A first acquisition module is used to acquire a set of perception data of the current vehicle, wherein the set of perception data includes: first perception data perceived by the first sensor, second perception data perceived by the second sensor, and vehicle speed; A first determination module, configured to determine a target judgment threshold corresponding to the current vehicle based on the perception data set, wherein the target judgment threshold is used to determine whether a first target perceived by the first sensor is the same as a second target perceived by the second sensor; a determination module, configured to compare the first sensing data and the second sensing data with the target determination threshold to obtain detection results of the first sensor and the second sensor, wherein the detection results are used to indicate whether the first sensor is successfully associated with the second sensor; The first determination module includes: a first determination unit, used to determine the target detection condition corresponding to the current vehicle based on the perception data set; an acquisition unit, used to acquire the perception data corresponding to the target detection condition from a preset mapping relationship, wherein the preset mapping relationship is used to characterize the data distribution corresponding to different detection conditions; a second determination unit, used to determine the target judgment threshold based on the data distribution; The device further includes: a second acquisition module, configured to acquire third sensing data and fourth sensing data obtained by respectively sensing the same third target by the first sensor and the second sensor under different detection conditions; acquiring a difference between the third sensing data and the fourth sensing data to obtain a first difference; a second determination module, configured to determine a data distribution of the first difference under the different detection conditions; and a generation module, configured to generate the preset mapping relationship based on the different detection conditions and the data distribution corresponding to the different detection conditions; The second determination unit includes: a first determination subunit, which is used to determine the target judgment threshold as the sum of the mean and the mean square error of the data distribution in response to the first detection of the first sensor and the second sensor in the current cycle; a second determination subunit, which is used to determine the target judgment threshold as the sum of the mean and a first product in response to the first sensor and the second sensor being detected not for the first time in the current cycle, and the first target and the second target being the same in a cycle before the current cycle, wherein the first product is the product of the mean square error and a first preset value; and a third determination subunit, which is used to determine the target judgment threshold as the sum of the mean and the second product in response to the first sensor and the second sensor being detected not for the first time in the current cycle, and the first target and the second target being the same in the two cycles before the current cycle, wherein the second product is the product of the mean square error and a second preset value.
7. A vehicle, characterized in that: include: one or more processors; A storage device for storing one or more programs; When the one or more programs are executed by the one or more processors, the one or more processors execute the target detection method described in any one of claims 1-5.
