A target detection method, device and equipment

Abstract

The application discloses a target detection method, device and equipment, comprising: acquiring detection information of a millimeter wave radar, wherein the detection information comprises a plurality of point cloud clusters; calculating the confidence of each point cloud cluster, and screening out a to-be-evaluated point cloud cluster with a confidence less than a threshold; acquiring image information of a region corresponding to the to-be-evaluated point cloud cluster, and identifying the image information to obtain an image identification result; and adjusting radar parameters of the to-be-evaluated point cloud cluster according to the image identification result to obtain a radar detection target. Based on the millimeter wave radar detection result, image information is introduced as an auxiliary prompt means for a weak target region to realize accurate identification of a weak reflection target, thereby enhancing the overall perception ability of the millimeter wave radar, avoiding dependence on high computing power and high bandwidth hardware, and effectively improving the detection rate of the system for the weak reflection target.

Description

Technical Field

[0001] This invention relates to the field of autonomous driving technology, and in particular to a vehicle target detection method, apparatus, and device. Background Technology

[0002] In the field of intelligent driving assistance systems for automobiles, millimeter-wave radar is widely used for functions such as forward collision warning and automatic emergency braking due to its all-weather operation and excellent metal object detection characteristics. Compared with traditional vision sensors, millimeter-wave radar is less sensitive to adverse weather conditions such as rain, fog, and snow, and has the advantages of long detection range and high velocity resolution. However, it also has obvious limitations, mainly manifested in short detection range and low recognition confidence for non-metallic materials or low-reflection targets, such as pedestrians, animals, wooden obstacles, and plastic cones, which can easily lead to missed detections or delayed response.

[0003] To overcome the shortcomings of millimeter-wave radar in detecting weakly reflective targets, existing technologies generally adopt two approaches: one is to improve hardware performance, such as using higher-resolution 4D imaging millimeter-wave radar to improve the distribution of angle and range point clouds, but this approach significantly increases the cost of sensors and computing modules; the other is deep fusion of multiple sensors, such as fusing radar data with camera and lidar data at the feature level or decision level, which improves detection accuracy, but usually relies on high-performance chips and high-bandwidth data transmission, resulting in increased power consumption and significantly increased hardware costs and system complexity. Summary of the Invention

[0004] This invention provides a target detection method to achieve accurate target detection.

[0005] According to a first aspect of the present invention, a target detection method is provided, comprising: acquiring detection information from a millimeter-wave radar, wherein the detection information includes a plurality of point cloud clusters;

[0006] Calculate the confidence level of each point cloud cluster and filter out the point cloud clusters to be evaluated whose confidence level is less than the threshold.

[0007] Obtain image information of the region corresponding to the point cloud cluster to be evaluated, and identify the image information to obtain image recognition results;

[0008] The radar parameters of the point cloud cluster to be evaluated are adjusted based on the image recognition results to obtain the radar detection target.

[0009] According to another aspect of the present invention, a target detection device is provided, comprising: a millimeter-wave radar detection module for acquiring detection information from the millimeter-wave radar, wherein the detection information includes multiple point cloud clusters;

[0010] The confidence level determination module is used to calculate the confidence level of each point cloud cluster and filter out the point cloud clusters to be evaluated whose confidence level is less than the threshold.

[0011] The image recognition module is used to acquire image information of the region corresponding to the point cloud cluster to be evaluated, and to recognize the image information to obtain image recognition results;

[0012] The target detection module is used to adjust the radar parameters of the point cloud cluster to be evaluated based on the image recognition results to obtain the radar detection target.

[0013] According to another aspect of the present invention, an electronic device is provided, the electronic device comprising:

[0014] At least one processor; and

[0015] A memory communicatively connected to the at least one processor; wherein,

[0016] The memory stores a computer program that can be executed by the at least one processor, the computer program being executed by the at least one processor to enable the at least one processor to perform the method described in any embodiment of the present invention.

[0017] According to another aspect of the present invention, a computer-readable storage medium is provided, the computer-readable storage medium storing computer instructions for causing a processor to execute and implement the method described in any embodiment of the present invention.

[0018] The technical solution of this invention introduces image information as an auxiliary prompting means for weak target areas based on the detection results of millimeter-wave radar, so as to achieve accurate identification of weak reflection targets, thereby enhancing the overall perception capability of millimeter-wave radar. While avoiding reliance on high computing power and high bandwidth hardware, it effectively improves the detection rate of weak reflection targets by the system.

[0019] It should be understood that the description in this section is not intended to identify key or essential features of the embodiments of the present invention, nor is it intended to limit the scope of the invention. Other features of the invention will become readily apparent from the following description. Attached Figure Description

[0020] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0021] Figure 1This is a flowchart of a target detection method provided in Embodiment 1 of the present invention;

[0022] Figure 2 This is a flowchart of a target detection method provided according to Embodiment 2 of the present invention;

[0023] Figure 3 This is a schematic diagram of the structure of a target detection device according to Embodiment 3 of the present invention;

[0024] Figure 4 This is a schematic diagram of the structure of the electronic device provided in Embodiment 4 of the present invention. Detailed Implementation

[0025] To enable those skilled in the art to better understand the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of the present invention.

[0026] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this invention are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of the invention described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover a non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0027] Example 1

[0028] Figure 1 This is a flowchart illustrating a method for recording vehicle test problem data according to Embodiment 1 of the present invention. This embodiment is applicable to situations where vehicle test problem data is recorded. This method can be executed by a vehicle test problem data recording device, which can be implemented in hardware and / or software. Figure 1 As shown, the method includes:

[0029] Step S101: Obtain the detection information from the millimeter-wave radar.

[0030] Optionally, the detection information of the millimeter-wave radar can be obtained, including: obtaining the point cloud detected by the millimeter-wave radar, and clustering the point cloud to obtain multiple point cloud clusters.

[0031] Specifically, this embodiment uses millimeter-wave radar as the main detection unit. It employs a scheme where the radar detection results actively trigger the camera to analyze key areas of the image. When the millimeter-wave radar detects clusters of points with insufficient confidence or abnormal features, it drives the camera to only acquire and process image information from the relevant areas. A lightweight, fast judgment algorithm generates area prompt information, which is then used to adjust subsequent radar point aggregation or judgment parameters to enhance the detection capability for non-metallic targets. Simultaneously, this embodiment also incorporates an adaptive adjustment strategy for the camera prompt ratio. By combining external lighting, weather conditions, and prompt effectiveness records, the frequency and weight of image intervention are intelligently and dynamically adjusted to ensure that detection performance and power consumption requirements are balanced under different environments.

[0032] Specifically, in this embodiment, millimeter-wave radar is used as the primary sensor to continuously output the detected point cloud. The front end clusters the detected point cloud to obtain multiple point cloud clusters, and each point cloud cluster represents a possible target. When performing clustering, clustering can be performed according to distance or according to feature similarity. This embodiment does not limit the specific clustering method used. As long as a possible target can be detected, it is within the protection scope of this application.

[0033] Step S102: Calculate the confidence level of each point cloud cluster and filter out the point cloud clusters to be evaluated whose confidence level is less than the threshold.

[0034] Optionally, the attribute scores of the point cloud cluster and the weights corresponding to each attribute score are obtained, wherein the attribute scores include intensity signal score, stability score and motion score; the product of the attribute scores and weights is calculated and the confidence score of the point cloud cluster is obtained.

[0035] Optionally, the attribute scores of the point cloud cluster are obtained, including: obtaining the sub-attribute scores of each point cloud in the point cloud coarse; calculating the mean of each sub-attribute score, and using the mean as the attribute score of the point cloud cluster.

[0036] Specifically, in this embodiment, after obtaining point cloud clusters through clustering, the confidence score corresponding to each point cloud cluster is calculated. The confidence score is a rating of the radar's certainty regarding the actual existence of the detected target. The radar calculates the confidence score based on three attributes of the point cloud cluster: intensity, stability, and motion. The signal strength score represents the target's reflectivity; metallic targets, such as cars, have strong reflectivity and therefore a high signal strength score, while non-metallic targets, such as pedestrians, have weak reflectivity and therefore a low signal strength score. The stability score represents the stability of target detection; for example, if the target appears in the same position several times in a row, the stability score is high, while if the target detection is intermittent, the stability score is low. The motion score represents the target's motion; for example, if the speed conforms to normal vehicle driving patterns, the motion score is high, while if there are significant speed changes, the motion score is low. Of course, this embodiment only uses the above three attribute scores as examples and does not limit the specific types of evaluation attributes involved in the confidence score.

[0037] It should be noted that in this embodiment, corresponding weights are pre-configured for each attribute score. For example, the weight corresponding to the intensity signal score is A1, the weight corresponding to the stability score is A2, and the weight corresponding to the motion score is A3. In this embodiment, the specific values ​​of each weight are not limited. As long as the sum is 1, it is within the protection scope of this application. The confidence level of each point cloud cluster can be calculated using the following formula (1):

[0038] Confidence level = (Intensity signal score × A1 + Stability score × A2 + Motion score × A3)

[0039] Specifically, when calculating the attribute score of each point cloud cluster, the sub-attribute scores of each point cloud in each cluster are obtained, and the average of the sub-attribute scores of all point cloud clusters is taken as the attribute score of the point cloud cluster. For example, when it is determined that a point cloud cluster contains 30 point clouds, when calculating the intensity signal score of the point cloud cluster, the sub-intensity signal scores of each point cloud are calculated, and the average of the sub-intensity signal scores of the 30 point clouds is taken as the intensity signal score of the point cloud cluster. Of course, this embodiment only uses the calculation of intensity signal score as an example for explanation. The calculation methods for stability score and motion score are roughly the same, and will not be elaborated on in this embodiment.

[0040] It should be noted that in this embodiment, a threshold is predetermined and point cloud clusters with confidence levels lower than the threshold are selected for evaluation. When the confidence level is relatively low, it means that the millimeter-wave radar cannot accurately detect whether there are obstacles in front of the vehicle. In this case, the image information captured by the image acquisition component is needed for assistance.

[0041] Step S103: Obtain image information of the region corresponding to the point cloud cluster to be evaluated, and identify the image information to obtain image recognition results.

[0042] Optionally, image information of the region corresponding to the point cloud cluster to be evaluated is acquired, and image recognition results are obtained by recognizing the image information, including: acquiring current environmental information through the environmental adjustment component; activating the image acquisition component when the environmental information meets the image assistance conditions; acquiring image information of the region corresponding to the point cloud cluster to be evaluated through the image acquisition component; and marking the identified target when the image information contains a target, and using the marked target as the image recognition result.

[0043] Specifically, in this embodiment, the image acquisition component will collect image information of the area corresponding to the point cloud cluster to be evaluated. Since the collected image information only corresponds to the local area of ​​the point cloud cluster to be evaluated, the processing pressure of the image information will be greatly reduced. When it is determined that the image information contains a target, the identified target will be marked, for example, by using a box to circle the existing target, and the marked target will be used as the image recognition result. If no target is identified, there is no need to mark it.

[0044] It should be noted that, to further reduce system power consumption and adapt to the fluctuations in image effectiveness under complex environments, this embodiment specifically controls the activation of the image acquisition component through an environmental adjustment component. This component acquires current environmental information, such as light intensity, weather conditions, and radar point stability indicators, to determine whether the current image acquisition has intervention value. In favorable environments with high image recognition value, the system allows the image intervention mechanism to participate in prompts at a higher frequency; however, in environments with significant image degradation, such as at night or in rain / fog, the system reduces the image acquisition frequency or even temporarily shuts down the image acquisition component to ensure power consumption is controlled. Simultaneously, during image intervention, the system also constructs a prompt effectiveness feedback evaluation mechanism, recording the changes in clustering accuracy resulting from each image prompt and periodically evaluating the cost-effectiveness of image participation to further optimize the working timing and frequency of the image acquisition component. This achieves a target recognition enhancement mechanism based on radar dominance, image prompts, and dynamic adjustment, which improves the detection capability of weak targets while controlling system cost and computing resources, achieving good recognition results without introducing expensive 4D radar or multimodal strong fusion structures.

[0045] Step S104: Adjust the radar parameters of the point cloud cluster to be evaluated based on the image recognition results to obtain radar detection targets.

[0046] Optionally, the radar parameters of the point cloud cluster to be evaluated are adjusted according to the image recognition results to obtain the radar detection target, including: converting the image recognition results to the radar coordinate system to obtain the coordinate-transformed image recognition results; obtaining a specified point cloud in the point cloud cluster to be evaluated that has the same position as the marked target and marking it to obtain the marked point cloud cluster; and adjusting the radar parameters of the marked point cloud cluster to obtain the radar detection target.

[0047] Optionally, the radar parameters of the marked point cloud cluster are adjusted to obtain the radar-detected target, including: adjusting the sub-attribute scores of the marked point cloud by a specified amount to obtain the adjusted sub-attribute scores; calculating the updated mean of the sub-attribute scores and the adjusted sub-attribute scores of each point cloud in the marked point cloud cluster; when the updated mean is greater than the confidence level, it is determined that a target has been detected in the marked point cloud cluster, and the detected target is output.

[0048] Specifically, in this embodiment, the image recognition result is converted to the radar coordinate system to obtain the image recognition result after coordinate transformation. This allows the marked target to be imaged onto the corresponding point cloud in the radar coordinate system, and the mapped point cloud is marked to obtain the marked point cloud cluster. Based on the image recognition result, the radar information is prompted. After obtaining the marked prompt information, the system does not directly make a fusion judgment. Instead, it intervenes in the radar's target association mechanism by adjusting the clustering module parameters, such as the point distance threshold and weight factor, so that the point cloud area marked by the image prompt gets a higher aggregation priority. This prompt-based intervention method effectively preserves the radar's dominance in target determination and guides it to more accurately lock the target area, thereby improving the detection success rate of non-metallic objects or weakly reflective targets, such as pedestrians, plastic road barriers, and thin wooden boards.

[0049] In one specific implementation, this method adjusts the radar parameters of the marked point cloud cluster to obtain the radar-detected target. The adjustment primarily involves increasing the sub-attribute scores of the marked point clouds by a specified margin. If a point cloud cluster contains 30 point clouds, and 25 of them are marked, it indicates that these 25 points represent a target based on the image information. Therefore, to improve the detection accuracy of the millimeter-wave radar, the sub-attribute scores of these 25 point clouds are increased. For example, if the attribute is an intensity signal, the sub-intensity signal of these 25 point clouds is increased. The intensity signal score is increased by 5 points according to a preset range. Specifically, when calculating the intensity signal score of the entire point cloud cluster, the average of the increased intensity signal scores of the 25 point clouds and the updated average of the individual intensity signal scores of the 5 point clouds is calculated. This updated average is used as the intensity signal score obtained after image information assistance. This embodiment only uses the adjustment of the intensity signal score as an example; the adjustment method for the stability score is roughly the same and will not be elaborated upon here. Since the image recognition result does not significantly affect the motion score, the motion score of the marked point cloud cluster remains unchanged in this embodiment. For example, before the image prompt, the intensity signal score of the blurred human figure is 30, the stability score is 20, and the motion score is 50, with a confidence level of 32%. Since this is less than the threshold of 40%, image assistance is triggered. After the image prompt, the parameters are adjusted and recalculated, resulting in an intensity signal score of 35, a stability score of 60, a motion score of 50, and a new confidence level of 49%, which meets the output criteria.

[0050] It is worth mentioning that in this embodiment, when the millimeter-wave radar detects low-confidence or weak-point areas, it only triggers the camera to acquire lightweight images in these areas and provides clues through a fast local image extraction algorithm, rather than performing traditional object detection or deep fusion on the entire scene image. This mechanism effectively reduces the image computation load and avoids the need for high computing power and high-cost hardware. Instead of directly fusing or discriminating the image recognition results with the radar output at the target level, it adjusts the radar clustering parameters, such as aggregation weights and thresholds, to increase the aggregation priority of corresponding points in a suggestive manner during the clustering stage. This indirect fusion mode maintains the radar's dominant position in the final target determination.

[0051] In this embodiment, based on the detection results of millimeter-wave radar, image information is introduced as an auxiliary prompting method for weak target areas to achieve accurate identification of weak reflective targets, thereby enhancing the overall perception capability of millimeter-wave radar. While avoiding reliance on high computing power and high bandwidth hardware, it effectively improves the system's detection rate of weak reflective targets.

[0052] Example 2

[0053] Figure 2The flowchart of a target detection method provided in Embodiment 2 of the present invention is as follows. This embodiment is based on the above embodiment. After adjusting the radar parameters of the point cloud cluster to be evaluated according to the image recognition results to obtain the radar detection target, it also includes verifying the radar detection target. Figure 2 As shown, the method includes:

[0054] Step S201: Obtain the detection information from the millimeter-wave radar.

[0055] Optionally, the detection information of the millimeter-wave radar can be obtained, including: obtaining the point cloud detected by the millimeter-wave radar, and clustering the point cloud to obtain multiple point cloud clusters.

[0056] Step S202: Calculate the confidence level of each point cloud cluster and filter out the point cloud clusters to be evaluated whose confidence level is less than the threshold.

[0057] Optionally, the attribute scores of the point cloud cluster and the weights corresponding to each attribute score are obtained, wherein the attribute scores include intensity signal score, stability score and motion score; the product of the attribute scores and weights is calculated and the confidence score of the point cloud cluster is obtained.

[0058] Optionally, the attribute scores of the point cloud cluster are obtained, including: obtaining the sub-attribute scores of each point cloud in the point cloud coarse; calculating the mean of each sub-attribute score, and using the mean as the attribute score of the point cloud cluster.

[0059] Step S203: Obtain image information of the region corresponding to the point cloud cluster to be evaluated, and identify the image information to obtain the image recognition result.

[0060] Optionally, image information of the region corresponding to the point cloud cluster to be evaluated is acquired, and image recognition results are obtained by recognizing the image information, including: acquiring current environmental information through the environmental adjustment component; activating the image acquisition component when the environmental information meets the image assistance conditions; acquiring image information of the region corresponding to the point cloud cluster to be evaluated through the image acquisition component; and marking the identified target when the image information contains a target, and using the marked target as the image recognition result.

[0061] Step S204: Adjust the radar parameters of the point cloud cluster to be evaluated based on the image recognition results to obtain radar detection targets.

[0062] Optionally, the radar parameters of the point cloud cluster to be evaluated are adjusted according to the image recognition results to obtain the radar detection target, including: converting the image recognition results to the radar coordinate system to obtain the coordinate-transformed image recognition results; obtaining a specified point cloud in the point cloud cluster to be evaluated that has the same position as the marked target and marking it to obtain the marked point cloud cluster; and adjusting the radar parameters of the marked point cloud cluster to obtain the radar detection target.

[0063] Optionally, the radar parameters of the marked point cloud cluster are adjusted to obtain the radar-detected target, including: adjusting the sub-attribute scores of the marked point cloud by a specified amount to obtain the adjusted sub-attribute scores; calculating the updated mean of the sub-attribute scores and the adjusted sub-attribute scores of each point cloud in the marked point cloud cluster; when the updated mean is greater than the confidence level, it is determined that a target has been detected in the marked point cloud cluster, and the detected target is output.

[0064] Step S205: Verify the radar-detected target.

[0065] Specifically, in this embodiment, after acquiring the radar-detected target, the radar-detected target is verified to ensure its accuracy. For example, if the radar detects a pedestrian in front of the vehicle, but image recognition or other sensing technologies determine that there is no obstacle ahead, then the radar-detected target is determined to be incorrect. Furthermore, this embodiment generates a detection report based on the verification results, which includes the verification results and the specific time of the verification.

[0066] It should be noted that in this embodiment, the generated verification report will be visually displayed on the front-end interface, allowing developers to intuitively and efficiently obtain the accuracy of the radar detection results. Furthermore, an alarm will be generated and played via voice or image when the verification fails, prompting developers to pay close attention and further improving the safety of vehicle development. Of course, this embodiment is merely illustrative and does not limit the specific display method of the verification report and alarm; anything that facilitates developer viewing falls within the scope of protection of this application.

[0067] In this embodiment, based on the detection results of millimeter-wave radar, image information is introduced as an auxiliary prompting method for weak target areas to achieve accurate identification of weak reflective targets, thereby enhancing the overall perception capability of millimeter-wave radar. While avoiding reliance on high computing power and high bandwidth hardware, it effectively improves the system's detection rate of weak reflective targets.

[0068] Example 3

[0069] Figure 3 A schematic diagram of the structure of a target detection device provided in Embodiment 3 of the present invention is shown. Figure 3 The device includes: a millimeter-wave radar detection module 310, a confidence judgment module 320, an image recognition module 330, and a target detection module 340.

[0070] Among them, the millimeter-wave radar detection module 310 is used to acquire the detection information of the millimeter-wave radar, which includes multiple point cloud clusters;

[0071] The confidence level judgment module 320 is used to calculate the confidence level of each point cloud cluster and filter out the point cloud clusters to be evaluated whose confidence level is less than the threshold.

[0072] The image recognition module 330 is used to acquire image information of the region corresponding to the point cloud cluster to be evaluated, and to recognize the image information to obtain the image recognition result;

[0073] The target detection module 340 is used to adjust the radar parameters of the point cloud cluster to be evaluated based on the image recognition results to obtain radar detection targets.

[0074] Optional, a millimeter-wave radar detection module is used to acquire the point cloud detected by the millimeter-wave radar.

[0075] Cluster the point cloud to obtain multiple point cloud clusters.

[0076] Optionally, a confidence judgment module is used to obtain the attribute scores of the point cloud cluster and the weights corresponding to each attribute score, wherein the attribute scores include intensity signal score, stability score and motion score;

[0077] Calculate the product of attribute scores and weights and obtain the confidence level of the point cloud cluster.

[0078] Optionally, the confidence score module is also used to obtain the sub-attribute scores of each point cloud in the point cloud cluster;

[0079] Calculate the mean of each sub-attribute score and use the mean as the attribute score of the point cloud cluster.

[0080] Optionally, an image recognition module is used to acquire current environmental information through an environmental adjustment component, and to activate an image acquisition component when it is determined that the environmental information meets the image assistance conditions.

[0081] Image information of the area corresponding to the point cloud cluster to be evaluated is acquired using the image acquisition component;

[0082] When it is determined that the image information contains a target, the identified target is marked, and the marked target is used as the image recognition result.

[0083] Optionally, a target detection module is used to convert the image recognition results to the radar coordinate system to obtain the image recognition results after coordinate transformation;

[0084] Obtain the specified point cloud within the point cloud cluster to be evaluated that has the same location as the marked target, and mark it to obtain the marked point cloud cluster;

[0085] Adjustments are made based on the radar parameters of the marked point cloud clusters to obtain radar-detected targets.

[0086] Optionally, the target detection module is also used to upscale the sub-attribute scores of the marked point cloud by a specified margin to obtain the upscaled sub-attribute scores.

[0087] Calculate the updated mean of the sub-attribute scores and up-adjusted sub-attribute scores for each point cloud in the labeled point cloud cluster;

[0088] When the updated mean is greater than the confidence level, it is determined that a target has been detected in the labeled point cloud cluster, and the detected target is output.

[0089] The target detection device provided in the embodiments of the present invention can execute the target detection method provided in any embodiment of the present invention, and has the corresponding functional modules and beneficial effects of the method execution.

[0090] Example 4

[0091] Figure 4 A schematic diagram of the structure of an electronic device 10 that can be used to implement embodiments of the present invention is provided. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device may also represent various forms of mobile devices, such as personal digital processors, cellular phones, smartphones, wearable devices (e.g., helmets, glasses, watches, etc.), and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely illustrative and are not intended to limit the implementation of the invention described and / or claimed herein.

[0092] like Figure 4 As shown, the electronic device 10 includes at least one processor 11 and a memory, such as a read-only memory (ROM) 12 or a random access memory (RAM) 13, communicatively connected to the at least one processor 11. The memory stores computer programs executable by the at least one processor. The processor 11 can perform various appropriate actions and processes based on the computer programs stored in the ROM 12 or loaded into the RAM 13 from storage unit 18. The RAM 13 may also store various programs and data required for the operation of the electronic device 10. The processor 11, ROM 12, and RAM 13 are interconnected via a bus 14. An input / output (I / O) interface 15 is also connected to the bus 14.

[0093] Multiple components in electronic device 10 are connected to I / O interface 15, including: input unit 16, such as keyboard, mouse, etc.; output unit 17, such as various types of displays, speakers, etc.; storage unit 18, such as disk, optical disk, etc.; and communication unit 19, such as network card, modem, wireless transceiver, etc. Communication unit 19 allows electronic device 10 to exchange information / data with other devices through computer networks such as the Internet and / or various telecommunications networks.

[0094] Processor 11 can be a variety of general-purpose and / or special-purpose processing components with processing and computing capabilities. Some examples of processor 11 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various special-purpose artificial intelligence (AI) computing chips, various processors running machine learning model algorithms, a digital signal processor (DSP), and any suitable processor, controller, microcontroller, etc. Processor 11 performs the various methods and processes described above, such as object detection methods.

[0095] In some embodiments, the target detection method may be implemented as a computer program tangibly contained in a computer-readable storage medium, such as storage unit 18. In some embodiments, part or all of the computer program may be loaded and / or mounted on electronic device 10 via ROM 12 and / or communication unit 19. When the computer program is loaded into RAM 13 and executed by processor 11, one or more steps of the target detection method described above may be performed. Alternatively, in other embodiments, processor 11 may be configured to perform the target detection method by any other suitable means (e.g., by means of firmware).

[0096] Various embodiments of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field-programmable gate arrays (FPGAs), application-specific integrated circuits (ASICs), application-specific standard products (ASSPs), systems-on-a-chip (SoCs), payload-programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various embodiments may include implementations in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which may be a dedicated or general-purpose programmable processor, capable of receiving data and instructions from a storage system, at least one input device, and at least one output device, and transmitting data and instructions to the storage system, the at least one input device, and the at least one output device.

[0097] Computer programs used to implement the methods of the present invention may be written in any combination of one or more programming languages. These computer programs may be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing device, such that when executed by the processor, the computer programs cause the functions / operations specified in the flowcharts and / or block diagrams to be performed. The computer programs may be executed entirely on a machine, partially on a machine, or as a standalone software package, partially on a machine and partially on a remote machine, or entirely on a remote machine or server.

[0098] In the context of this invention, a computer-readable storage medium can be a tangible medium that may contain or store a computer program for use by or in conjunction with an instruction execution system, apparatus, or device. A computer-readable storage medium may include, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination thereof. Alternatively, a computer-readable storage medium may be a machine-readable signal medium. More specific examples of machine-readable storage media include electrical connections based on one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fibers, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof.

[0099] To provide interaction with a user, the systems and techniques described herein can be implemented on an electronic device having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and pointing device (e.g., a mouse or trackball) through which the user provides input to the electronic device. Other types of devices can also be used to provide interaction with the user; for example, feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including sound input, voice input, or tactile input).

[0100] The systems and technologies described herein can be implemented in computing systems that include backend components (e.g., as data servers), or computing systems that include middleware components (e.g., application servers), or computing systems that include frontend components (e.g., user computers with graphical user interfaces or web browsers through which users can interact with implementations of the systems and technologies described herein), or any combination of such backend, middleware, or frontend components. The components of the system can be interconnected via digital data communication of any form or medium (e.g., communication networks). Examples of communication networks include local area networks (LANs), wide area networks (WANs), blockchain networks, and the Internet.

[0101] A computing system can include clients and servers. Clients and servers are generally located far apart and typically interact through communication networks. The client-server relationship is created by computer programs running on the respective computers and having a client-server relationship with each other. The server can be a cloud server, also known as a cloud computing server or cloud host, which is a hosting product within the cloud computing service system to address the shortcomings of traditional physical hosts and VPS services, such as high management difficulty and weak business scalability.

[0102] It should be understood that the various forms of processes shown above can be used, with steps reordered, added, or deleted. For example, the steps described in this invention can be executed in parallel, sequentially, or in different orders, as long as the desired result of the technical solution of this invention can be achieved, and this is not limited herein.

[0103] The specific embodiments described above do not constitute a limitation on the scope of protection of this invention. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this invention should be included within the scope of protection of this invention.

Claims

1. A target detection method, characterized in that the method include: Acquire detection information from millimeter-wave radar, wherein the detection information includes multiple point cloud clusters; Calculate the confidence level of each point cloud cluster and filter out the point cloud clusters to be evaluated whose confidence level is less than the threshold. Obtain image information of the region corresponding to the point cloud cluster to be evaluated, and identify the image information to obtain image recognition results; The radar parameters of the point cloud cluster to be evaluated are adjusted based on the image recognition results to obtain the radar detection target; Adjusting the radar parameters of the point cloud cluster to be evaluated based on the image recognition results to obtain radar detection targets includes: converting the image recognition results to a radar coordinate system to obtain coordinate-transformed image recognition results; obtaining and marking designated point clouds in the point cloud cluster to be evaluated that have the same location as the marked target to obtain marked point cloud clusters; and adjusting the radar parameters of the marked point cloud clusters to obtain radar detection targets. The step of adjusting the radar parameters of the marked point cloud cluster to obtain radar-detected targets includes: increasing the sub-attribute scores of the marked point cloud by a specified amount to obtain increased sub-attribute scores; calculating the updated mean of the sub-attribute scores and increased sub-attribute scores of each point cloud in the marked point cloud cluster; and when the confidence level recalculated based on the updated mean is greater than the threshold, determining that a target has been detected in the marked point cloud cluster and outputting the detected target.

2. The method according to claim 1, characterized in that, The acquisition of millimeter-wave radar detection information includes: Acquire the point cloud detected by the millimeter-wave radar. Cluster the point cloud to obtain multiple point cloud clusters.

3. The method according to claim 1, characterized in that, The calculation of the confidence level of each point cloud cluster includes: Obtain the attribute scores of the point cloud cluster and the weights corresponding to each attribute score, wherein the attribute scores include intensity signal score, stability score and motion score; Calculate the product of the attribute score and the weight, and obtain the confidence level of the point cloud cluster.

4. The method according to claim 3, characterized in that, The acquisition of the attribute scores of the point cloud cluster includes: Obtain the sub-attribute scores of each point cloud in the point cloud cluster; Calculate the mean of each of the sub-attribute scores, and use the mean as the attribute score of the point cloud cluster.

5. The method according to claim 1, characterized in that, The step of acquiring image information of the region corresponding to the point cloud cluster to be evaluated, and recognizing the image information to obtain image recognition results, includes: The current environmental information is obtained through the environmental adjustment component, and the image acquisition component is activated when it is determined that the environmental information meets the image assistance conditions. The image acquisition component acquires the image information of the region corresponding to the point cloud cluster to be evaluated. When it is determined that the image information contains a target, the identified target is marked, and the marked target is used as the image recognition result.

6. A target detection device, characterized in that, include: A millimeter-wave radar detection module is used to acquire detection information from a millimeter-wave radar, wherein the detection information includes multiple point cloud clusters; The confidence level determination module is used to calculate the confidence level of each point cloud cluster and filter out the point cloud clusters to be evaluated whose confidence level is less than the threshold. The image recognition module is used to acquire image information of the region corresponding to the point cloud cluster to be evaluated, and to recognize the image information to obtain image recognition results; The target detection module is used to adjust the radar parameters of the point cloud cluster to be evaluated based on the image recognition results to obtain the radar detection target; The target detection module is used to convert the image recognition result to a radar coordinate system to obtain the image recognition result after coordinate transformation; to obtain and mark the specified point cloud in the point cloud cluster to be evaluated that has the same position as the marked target, so as to obtain the marked point cloud cluster; and to adjust the radar parameters according to the marked point cloud cluster to obtain the radar detection target. The target detection module is further configured to: increase the sub-attribute scores of the marked point cloud by a specified margin to obtain the increased sub-attribute scores; calculate the updated mean of the sub-attribute scores and the increased sub-attribute scores of each point cloud in the marked point cloud cluster; and when the confidence level recalculated based on the updated mean is greater than the threshold, determine that a target has been detected in the marked point cloud cluster and output the detected target.

7. An electronic device, characterized in that, The electronic device includes: At least one processor; and A memory communicatively connected to the at least one processor; wherein, The memory stores a computer program that is executed by the at least one processor to enable the at least one processor to perform the method of any one of claims 1-5.

8. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer instructions that are used to cause a processor to execute the method of any one of claims 1-5.

Images