Parameter tuning method, apparatus, device, medium, and product

Abstract

The application discloses a parameter tuning method and device, equipment, medium and product, and relates to the technical field of data processing. The method comprises the following steps: receiving radar collection parameters, video data and mode conversion ADC data sent by a data collection device; generating a final upgrade firmware according to the radar collection parameters, the video data and the ADC data; and sending the final upgrade firmware to the data collection device, so that the data collection device performs parameter upgrade on a millimeter wave radar of the data collection device according to the final upgrade firmware to realize tuning and obtain a tuning result. Thus, the problem that the accuracy and efficiency of manual debugging in the existing millimeter wave radar signal processing parameter tuning process are not high is solved, and the efficiency of parameter tuning is improved.

Description

Technical Field

[0001] This application relates to the field of data processing technology, and in particular to a parameter optimization method, apparatus, device, medium, and product. Background Technology

[0002] With the rapid development of autonomous driving, intelligent transportation and industrial automation technologies, millimeter-wave radar, as an important sensor capable of target detection and tracking in complex environments, is widely used in many fields such as automotive collision avoidance, lane keeping and drone obstacle avoidance. Millimeter-wave radar has strong penetration capabilities and can work in adverse weather conditions, such as heavy fog, rain and snow. Therefore, its importance in autonomous driving systems is becoming increasingly prominent.

[0003] In the application of millimeter-wave radar, signal processing algorithms have a significant impact on the radar's detection performance. The signal processing parameters of millimeter-wave radar include filtering, detection thresholds, etc. The selection of these parameters directly affects the radar's detection capability, false detection rate, and target positioning accuracy. Traditional parameter setting relies on human experience, and the optimal parameters are determined through a large number of experiments and adjustments. Furthermore, data feedback depends on repeated manual operations. This manual adjustment process is not only time-consuming and labor-intensive, but also makes it difficult to obtain the optimal parameter combination in complex scenarios. As a result, the detection performance of millimeter-wave radar cannot be fully utilized. Moreover, since the experimental results in real scenarios lack effective true values, it is impossible to effectively evaluate the results.

[0004] The above content is only used to help understand the technical solution of this application and does not represent an admission that the above content is prior art. Summary of the Invention

[0005] The main purpose of this application is to provide a parameter tuning method, apparatus, equipment, medium, and product, which aims to solve the technical problems of insufficient accuracy and efficiency of manual tuning in the existing millimeter-wave radar signal processing parameter tuning process.

[0006] To achieve the above objectives, this application proposes a parameter tuning method, which is applied to a host computer and includes:

[0007] Receive radar acquisition parameters, video data, and mode conversion ADC data sent by the data acquisition equipment;

[0008] The final upgrade firmware is generated based on the radar acquisition parameters, video data, and ADC data.

[0009] The final upgrade firmware is sent to the data acquisition device, which then upgrades the parameters of its millimeter-wave radar based on the final upgrade firmware to achieve optimization and obtain the optimization result.

[0010] In one embodiment, the step of generating the final upgrade firmware based on the radar acquisition parameters, video data, and ADC data includes:

[0011] Based on the radar acquisition parameters, several sets of optimization parameter values ​​are generated by optimizing the parameters.

[0012] The video data is frame-by-frame extracted to obtain an image set, and the image set is labeled to obtain the scene target ground truth.

[0013] Write the aforementioned sets of tuning parameter values ​​into a configuration file, and compile based on the configuration file to obtain several upgrade firmwares;

[0014] Send the aforementioned upgraded firmware to the data acquisition device;

[0015] The millimeter-wave radar is upgraded according to the aforementioned several upgrade firmware, and the ADC data is fed back to the millimeter-wave radar after each upgrade. The millimeter-wave radar is used to process the ADC data to obtain several millimeter-wave radar point cloud data groups.

[0016] The system receives several millimeter-wave radar point cloud data sets sent by the data acquisition device, and matches them with the scene target ground truth based on the several millimeter-wave radar point cloud data sets to obtain several matching results.

[0017] The matching results are evaluated to obtain several evaluation results, and the final upgrade firmware is determined based on the several evaluation results.

[0018] In one embodiment, the step of extracting frames from the video data to obtain an image set, and then labeling the image set to obtain the ground truth value of the scene target includes:

[0019] The video data is frame-by-frame extracted based on a preset frame rate to obtain an image set;

[0020] The image set is annotated using an annotation tool to obtain annotation information, which includes the detected target, object bounding box, object category, and spatial location information.

[0021] Generate the scene target truth value based on the annotation information.

[0022] In one embodiment, the step of matching the scene target ground truth values ​​with the plurality of millimeter-wave radar point cloud data groups to obtain a plurality of matching results includes:

[0023] By transforming the coordinate system, the true value of the scene target is projected onto the radar coordinate system to obtain the position information of the true value of the scene target in the radar coordinate system;

[0024] Based on the location information, the several millimeter radar point cloud data groups are matched using maximum weight matching until each millimeter radar point cloud data group is matched, resulting in several matching results.

[0025] In one embodiment, the step of evaluating the plurality of matching results to obtain a plurality of evaluation results, and determining the final upgrade firmware based on the plurality of evaluation results includes:

[0026] The detection rate, false detection rate, and distance cost of each matching result are calculated.

[0027] Based on the detection rate, false detection rate, and distance cost of the several matching results, the several matching results are evaluated using preset evaluation weights to obtain several evaluation results;

[0028] The final upgrade firmware among the several upgrade firmware is determined based on the evaluation results.

[0029] Furthermore, to achieve the above objectives, this application also proposes a parameter optimization method, which is applied to a data acquisition device, including a millimeter-wave radar and a camera device, and the method includes:

[0030] Based on the radar acquisition parameters of the millimeter-wave radar, ADC data is converted through the acquisition mode of the millimeter-wave radar, and video data is acquired through the camera device;

[0031] The radar acquisition parameters, video data, and ADC data are sent to the host computer, which then generates the final upgrade firmware based on the radar acquisition parameters, video data, and ADC data, and sends the final upgrade firmware to the data acquisition device.

[0032] Based on the final upgraded firmware, the parameters of the millimeter-wave radar are upgraded to achieve optimization, and the optimization results are obtained.

[0033] Furthermore, to achieve the above objectives, this application also proposes a parameter tuning device, which is applied to a host computer and includes:

[0034] The receiving module is used to receive radar acquisition parameters, video data, and ADC data sent by the data acquisition device.

[0035] The generation module is used to generate the final upgrade firmware based on the radar acquisition parameters, video data, and ADC data.

[0036] The tuning module is used to send the final upgrade firmware to the data acquisition device, and the data acquisition device upgrades the parameters of the millimeter-wave radar of the data acquisition device according to the upgrade firmware to achieve tuning and obtain the tuning result.

[0037] In addition, to achieve the above objectives, this application also proposes a parameter tuning device, the device comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor, the computer program being configured to implement the steps of the parameter tuning method as described above.

[0038] In addition, to achieve the above objectives, this application also proposes a storage medium, which is a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, it implements the steps of the parameter tuning method described above.

[0039] In addition, to achieve the above objectives, this application also provides a computer program product, which includes a computer program that, when executed by a processor, implements the steps of the parameter tuning method described above.

[0040] One or more technical solutions proposed in this application have at least the following technical effects:

[0041] This application proposes a parameter tuning method, apparatus, device, storage medium, and computer program product. It receives radar acquisition parameters, video data, and mode conversion ADC data sent by a data acquisition device; generates a final upgrade firmware based on the radar acquisition parameters, video data, and ADC data; sends the final upgrade firmware to the data acquisition device, and the data acquisition device upgrades the parameters of its millimeter-wave radar according to the final upgrade firmware to achieve tuning, obtaining the tuning result. Thus, the host computer analyzes and verifies the radar acquisition parameters, video data, and ADC data sent by the data acquisition device to obtain the best-performing final upgrade firmware, and then sends the final upgrade firmware to the data acquisition device, which upgrades the parameters of the millimeter-wave radar to achieve tuning, obtaining the tuning result. This solves the problem of insufficient accuracy and efficiency of manual debugging in existing millimeter-wave radar signal processing parameter tuning processes, improving the efficiency of parameter tuning. Attached Figure Description

[0042] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this application and, together with the description, serve to explain the principles of this application.

[0043] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, for those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0044] Figure 1 This is a flowchart illustrating the parameter tuning method of this application in Embodiment 1.

[0045] Figure 2 This is a schematic diagram illustrating the image ground truth annotation involved in the parameter tuning method of this application;

[0046] Figure 3 The parameter tuning method of this application involves projecting the coordinates of point cloud data in the radar coordinate system onto a 2D image plane;

[0047] Figure 4 This is a schematic diagram of the 2D ground truth bounding box projected onto the radar coordinate system for the parameter tuning method of this application;

[0048] Figure 5 This is a flowchart illustrating Embodiment 2 of the parameter tuning method of this application;

[0049] Figure 6 A simplified flowchart illustrating the parameter tuning method provided in Embodiment 2 of this application;

[0050] Figure 7 This is a schematic diagram of the module structure of the parameter optimization device in an embodiment of this application;

[0051] Figure 8 This is a schematic diagram of the device structure of the hardware operating environment involved in the parameter tuning method in the embodiments of this application.

[0052] The purpose, features, and advantages of this application will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0053] It should be understood that the specific embodiments described herein are merely illustrative of the technical solutions of this application and are not intended to limit this application.

[0054] To better understand the technical solution of this application, a detailed description will be provided below in conjunction with the accompanying drawings and specific implementation methods.

[0055] The main solution of this application embodiment is as follows: Based on the radar acquisition parameters, generate several sets of optimization parameter values ​​through optimization parameters; extract frames from the video data to obtain an image set, and annotate the image set to obtain the scene target ground truth; write the several sets of optimization parameter values ​​into a configuration file, and compile based on the configuration file to obtain several upgrade firmware; send the several upgrade firmware to the data acquisition device; upgrade the millimeter-wave radar according to the several upgrade firmware, and after each upgrade, feed the ADC data back to the millimeter-wave radar, and process the ADC data back to the millimeter-wave radar to obtain several millimeter-wave radar point cloud data sets; receive the several millimeter-wave radar point cloud data sets sent by the data acquisition device, and match them with the scene target ground truth based on the several millimeter-wave radar point cloud data sets to obtain several matching results; evaluate the several matching results to obtain several evaluation results, and determine the final upgrade firmware based on the several evaluation results. The video data is frame-sampling based on a preset frame rate to obtain an image set. The image set is then annotated using a labeling tool to obtain annotation information, including the detected target, object bounding box, object category, and spatial location information. A scene target ground truth is generated based on the annotation information. This scene target ground truth is projected onto the radar coordinate system through coordinate system transformation to obtain the position information of the scene target ground truth in the radar coordinate system. Based on this position information, the several millimeter radar point cloud data groups are matched using maximum weight matching until each group of millimeter radar point cloud data is matched, resulting in several matching results. The detection rate, false detection rate, and distance cost of each matching result are calculated. Based on these results, the matching results are evaluated using preset evaluation weights to obtain several evaluation results. The final upgrade firmware among the several upgrade firmware is determined based on these evaluation results. Based on the radar acquisition parameters of the millimeter-wave radar, ADC data is converted through the millimeter-wave radar acquisition mode, and video data is acquired through the camera device. The radar acquisition parameters, video data, and ADC data are sent to a host computer, which generates a final upgrade firmware based on the radar acquisition parameters, video data, and ADC data, and sends the final upgrade firmware to the data acquisition device. Based on the final upgrade firmware, the millimeter-wave radar parameters are upgraded to achieve optimization, resulting in an optimized tuning result. This solves the problems of insufficient accuracy and efficiency in manual tuning of existing millimeter-wave radar signal processing parameters, achieving parameter optimization and improving the efficiency of parameter tuning.Based on the present invention, addressing the problem that traditional parameter setting relies on human experience, involves extensive experimentation and debugging to determine optimal parameters, and data feedback depends on repeated manual operations, making it difficult to obtain the optimal parameter combination and resulting in low efficiency, a parameter tuning method is designed. The effectiveness of the parameter tuning method of the present invention is verified during parameter tuning, and the efficiency of parameter tuning is significantly improved by the present invention.

[0056] In this embodiment, for ease of description, the parameter tuning device will be used as the execution subject in the following description.

[0057] Due to the rapid development of autonomous driving, intelligent transportation, and industrial automation technologies, millimeter-wave radar, as an important sensor capable of target detection and tracking in complex environments, is widely used in various fields such as automotive collision avoidance, lane keeping, and drone obstacle avoidance. Millimeter-wave radar has strong penetration capabilities and can operate in adverse weather conditions, such as heavy fog, rain, and snow. Therefore, its importance in autonomous driving systems is becoming increasingly prominent. In the application of millimeter-wave radar, signal processing algorithms have a significant impact on its detection performance. Signal processing parameters for millimeter-wave radar include filtering and detection thresholds. The selection of these parameters directly affects the radar's detection capability, false detection rate, and target positioning accuracy. Traditional parameter settings rely on human experience, requiring extensive experimentation and debugging to determine the optimal parameters. Furthermore, data feedback depends on repeated manual operations. This manual debugging process is not only time-consuming and labor-intensive, but also makes it difficult to obtain the optimal parameter combination in complex scenarios, resulting in the millimeter-wave radar's detection performance not being fully utilized. Moreover, because experimental results in real-world scenarios lack effective true values, the results cannot be effectively evaluated, leading to a decrease in radar detection efficiency.

[0058] This application provides a solution in which a host computer receives radar acquisition parameters, video data, and mode conversion ADC data sent by a data acquisition device. Simultaneously, the obtained parameters and data are used to determine the optimal parameters, and the parameters are used to generate the final upgrade firmware. The data acquisition device then upgrades the parameters of the millimeter-wave radar based on the final upgrade firmware to achieve optimization, thereby obtaining the optimization results and providing users with better services.

[0059] As can be seen from the above embodiments, this application receives radar acquisition parameters, video data, and mode conversion ADC data sent by a data acquisition device; generates a final upgrade firmware based on the radar acquisition parameters, video data, and ADC data; sends the final upgrade firmware to the data acquisition device, and the data acquisition device upgrades the parameters of its millimeter-wave radar according to the final upgrade firmware to achieve optimization, obtaining the optimization result. Thus, the host computer analyzes and verifies the radar acquisition parameters, video data, and ADC data sent by the data acquisition device to obtain the best-performing final upgrade firmware, and then sends the final upgrade firmware to the data acquisition device. The data acquisition device then upgrades the parameters of the millimeter-wave radar to achieve optimization, obtaining the optimization result. This solves the problem of insufficient accuracy and efficiency of manual debugging in the existing millimeter-wave radar signal processing parameter optimization process, improving the efficiency of parameter optimization.

[0060] It should be noted that the executing entity in this embodiment can be a computing service device with data processing, network communication, and program execution functions, such as a tablet computer, personal computer, or mobile phone, or an electronic device or parameter tuning device capable of performing the above functions. The following description uses a parameter tuning device as an example to illustrate this embodiment and the subsequent embodiments.

[0061] Based on this, the embodiments of this application provide a parameter tuning method, referring to... Figure 1 , Figure 1 This is a flowchart illustrating the first embodiment of the parameter tuning method of this application.

[0062] In this embodiment, the parameter tuning method is applied to a host computer, and the method includes steps S01 to S03:

[0063] Step S01: Receive radar acquisition parameters, video data, and mode conversion ADC data sent by the data acquisition device;

[0064] Before the present embodiment begins to be described, it should be clear that in the prior art, with the rapid development of autonomous driving, intelligent transportation and industrial automation technologies, millimeter-wave radar, as an important sensor that can achieve target detection and tracking in complex environments, is widely used in many fields such as automobile collision avoidance, lane keeping and drone obstacle avoidance. Millimeter-wave radar has strong penetration capabilities and can work in adverse weather conditions, such as heavy fog, rain and snow. Therefore, its importance in autonomous driving systems is becoming increasingly prominent.

[0065] In practical applications of millimeter-wave radar, signal processing algorithms have a significant impact on its detection performance. Signal processing parameters for millimeter-wave radar include filtering and detection thresholds. The selection of these parameters directly affects the radar's detection capability, false detection rate, and target positioning accuracy. Traditional parameter settings rely on manual experience, requiring extensive experimental debugging to determine the optimal parameters. Furthermore, data feedback depends on repeated manual operations. This manual debugging process is not only time-consuming and labor-intensive, but also makes it difficult to obtain the optimal parameter combination in complex scenarios. Consequently, the detection performance of millimeter-wave radar cannot be fully utilized. Moreover, due to the lack of effective true values ​​in real-world experimental results, effective evaluation of the results is impossible.

[0066] Therefore, in this embodiment, the host computer receives radar acquisition parameters, video data, and mode conversion ADC data sent by the data acquisition device. The host computer refers to the central computer or control unit in the vehicle system used to manage and monitor various on-board devices, control systems, and data processing. It is responsible for coordinating and processing the operations between different subsystems in the vehicle to ensure the efficient and safe operation of the vehicle system. The data acquisition device refers to various sensors in the vehicle. In this embodiment, the data acquisition device also includes millimeter-wave radar, data acquisition board, and camera equipment. In the vehicle system, the host computer and the data acquisition device establish a normal bidirectional communication link (a bidirectional communication link refers to a communication method that enables two devices to transmit data to each other. In the vehicle system, bidirectional communication links are usually used between various electronic control units (ECUs) of the vehicle to ensure that data can be transmitted bidirectionally between the host computer and the slave computer).

[0067] Step S02: Generate the final upgrade firmware based on the radar acquisition parameters, video data, and ADC data;

[0068] After obtaining the radar acquisition parameters, video data, and ADC data from the data acquisition device, corresponding analysis can be performed. Among them, the application of radar acquisition parameters in vehicle systems is mainly for environmental perception and obstacle detection. Especially in autonomous driving and advanced driver assistance systems (ADAS), radar is a key technology. Radar sensors acquire information about the surrounding environment by emitting electromagnetic waves and receiving return signals. The parameters acquired are crucial for subsequent decision-making and control. The focus of this embodiment is to optimize the radar acquisition parameters, while the video data and ADC data are used for subsequent verification of the effectiveness of the radar acquisition parameters.

[0069] Before obtaining the final upgrade firmware, this embodiment also needs to generate numerous sets of tuning parameter values ​​to generate the upgrade firmware. The generated upgrade firmware is then used to iteratively test the millimeter-wave radar. After all the upgrade firmware has been tested, the final upgrade firmware with the best performance can be obtained.

[0070] Step S03: The final upgrade firmware is sent to the data acquisition device, and the data acquisition device upgrades the parameters of the millimeter-wave radar according to the final upgrade firmware to achieve optimization and obtain the optimization result.

[0071] Once the final upgrade firmware is determined, it is sent to the data acquisition device. The data acquisition device then upgrades the parameters of the millimeter-wave radar according to the final upgrade firmware to achieve optimization and obtain the optimization results. Millimeter-wave radar (MMW radar) is a radar system that operates in the millimeter-wave frequency band (usually between 30 GHz and 300 GHz). Due to its shorter wavelength, millimeter-wave radar can provide higher resolution and is therefore widely used in autonomous driving, vehicle driver assistance systems (ADAS), radar imaging, object detection and tracking, and other fields. The working principle of millimeter-wave radar is similar to that of traditional radar. It detects the position, speed and distance of the target by emitting high-frequency electromagnetic waves (usually millimeter waves) and receiving the signals reflected from the target. Because of the high frequency of millimeter-wave radar, its wavelength is very short (usually between 1 mm and 10 mm), which gives it high spatial resolution and allows for more accurate identification of small objects, details and targets in complex environments.

[0072] Specifically, step S02 above, which involves generating the final upgrade firmware based on the radar acquisition parameters, video data, and ADC data, includes:

[0073] Step S021: Based on the radar acquisition parameters, generate several sets of optimization parameter values ​​through optimization parameters;

[0074] Step S022: Extract frames from the video data to obtain an image set, and annotate the image set to obtain the scene target ground truth.

[0075] Step S023: Write the several sets of tuning parameter values ​​into a configuration file, and compile based on the configuration file to obtain several upgrade firmware;

[0076] Step S024: Send the aforementioned upgraded firmware to the data acquisition device;

[0077] Step S025: Upgrade the millimeter-wave radar according to the several upgrade firmware, and feed back the ADC data to the millimeter-wave radar after each upgrade. The millimeter-wave radar processes the ADC data to obtain several millimeter-wave radar point cloud data groups.

[0078] Step S026: Receive several millimeter-wave radar point cloud data groups sent by the data acquisition device; match the scene target ground truth values ​​based on the several millimeter-wave radar point cloud data groups to obtain several matching results.

[0079] Step S027: Evaluate the several matching results to obtain several evaluation results, and determine the final upgrade firmware based on the several evaluation results.

[0080] To obtain more accurate acquisition parameters, this embodiment generates multiple sets of parameter values ​​for parameters to be tuned through an optimization algorithm (the optimization algorithm refers to a genetic algorithm). Then, each set of parameter values ​​is written into a configuration file in sequence, and the source program is compiled to obtain the firmware to be upgraded. The compilation operation, OTA upgrade, and ADC data backfeed on the host computer can all be operated using an automated window control tool written in Python. The OTA upgrade pushes the firmware with the new parameters to the millimeter-wave radar device for upgrade. By backfeeding the original ADC data, point cloud-level data under the new parameter combination can be obtained to verify the effect of millimeter-wave radar signal processing under different parameter combinations.

[0081] To verify the effectiveness of parameter acquisition, this embodiment also obtains the true value of the scene target for effect verification. After the upgrade firmware is sent to the millimeter-wave radar and the ADC data is fed back, the millimeter-wave radar in the data acquisition device can be used to process the ADC data with new parameters to obtain the millimeter-wave radar point cloud data group for each upgrade firmware. Then the obtained millimeter-wave radar point cloud data group is sent to the host computer.

[0082] The host computer performs matching and evaluation based on the millimeter-wave radar point cloud data set and the scene target ground truth to obtain the parameter set with the best effect, and determines the upgrade firmware corresponding to the parameter set with the best effect as the final upgrade firmware.

[0083] More specifically, step S22 above, which involves extracting frames from the video data to obtain an image set and labeling the image set to obtain the ground truth value of the scene target, includes the following specific steps:

[0084] Step S0221: Extract frames from the video data based on a preset frame rate to obtain an image set;

[0085] Step S0222: The image set is annotated using an annotation tool to obtain annotation information, which includes the detected target, object bounding box, object category, and spatial location information.

[0086] Step S0223: Generate the scene target ground truth value based on the annotation information.

[0087] In fields such as object detection, autonomous driving, and computer vision, "ground truth" refers to the true labeled data of the location, category, and related attributes of objects in a specific scene or image. It is usually obtained through manual annotation or other precise measurement methods and serves as the standard answer for training, validating, and testing machine learning models. Therefore, to obtain ground truth for result verification, this embodiment performs frame extraction on the video data at a fixed frame rate to obtain several representative video frame images, ensuring that the number of frames is dense enough to cover the entire time period of interest. After obtaining the synchronized video frames, such as... Figure 2 As shown, professional annotation tools are used to annotate the scene targets in each frame of the image. The annotated targets are the main detection objects of the millimeter-wave radar, such as vehicles and pedestrians. The ground truth data of each target should include the accurate object bounding box, category, and spatial location information. These annotated images will serve as reference standards for subsequent algorithm optimization.

[0088] Furthermore, in order to obtain the optimal radar parameters, this embodiment performs matching and evaluation on the millimeter-wave radar point cloud data sets. Therefore, step S026 above, which involves matching the scene target ground truth values ​​with the several millimeter-wave radar point cloud data sets to obtain several matching results, includes:

[0089] Step S0261: Project the true value of the scene target to the radar coordinate system through coordinate system transformation to obtain the position information of the true value of the scene target in the radar coordinate system;

[0090] Step S0262: Based on the location information, the several millimeter radar point cloud data groups are matched by maximum weight matching until each group of millimeter radar point cloud data in the several millimeter radar point cloud data groups is matched, and several matching results are obtained.

[0091] Point cloud data is obtained by feeding back the original ADC data. By comparing the timestamps of the video data and the millimeter-wave radar point cloud data, the millimeter-wave radar point cloud data corresponding to each frame of the video image is accurately indexed. The ground truth data of the target in the video image uses the image coordinate system, while the millimeter-wave radar point cloud data is in the radar coordinate system. Through coordinate system transformation, the bounding box of the target ground truth in the image coordinate system is projected onto the radar coordinate system. Based on the position information of the ground truth bounding box in the radar coordinate system and the point cloud position information, the maximum weight matching method is used to perform a one-to-one matching between the point cloud and the ground truth, and the matching result of each set of millimeter-wave point cloud data is obtained.

[0092] In this embodiment, the specific matching can be as follows: After calibrating the intrinsic and extrinsic parameter matrices, the projection of the 3D point in the world coordinate system onto the image coordinate system is calculated. However, the matrix operation of transforming from the image coordinate system to the radar coordinate system is irreversible. Therefore, a transmission transformation is used instead of the inverse operation. First, a frame of example point cloud data is used to calculate the corresponding 2D coordinates in the image coordinate system through intrinsic and extrinsic parameters, resulting in multiple 2D coordinate and radar coordinate matching pairs. Then, using the 2D coordinates as the source coordinates and the radar coordinates as the target coordinates, the transmission transformation matrix H is calculated to obtain the coordinate mapping relationship from 2D to radar point cloud. The coordinates of the point cloud in the radar coordinate system are projected onto the 2D image plane as shown below. Figure 3 As shown, the red dots represent radar point clouds, and the 2D ground truth bounding boxes projected onto the radar coordinate system are as follows: Figure 4 As shown in the figure, the purple polygon represents the ground truth bounding box, the blue triangle represents the center of the ground truth bounding box, the red dot represents the radar point cloud, and the marked red dot represents the radar point cloud that successfully matches the ground truth bounding box.

[0093] It should be noted that after obtaining the position information of the ground truth bounding box coordinate transformation, a one-to-one matching between the point cloud and the ground truth bounding box is performed using the maximum weight matching method based on the position information of the ground truth bounding box in the radar coordinate system and the point cloud position information. The entire matching process can be divided into two maximum weight matching operations. The detailed process of the first maximum weight matching is as follows:

[0094] (1) Graph construction: Create an undirected graph to represent the relationship between ground truth bounding boxes and point clouds. Traverse each ground truth bounding box and create a node for each bounding box. The node is named gt_x, where x is the index of the ground truth.

[0095] (2) Edge addition: Traverse all point clouds. For each point cloud, check if it is within a ground truth bounding box. If the point cloud is within a ground truth bounding box, calculate the distance between the point and the center of the bounding box, and calculate a score based on the distance.

[0096]

[0097] If the point cloud is not within the bounding box of the ground truth node, score = None. For points with scores other than None, add an edge from the ground truth node to the point cloud node in the graph, with the weight of the edge being the calculated score.

[0098] (3) Maximum Weight Matching: The maximum weight matching method is used to find the maximum weight match and obtain the point cloud index that matches the ground truth bounding box. The objects of the second maximum weight matching are the ground truth bounding boxes and point clouds that failed to match in the first match. The rules for adding edges are changed in the second match. Specifically, the length of the preset search radius is determined according to the category of the target ground truth box. For large targets such as trucks or buses, a larger search radius is set. Then, with the ground truth box as the center, for each point cloud that has not yet been successfully matched, it is checked whether its coordinates are inside the circle established by a ground truth bounding box. If the point cloud is inside the ground truth bounding box, the distance between the point and the center of the bounding box is calculated, and a score is calculated based on the distance.

[0099]

[0100] If the point cloud is not inside the circle created by the ground truth bounding box, then score = None. For points with scores other than None, add an edge from the ground truth node to the point cloud node in the graph, with the weight of the edge being the calculated score.

[0101] The above step S027, which evaluates the several matching results to obtain several evaluation results, and determines the final upgrade firmware based on the several evaluation results, includes:

[0102] Step S0271: Calculate the detection rate, false detection rate, and distance cost for each matching result.

[0103] Step S0272: Based on the detection rate, false detection rate and distance cost of the several matching results, evaluate the several matching results through preset evaluation weights to obtain several evaluation results;

[0104] Step S0273: Determine the final upgrade firmware among the several upgrade firmware based on the several evaluation results.

[0105] The evaluation results of the new set of parameters are calculated based on the matching results of each frame. The evaluation results are mainly divided into three indicators: detection rate, false detection rate, and distance cost. The detection rate refers to the proportion of the target ground truth that is successfully matched with the point cloud. This indicator is used to evaluate the radar's effective detection capability for the target under the current parameter settings. The higher the value, the more accurately the radar can detect the target. The false detection rate refers to the proportion of the point cloud that is not successfully matched with the ground truth target. This indicator reflects the false detection situation in radar signal processing, that is, whether the radar has detected incorrect point clouds or noise. The lower the false detection rate, the cleaner and more accurate the radar data under the parameter settings. The distance cost is the sum of the distances of the point cloud from the center of the ground truth box in each matching pair. The smaller the distance cost, the more accurately the point cloud data reflects the position of the target ground truth.

[0106] The distance cost calculation formula is as follows:

[0107]

[0108] Among them, dist(P i G i ) represents the point cloud point P in the i-th matching pair. i With the true value of the target G i The smaller the distance cost, the more accurately the point cloud data reflects the true location of the target.

[0109] This embodiment also describes the evaluation weights, that is, for different usage needs, corresponding weights can be set for detection rate, false detection rate and distance cost to obtain radar detection parameters that better meet user needs. In addition, this embodiment can also use judgment rules. The optimal judgment rule is to prioritize the parameter with the highest detection rate. If the detection rates are the same, then select the parameter with the lowest false detection rate. If the false detection rates are also the same, then select the parameter with the lowest distance cost.

[0110] This embodiment, through the above-described scheme, specifically receives radar acquisition parameters, video data, and mode conversion ADC data sent by a data acquisition device; generates a final upgrade firmware based on the radar acquisition parameters, video data, and ADC data; sends the final upgrade firmware to the data acquisition device, which then upgrades the parameters of its millimeter-wave radar according to the final upgrade firmware to achieve optimization, obtaining the optimization result. Thus, the host computer analyzes and verifies the radar acquisition parameters, video data, and ADC data sent by the data acquisition device to obtain the best-performing final upgrade firmware, and then sends the final upgrade firmware to the data acquisition device, which then upgrades the parameters of the millimeter-wave radar to achieve optimization, obtaining the optimization result. This solves the problem of insufficient accuracy and efficiency of manual debugging in the existing millimeter-wave radar signal processing parameter optimization process, improving the efficiency of parameter optimization.

[0111] Based on the first embodiment of this application, in the second embodiment of this application, the content that is the same as or similar to that in the first embodiment described above can be referred to the above description, and will not be repeated hereafter. Based on this, please refer to... Figure 5 The parameter optimization method is also applied to data acquisition equipment, which includes millimeter-wave radar and camera equipment. The method includes steps S04 to S06:

[0112] Step S04: Based on the radar acquisition parameters of the millimeter-wave radar, convert the ADC data through the millimeter-wave radar acquisition mode, and acquire video data through the camera device;

[0113] Step S05: The radar acquisition parameters, video data, and ADC data are sent to the host computer, which generates the final upgrade firmware based on the radar acquisition parameters, video data, and ADC data, and sends the final upgrade firmware to the data acquisition device.

[0114] Step S06: Based on the final upgraded firmware, upgrade the parameters of the millimeter-wave radar to achieve optimization and obtain the optimization result.

[0115] It should be clear that the data acquisition device and the host computer in this embodiment together form a general control system, and the two cooperate with each other. The data acquisition device in this embodiment includes millimeter-wave radar, data acquisition board and camera equipment, etc.

[0116] In this embodiment, the control software of the host computer is used to configure the radar acquisition parameters to ensure that the millimeter-wave radar can correctly output the raw ADC data. At the same time, video data is captured in real time. The host computer receives the raw ADC data of the millimeter-wave radar and the video data of the camera in real time and stores the data in the local file system for subsequent processing. Among them, ADC (Analog-to-Digital Converter) is an electronic device that converts continuous analog signals into discrete digital signals. In electronic systems, ADC is a key component that converts analog signals (such as temperature, voltage, sound, etc.) into digital formats that can be processed by digital processors (such as microcontrollers, computers, etc.).

[0117] The basic principle of ADC is:

[0118] ① Input signal: Usually an analog signal, such as voltage;

[0119] ② Sampling: The ADC samples the analog signal at fixed time intervals;

[0120] ③ Quantization: After sampling, the amplitude of the analog signal is divided into several discrete values. The accuracy of quantization is determined by the resolution of the ADC.

[0121] ④ Encoding: Convert the quantized signal into a digital signal in binary format.

[0122] The video data and ADC data are then sent to the host computer, which iteratively optimizes the radar acquisition parameters, video data, and the ADC and millimeter-wave radar to determine the upgrade firmware corresponding to the acquisition parameters with the best effect as the final upgrade firmware.

[0123] Finally, the data acquisition equipment adjusts the acquisition parameters of the millimeter-wave radar through a final firmware upgrade to obtain the optimization results.

[0124] To help understand the implementation process of the parameter tuning method obtained in this embodiment combined with the above embodiment one, please refer to... Figure 6 , Figure 6 A simplified flowchart of a parameter tuning method is provided, specifically:

[0125] Step 1: Enable the host computer to communicate normally with the millimeter-wave radar and camera, and use the millimeter-wave radar and camera to collect the raw ADC data and video data of the millimeter-wave radar;

[0126] Step 2: Extract multiple frames from the video and index the millimeter-wave radar ADC data at the corresponding time according to the timestamp, and label the scene target true value on the extracted frame images;

[0127] Step 3: Generate parameter values ​​for the parameters to be tuned using an optimization algorithm and write them to a configuration file. Then, use automated tools to perform compilation, OTA upgrades, and ADC data refeedback.

[0128] Step 4: Match the point cloud data obtained from the data backfeeding with the target ground truth using the maximum weight matching method;

[0129] Step 5: Evaluate the matching results and provide the evaluation results for this set of parameter values;

[0130] Step 6: Jump back to step 3, iterate again until the termination condition is met, save the parameter values ​​of the optimal evaluation result, and output the final upgrade firmware.

[0131] This embodiment, through the above-described scheme, specifically utilizes the radar acquisition parameters of the millimeter-wave radar, converts ADC data through the millimeter-wave radar acquisition mode, and acquires video data through the camera device. The radar acquisition parameters, video data, and ADC data are sent to a host computer, which generates a final upgrade firmware based on these parameters and sends it to the data acquisition device. Based on the final upgrade firmware, the millimeter-wave radar parameters are upgraded to achieve optimization, yielding the optimization result. Thus, the host computer analyzes and verifies the radar acquisition parameters, video data, and ADC data sent by the data acquisition device to obtain the best-performing final upgrade firmware. This final upgrade firmware is then sent to the data acquisition device, which upgrades the millimeter-wave radar parameters to achieve optimization, obtaining the optimization result. This solves the problem of insufficient accuracy and efficiency in manual debugging during existing millimeter-wave radar signal processing parameter optimization processes, improving the efficiency of parameter optimization.

[0132] It should be noted that the above examples are only for understanding this application and do not constitute a limitation on the parameter tuning method of this application. Any simple modifications based on this technical concept are within the protection scope of this application.

[0133] This application also provides a parameter tuning device, please refer to... Figure 7 The parameter tuning device is applied to the host computer, and the device includes:

[0134] The receiving module is used to receive radar acquisition parameters, video data, and ADC data sent by the data acquisition device.

[0135] The generation module is used to generate the final upgrade firmware based on the radar acquisition parameters, video data, and ADC data.

[0136] The tuning module is used to send the final upgrade firmware to the data acquisition device, and the data acquisition device upgrades the parameters of the millimeter-wave radar of the data acquisition device according to the upgrade firmware to achieve tuning and obtain the tuning result.

[0137] The parameter tuning device provided in this application, employing the parameter tuning method described in the above embodiments, can solve the technical problems of insufficient accuracy and efficiency in manual tuning during existing millimeter-wave radar signal processing parameter tuning. Compared with the prior art, the beneficial effects of the parameter tuning device provided in this application are the same as those of the parameter tuning method provided in the above embodiments, and other technical features in the parameter tuning device are the same as those disclosed in the methods of the above embodiments, and will not be repeated here.

[0138] This application provides a parameter tuning device, which includes: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, which are executed by the at least one processor to enable the at least one processor to perform the parameter tuning method in the first embodiment described above.

[0139] The following is for reference. Figure 8 The diagram illustrates a structural schematic suitable for implementing the parameter tuning device in the embodiments of this application. The parameter tuning device in the embodiments of this application may include, but is not limited to, mobile terminals such as mobile phones, laptops, digital broadcast receivers, PDAs (Personal Digital Assistants), PADs (Portable Application Description), PMPs (Portable Media Players), in-vehicle terminals (e.g., in-vehicle navigation terminals), and fixed terminals such as digital TVs and desktop computers. Figure 8The parameter tuning device shown is merely an example and should not impose any limitations on the functionality and scope of use of the embodiments of this application.

[0140] like Figure 8 As shown, the parameter tuning device may include a processing unit 1001 (e.g., a central processing unit, a graphics processing unit, etc.) that can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 1002 or a program loaded from a storage device 1003 into a random access memory (RAM) 1004. The RAM 1004 also stores various programs and data required for the operation of the parameter tuning device. The processing unit 1001, ROM 1002, and RAM 1004 are interconnected via a bus 1005. An input / output (I / O) interface 1006 is also connected to the bus. Typically, the following systems can be connected to the I / O interface 1006: input devices 1007 including, for example, a touchscreen, touchpad, keyboard, mouse, image sensor, microphone, accelerometer, gyroscope, etc.; output devices 1008 including, for example, a liquid crystal display (LCD), speaker, vibrator, etc.; storage devices 1003 including, for example, magnetic tape, hard disk, etc.; and communication devices 1009. Communication device 1009 allows the parameter tuning device to communicate wirelessly or wiredly with other devices to exchange data. Although the figure shows parameter tuning devices with various systems, it should be understood that implementation or possession of all the systems shown is not required. More or fewer systems may be implemented alternatively.

[0141] Specifically, according to the embodiments disclosed in this application, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, embodiments disclosed in this application include a computer program product comprising a computer program carried on a computer-readable medium, the computer program containing program code for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via a communication device, or installed from storage device 1003, or installed from ROM 1002. When the computer program is executed by processing device 1001, it performs the functions defined in the methods of the embodiments disclosed in this application.

[0142] The parameter tuning device provided in this application, employing the parameter tuning method described in the above embodiments, can solve the technical problems of insufficient accuracy and efficiency in manual tuning during existing millimeter-wave radar signal processing parameter tuning. Compared with the prior art, the beneficial effects of the parameter tuning device provided in this application are the same as those of the parameter tuning method provided in the above embodiments, and other technical features of this parameter tuning device are the same as those disclosed in the previous embodiment method, and will not be repeated here.

[0143] It should be understood that the various parts disclosed in this application can be implemented using hardware, software, firmware, or a combination thereof. In the description of the above embodiments, specific features, structures, materials, or characteristics can be combined in any suitable manner in one or more embodiments or examples.

[0144] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

[0145] This application provides a computer-readable storage medium having computer-readable program instructions (i.e., a computer program) stored thereon, the computer-readable program instructions being used to execute the parameter tuning method in the above embodiments.

[0146] The computer-readable storage medium provided in this application may be, for example, a USB flash drive, but is not limited to, electrical, magnetic, optical, electromagnetic, infrared, or semiconductor systems, devices, or any combination thereof. More specific examples of computer-readable storage media may include, but are not limited to: electrical connections having 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 fiber, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof. In this embodiment, the computer-readable storage medium may be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, system, or device. The program code contained on the computer-readable storage medium may be transmitted using any suitable medium, including but not limited to: wires, optical cables, RF (Radio Frequency), etc., or any suitable combination thereof.

[0147] The aforementioned computer-readable storage medium may be included in the parameter tuning device; or it may exist independently and not be assembled into the parameter tuning device.

[0148] The aforementioned computer-readable storage medium carries one or more programs. When these programs are executed by the parameter tuning device, the parameter tuning device: receives radar acquisition parameters, video data, and mode conversion ADC data sent by the data acquisition device; generates a final upgrade firmware based on the radar acquisition parameters, video data, and ADC data; sends the final upgrade firmware to the data acquisition device, and the data acquisition device upgrades the parameters of its millimeter-wave radar according to the final upgrade firmware to achieve tuning, thereby obtaining a tuning result.

[0149] Computer program code for performing the operations of this application can be written in one or more programming languages ​​or a combination thereof, including object-oriented programming languages ​​such as Java, Smalltalk, and C++, and conventional procedural programming languages ​​such as the "C" language or similar programming languages. The program code can be executed entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving remote computers, the remote computer can be connected to the user's computer via any type of network—including a Local Area Network (LAN) or a Wide Area Network (WAN)—or can be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0150] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of this application. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, can be implemented using a dedicated hardware-based system that performs the specified function or operation, or using a combination of dedicated hardware and computer instructions.

[0151] The modules described in the embodiments of this application can be implemented in software or hardware. The names of the modules do not necessarily limit the functionality of the unit itself.

[0152] The readable storage medium provided in this application is a computer-readable storage medium that stores computer-readable program instructions (i.e., a computer program) for executing the above-described parameter tuning method. This solves the technical problems of insufficient accuracy and efficiency in manual tuning during existing millimeter-wave radar signal processing parameter tuning processes. Compared with the prior art, the beneficial effects of the computer-readable storage medium provided in this application are the same as those of the parameter tuning method provided in the above embodiments, and will not be repeated here.

[0153] This application also provides a computer program product, including a computer program that, when executed by a processor, implements the steps of the parameter tuning method described above.

[0154] The computer program product provided in this application can solve the technical problems of insufficient accuracy and efficiency in manual tuning of existing millimeter-wave radar signal processing parameters. Compared with the prior art, the beneficial effects of the computer program product provided in this application are the same as those of the parameter tuning method provided in the above embodiments, and will not be repeated here.

[0155] The above description is only a part of the embodiments of this application and does not limit the patent scope of this application. All equivalent structural transformations made under the technical concept of this application and using the contents of the specification and drawings of this application, or direct / indirect applications in other related technical fields, are included in the patent protection scope of this application.

Claims

1. A parameter tuning method, characterized in that, The method is applied to a host computer, and the method includes: Receive radar acquisition parameters, video data, and mode conversion ADC data sent by the data acquisition equipment; The final upgrade firmware is generated based on the radar acquisition parameters, video data, and ADC data. The step of generating the final upgrade firmware based on the radar acquisition parameters, video data, and ADC data includes: Based on the radar acquisition parameters, several sets of optimization parameter values ​​are generated by optimizing the parameters. The video data is frame-by-frame extracted to obtain an image set, and the image set is labeled to obtain the scene target ground truth. The steps of extracting frames from the video data to obtain an image set, and then labeling the image set to obtain the ground truth value of the scene target include: The video data is frame-by-frame extracted based on a preset frame rate to obtain an image set; The image set is annotated using an annotation tool to obtain annotation information, which includes the detected target, object bounding box, object category, and spatial location information. Generate scene target truth values ​​based on the annotation information; Write the aforementioned sets of tuning parameter values ​​into a configuration file, and compile based on the configuration file to obtain several upgrade firmwares; Send the aforementioned upgraded firmware to the data acquisition device; The millimeter-wave radar is upgraded according to the aforementioned firmware, and the ADC data is fed back to the millimeter-wave radar after each upgrade. The millimeter-wave radar is used to process the ADC data to obtain several millimeter-wave radar point cloud data groups. The system receives several millimeter-wave radar point cloud data sets sent by the data acquisition device, and matches them with the scene target ground truth based on the several millimeter-wave radar point cloud data sets to obtain several matching results. The step of matching the scene target ground truth values ​​with the plurality of millimeter-wave radar point cloud data groups to obtain a plurality of matching results includes: By transforming the coordinate system, the true value of the scene target is projected onto the radar coordinate system to obtain the position information of the true value of the scene target in the radar coordinate system; Based on the location information, the several millimeter radar point cloud data groups are matched by maximum weight matching until each millimeter radar point cloud data group in the several millimeter radar point cloud data groups is matched, and several matching results are obtained. The step of matching the plurality of millimeter radar point cloud data groups based on the location information using maximum weight matching includes: Create an undirected graph to represent the relationship between ground truth bounding boxes and point clouds. Traverse each ground truth bounding box and create a node for each bounding box, named gt_x, where x is the index of the ground truth. Iterate through all point clouds and check whether each point cloud is within a bounding box of a certain truth value. If the point cloud is within the bounding box of the ground truth, then calculate the distance between the point cloud and the center of the bounding box. And calculate a score based on the distance. ,in, If the point cloud is not within the bounding box of the truth value, then For ratings that are not For each point, add an edge from the truth node to the point cloud node in the graph, with the weight of the edge being the calculated score. The maximum weight matching method is used to find the maximum weight match and obtain the point cloud index that matches the ground truth bounding box. The object of the second maximum weight matching is the ground truth bounding box and point cloud that failed to match in the first match. The second maximum weight matching specifically determines the length of the preset search radius according to the category of the target ground truth box. A larger search radius will be set for large targets. Using the truth box as the center, for each point cloud that has not yet been successfully matched, check whether its coordinates are inside the circle established by a certain truth bounding box; If the point cloud is within the bounding box of the ground truth, then calculate the distance between the point cloud and the center of the bounding box. And calculate a score based on the distance. ,in, If the point cloud is not inside the circle defined by the truth bounding box, then For ratings that are not For each point, add an edge from the truth node to the point cloud node in the graph, with the weight of the edge being the calculated score. The matching results are evaluated to obtain several evaluation results, and the final upgrade firmware is determined based on the several evaluation results; The final upgrade firmware is sent to the data acquisition device, which then upgrades its millimeter-wave radar according to the final upgrade firmware to optimize parameters and obtain the optimization results.

2. The method as described in claim 1, characterized in that, The step of evaluating the several matching results to obtain several evaluation results, and determining the final upgrade firmware based on the several evaluation results, includes: The detection rate, false detection rate, and distance cost of each matching result are calculated. Based on the detection rate, false detection rate, and distance cost of the several matching results, the several matching results are evaluated using preset evaluation weights to obtain several evaluation results; The final upgrade firmware among the several upgrade firmware is determined based on the evaluation results.

3. A parameter tuning method, characterized in that, The method is applied to a data acquisition device, which includes millimeter-wave radar and camera equipment, and the method includes: Based on the radar acquisition parameters of the millimeter-wave radar, ADC data is converted through the acquisition mode of the millimeter-wave radar, and video data is acquired through the camera device; The radar acquisition parameters, video data, and ADC data are sent to the host computer, which then generates the final upgrade firmware based on the radar acquisition parameters, video data, and ADC data, and sends the final upgrade firmware to the data acquisition device. The step of having the host computer generate the final upgrade firmware based on the radar acquisition parameters, video data, and ADC data, and then sending the final upgrade firmware to the data acquisition device includes: Based on the radar acquisition parameters, the host computer generates several sets of tuning parameter values ​​by optimizing the parameters. The host computer extracts frames from the video data to obtain an image set, and then labels the image set to obtain the scene target ground truth. The step of extracting frames from the video data by the host computer to obtain an image set, and then labeling the image set to obtain the ground truth value of the scene target includes: The host computer extracts frames from the video data based on a preset frame rate to obtain an image set; The host computer annotates the image set using an annotation tool to obtain annotation information, which includes the detected target, object bounding box, object category, and spatial location information. The host computer generates the scene target true value based on the annotation information; The aforementioned sets of tuning parameter values ​​are written into a configuration file, and the host computer compiles the configuration file to obtain several upgrade firmware files. The host computer sends the several upgraded firmware files to the data acquisition device; The host computer upgrades the millimeter-wave radar according to the several upgrade firmware, and after each upgrade, the ADC data is fed back to the millimeter-wave radar. The millimeter-wave radar processes the ADC data to obtain several millimeter-wave radar point cloud data groups. The host computer receives several millimeter-wave radar point cloud data sets sent by the data acquisition device. Based on the several millimeter-wave radar point cloud data sets, the host computer performs matching through the scene target ground truth to obtain several matching results. The step of obtaining several matching results by the host computer matching the scene target ground truth values ​​based on the plurality of millimeter-wave radar point cloud data groups includes: The host computer projects the true value of the scene target onto the radar coordinate system through coordinate system transformation, thereby obtaining the position information of the true value of the scene target in the radar coordinate system. Based on the location information, the host computer performs maximum weight matching on the several millimeter radar point cloud data groups until each millimeter radar point cloud data group is matched, and several matching results are obtained. The step of matching the several millimeter radar point cloud data groups by the host computer using maximum weight matching based on the location information includes: An undirected graph is created by the host computer to represent the relationship between ground truth bounding boxes and point clouds. Each ground truth bounding box is traversed, and a node is created for each bounding box. The node is named gt_x, where x is the index of the ground truth. The host computer traverses all point clouds and checks whether each point cloud is within a certain true bounding box. If the point cloud is within the bounding box of the ground truth, the host computer calculates the distance between the point cloud and the center of the bounding box. And calculate a score based on the distance. ,in, If the point cloud is not within the bounding box of the truth value, then The host computer determines whether the score is not... For each point, add an edge from the truth node to the point cloud node in the graph, with the weight of the edge being the calculated score. The host computer uses the maximum weight matching method to find the maximum weight match and obtains the point cloud index that matches the ground truth bounding box. The object of the second maximum weight matching is the ground truth bounding box and point cloud that failed to match in the first match. The second maximum weight matching specifically determines the length of the preset search radius according to the category of the target ground truth box. For large targets, a larger search radius will be set. Using the truth box as the center, the host computer checks whether the coordinates of each point cloud that has not yet been successfully matched are within the circle established by a certain truth bounding box. If the point cloud is within the bounding box of the ground truth, the host computer calculates the distance between the point cloud and the center of the bounding box. And calculate a score based on the distance. ,in, If the point cloud is not inside the circle defined by the truth bounding box, then The host computer determines whether the score is not... For each point, add an edge from the truth node to the point cloud node in the graph, with the weight of the edge being the calculated score. The host computer evaluates the matching results to obtain several evaluation results, and determines the final upgrade firmware based on the several evaluation results; Based on the final upgraded firmware, the parameters of the millimeter-wave radar are upgraded to achieve optimization, and the optimization results are obtained.

4. A parameter tuning device, characterized in that, The device is used in a host computer, and the device includes: The receiving module is used to receive radar acquisition parameters, video data, and ADC data sent by the data acquisition device. The generation module is used to generate the final upgrade firmware based on the radar acquisition parameters, video data, and ADC data. The generation module is further configured to generate several sets of optimization parameter values ​​based on the radar acquisition parameters through optimization parameters. The video data is frame-by-frame extracted to obtain an image set, and the image set is labeled to obtain the scene target ground truth. The video data is frame-by-frame extracted based on a preset frame rate to obtain an image set; The image set is annotated using an annotation tool to obtain annotation information, which includes the detected target, object bounding box, object category, and spatial location information. Generate scene target truth values ​​based on the annotation information; Write the aforementioned sets of tuning parameter values ​​into a configuration file, and compile based on the configuration file to obtain several upgrade firmwares; Send the aforementioned upgraded firmware to the data acquisition device; The millimeter-wave radar is upgraded according to the aforementioned firmware, and the ADC data is fed back to the millimeter-wave radar after each upgrade. The millimeter-wave radar is used to process the ADC data to obtain several millimeter-wave radar point cloud data groups. The system receives several millimeter-wave radar point cloud data sets sent by the data acquisition device, and matches them with the scene target ground truth based on the several millimeter-wave radar point cloud data sets to obtain several matching results. By transforming the coordinate system, the true value of the scene target is projected onto the radar coordinate system to obtain the position information of the true value of the scene target in the radar coordinate system; Based on the location information, the several millimeter radar point cloud data groups are matched by maximum weight matching until each millimeter radar point cloud data group in the several millimeter radar point cloud data groups is matched, and several matching results are obtained. Create an undirected graph to represent the relationship between ground truth bounding boxes and point clouds. Traverse each ground truth bounding box and create a node for each bounding box, named gt_x, where x is the index of the ground truth. Iterate through all point clouds and check whether each point cloud is within a bounding box of a certain truth value. If the point cloud is within the bounding box of the ground truth, then calculate the distance between the point cloud and the center of the bounding box. And calculate a score based on the distance. ,in, If the point cloud is not within the bounding box of the truth value, then For ratings that are not For each point, add an edge from the truth node to the point cloud node in the graph, with the weight of the edge being the calculated score. The maximum weight matching method is used to find the maximum weight match and obtain the point cloud index that matches the ground truth bounding box. The object of the second maximum weight matching is the ground truth bounding box and point cloud that failed to match in the first match. The second maximum weight matching specifically determines the length of the preset search radius according to the category of the target ground truth box. A larger search radius will be set for large targets. Using the truth box as the center, for each point cloud that has not yet been successfully matched, check whether its coordinates are inside the circle established by a certain truth bounding box; If the point cloud is within the bounding box of the ground truth, then calculate the distance between the point cloud and the center of the bounding box. And calculate a score based on the distance. ,in, If the point cloud is not inside the circle defined by the truth bounding box, then For ratings that are not For each point, add an edge from the truth node to the point cloud node in the graph, with the weight of the edge being the calculated score. The matching results are evaluated to obtain several evaluation results, and the final upgrade firmware is determined based on the several evaluation results; The tuning module is used to send the final upgrade firmware to the data acquisition device, and the data acquisition device upgrades the parameters of the millimeter-wave radar of the data acquisition device according to the upgrade firmware to achieve tuning and obtain the tuning result.

5. A parameter tuning device, characterized in that, The device includes: a memory, a processor, and a computer program stored in the memory and executable on the processor, the computer program being configured to implement the steps of the parameter tuning method as described in any one of claims 1 to 3.

6. A storage medium, characterized in that, The storage medium is a computer-readable storage medium, and a computer program is stored on the storage medium. When the computer program is executed by a processor, it implements the steps of the parameter tuning method as described in any one of claims 1 to 3.

7. A computer program product, characterized in that, The computer program product includes a computer program that, when executed by a processor, implements the steps of the parameter tuning method as described in any one of claims 1 to 3.

Images