Information processing method for multiple recognition devices, control device, and storage medium

Abstract

The present application relates to a multi-identification device information processing method, a control device and a storage medium. The method comprises: obtaining first target data collected by each device, wherein the first target data at least includes data of a target detected by a corresponding device; obtaining a target detection result based on the first target data; obtaining a detection trust degree based on the first target data and the detection characteristics of each device; performing trust degree fusion based on the target detection result and the trust degree of each device to obtain a confidence degree; and selectively obtaining second target data fused based on the first target data based on the confidence degree. In the technical solution of the present application, the recognition result conflict problem caused by the inherent working principle and characteristics of different sensors is effectively solved. Through this multi-sensor trust degree fusion strategy, the unmanned device can more accurately judge and identify the target, greatly improving the accuracy and robustness of target identification.

Description

Technical Field

[0001] This invention relates to the field of automatic identification, and specifically provides an information processing method, control device, and storage medium for multiple identification devices. Background Technology

[0002] With the widespread use of unmanned equipment, the requirements for recognition accuracy are becoming increasingly stringent.

[0003] In current unmanned equipment identification processes, different types of sensors, due to their inherent working principles and characteristics, are likely to provide different detection results for the same target. This can lead to conflicting results from different sensors in applications such as unmanned equipment, preventing the device from making the most appropriate judgment.

[0004] Accordingly, a new identification scheme is needed in this field to solve the above problems. Summary of the Invention

[0005] In order to overcome the above-mentioned defects, the present invention is proposed to provide a solution or at least a partial solution to the technical problem in the prior art where multi-sensor results may conflict with each other, making it impossible for the device to make a judgment.

[0006] In a first aspect, the present invention provides an information processing method for multiple identification devices, wherein the method includes: acquiring first target data collected by each device, wherein the first target data includes at least one target data detected by a corresponding device; obtaining a target detection result based on the first target data; obtaining a detection confidence level based on the first target data and the detection characteristics of each device; performing confidence level fusion based on the target detection results and confidence levels of each device to obtain a confidence level; and selectively obtaining second target data fused based on the confidence level and the first target data.

[0007] As an alternative or supplement to the above solutions, in a method according to an embodiment of the present invention, the identification device includes a millimeter-wave radar, a lidar, and a camera device. The first target data includes: third target data collected by the millimeter-wave radar, fourth target data collected by the lidar, and fifth target data collected by the camera device. The step of "obtaining target detection results based on the first target data" includes: obtaining a first target matching result for the targets identified by the two devices based on the third target data and the fourth target data; obtaining a second target matching result for the targets identified by the two devices based on the third target data and the fifth target data; obtaining a third target matching result for the targets identified by the two devices based on the fourth target data and the fifth target data; and obtaining a target detection result for the target detected by each device based on the first target matching result, the second target matching result, and the third target matching result.

[0008] As an alternative or supplement to the above solutions, in a method according to an embodiment of the present invention, wherein the third target data and the fourth target data are three-dimensional target data, and the fifth target data is two-dimensional target data, the method further includes: performing coordinate transformation on the third target data and the fourth target data to obtain corresponding third two-dimensional target data and fourth two-dimensional target data; "obtaining a second target matching result for the targets identified by the two corresponding devices based on the third target data and the fifth target data" includes: obtaining a second target matching result for the targets identified by the two corresponding devices based on the third two-dimensional target data and the fifth target data; "obtaining a third target matching result for the targets identified by the two corresponding devices based on the fourth target data and the fifth target data" includes: obtaining a third target matching result for the targets identified by the two corresponding devices based on the fourth two-dimensional target data and the fifth target data.

[0009] As an alternative or supplement to the above solutions, in a method according to an embodiment of the present invention, the trust level includes a first trust level corresponding to the millimeter-wave radar setting and a second trust level corresponding to the lidar setting. "Obtaining the detection trust level based on the first target data and the detection characteristics of each device" includes the following steps: obtaining a first target distance and a first target size based on third target data; obtaining a first target distance weight and a first target size weight based on the first target distance and the first target size; obtaining a first target distance threshold and a first target size threshold based on the first target size; obtaining the first trust level of the millimeter-wave radar for the target based on the first target distance, the first target size, the first target distance weight, the first target size weight, the first target distance threshold, and the first target size threshold; obtaining a second target distance and a second target size based on fourth target data; obtaining a second target distance weight and a second target size weight based on the second target distance and the second target size; obtaining a second target size threshold and a target distance attenuation factor obtained based on the second target distance; obtaining the second trust level of the lidar for the target based on the second target distance, the second target size, the second target distance weight, the second target size weight, the second target size threshold, and the target distance attenuation factor.

[0010] As an alternative or supplement to the above solution, in a method according to an embodiment of the present invention, "obtaining the trust level of the corresponding device based on the first target data and the detection characteristics of each device" further includes: selectively assigning a trust level to the device that did not detect the target based on the target detection result.

[0011] As an alternative or supplement to the above solutions, in a method according to an embodiment of the present invention, "selectively assigning a value to the trust level of a device that has not detected a target based on the target detection result" includes: filtering out targets whose detection results for the same target are inconsistent among devices based on the target detection result; for targets with inconsistent detection results, obtaining a first device that has not detected the target; and using the minimum trust level preset for this device as the trust level of the first device for the target.

[0012] As an alternative or supplement to the above solutions, in a method according to an embodiment of the present invention, "fusing trust scores based on the target detection results and trust scores of each device to obtain a confidence score" includes: fusing the trust scores of each device based on the Dempster combination rule to obtain a confidence score.

[0013] As an alternative or supplement to the above solutions, in a method according to an embodiment of the present invention, "selectively obtaining second target data based on the fusion of first target data based on confidence level" includes: if the confidence level is higher than a preset threshold, then obtaining second target data from third target data and fourth target data.

[0014] In a second aspect, a control device is provided, comprising a processor and a storage device, the storage device being adapted to store a plurality of computer programs, the computer programs being adapted to be loaded and run by the processor to perform the information processing method for a multi-identification device as described in any of the above-described technical solutions.

[0015] In a third aspect, a computer-readable storage medium is provided, wherein a plurality of computer programs are stored therein, the computer programs being adapted to be loaded and run by a processor to perform the information processing method for a multi-identification device as described in any of the above-described technical solutions.

[0016] The present invention comprises one or more of the following technical solutions:

[0017] Beneficial effects:

[0018] In implementing the technical solution of this invention, by collecting and integrating first target data from multiple devices, calculating the confidence level based on the detection characteristics of the devices, and further fusing the confidence levels of each device to obtain the final confidence level, this solution effectively solves the problem of conflicting recognition results caused by the inherent working principles and characteristics of different sensors. Through this multi-sensor confidence fusion strategy, unmanned equipment can more accurately judge and identify targets, significantly improving the accuracy and robustness of target recognition. Furthermore, this technology also helps reduce the decision-making risks caused by misjudgment or missed detection by a single sensor, thereby ensuring the safe and efficient operation of unmanned equipment in complex environments. Attached Figure Description

[0019] The disclosure of this invention will become more readily understood with reference to the accompanying drawings. It will be readily understood by those skilled in the art that these drawings are for illustrative purposes only and are not intended to limit the scope of protection of this invention. Furthermore, similar numbers in the drawings are used to denote similar components, wherein:

[0020] Figure 1 This is a schematic flowchart of the main steps of an information processing method for a multi-identification device according to an embodiment of the present invention;

[0021] Figure 2 This is a flowchart illustrating the minor steps of an information processing method for a multi-identification device according to an embodiment of the present invention.

[0022] Figure 3 This is a flowchart illustrating the secondary steps of an information processing method for a multi-identification device according to an embodiment of the present invention. Detailed Implementation

[0023] Some embodiments of the present invention will now be described with reference to the accompanying drawings. Those skilled in the art should understand that these embodiments are merely illustrative of the technical principles of the present invention and are not intended to limit the scope of protection of the present invention.

[0024] In the description of this invention, "module" and "processor" can include hardware, software, or a combination of both. A module can include hardware circuitry, various suitable sensors, communication ports, memory, and may also include software components, such as computer programs, or a combination of software and hardware. A processor can be a central processing unit, microprocessor, image processor, digital signal processor, or any other suitable processor. The processor has data and / or signal processing capabilities. The processor can be implemented in software, in hardware, or a combination of both. Non-transitory computer-readable storage media includes any suitable medium capable of storing computer programs, such as magnetic disks, hard disks, optical disks, flash memory, read-only memory, random access memory, etc. The term "A and / or B" means all possible combinations of A and B, such as only A, only B, or A and B. The terms "at least one A or B" or "at least one of A and B" have a similar meaning to "A and / or B" and can include only A, only B, or A and B. The singular terms "a" or "this" can also include plural forms.

[0025] Here we will first explain some of the terms involved in this invention.

[0026] Dempster-Shafer (DS) evidence theory is a mathematical theory for dealing with uncertain and incomplete information. It provides a framework that allows for uncertainty and conflict between data sources and offers tools and methods for merging information from different sources.

[0027] Dempster's Combination Rule: Dempster's combination rule is a method in Dempster's evidence theory used to combine information or evidence from two or more sources. It arrives at a comprehensive level of trust by considering conflicts between two pieces of evidence and weighting their trust levels. This rule is particularly suitable for situations where there are inconsistencies or conflicts between sources.

[0028] Example 1:

[0029] See appendix Figure 1 , Figure 1 This is a schematic flowchart illustrating the main steps of an information processing method for a multi-identification device according to an embodiment of the present invention. Figure 1 As shown, the information processing method for multiple identification devices in this embodiment of the invention mainly includes the following steps S10-S50.

[0030] Step S10: Obtain the first target data collected by each device.

[0031] In this embodiment, the first target data includes at least one target data detected by a corresponding device.

[0032] In one implementation, acquiring the first target data collected by each device is to ensure comprehensive information is obtained from multiple data sources. In real-world detection and identification scenarios, due to the differences in the physical principles and working mechanisms of various devices, each device may have unique advantages and limitations in identifying certain types of targets.

[0033] For example, millimeter-wave radar has excellent penetration capabilities and can operate in adverse weather conditions such as fog, rain, or snow; while lidar (LiDAR) can provide high-resolution 3D information, but may be affected by certain weather conditions. Camera devices can capture detailed information such as the color, shape, and texture of a target, but may be less effective in low light or obstructed conditions. Using multiple devices in combination can fully leverage their respective strengths and compensate for their weaknesses, thereby achieving more accurate and comprehensive target detection and identification.

[0034] In a preferred embodiment, the identification system employs three main devices: millimeter-wave radar, lidar, and a camera to capture and collect target data.

[0035] In this embodiment, the first target data includes: third target data, fourth target data, and fifth target data. The data collected by the millimeter-wave radar is referred to as the third target data, while the data acquired by the lidar is designated as the fourth target data. Both types of data are in three-dimensional form and can provide some target information. In this embodiment, the target information includes the first target range d provided by the millimeter-wave radar. R (i), First target orientation a R (i), the length of the first target l R (i), the first target width w R (i) Distance d from the second target provided by the lidar L (j), second target orientation a L (j), second target length l L (j), second target width w L (j), second target height h L (j) and other information. Such three-dimensional data is crucial for more accurate target location and identification, especially in complex environments requiring depth perception.

[0036] At the same time, the camera device provides a fifth type of target data, which is two-dimensional target data. Although it does not carry depth information, it can capture detailed features of the target such as color, texture, and contour. These features, when combined with three-dimensional data, can enhance the accuracy of target recognition, especially in scenes with rich variations in color and texture.

[0037] In summary, by using the data obtained from these three devices, I am able to gain a more comprehensive and detailed understanding and analysis of the target's characteristics, thus providing a solid data foundation for subsequent analysis and judgment.

[0038] Step S20: Obtain the target detection result based on the first target data.

[0039] In this embodiment, the target detection result is a compilation and comparison of the target identification results of the devices. In this embodiment, at the same time, m targets will be detected by millimeter-wave radar, n targets will be detected by lidar, and v targets will be detected by camera devices (m, n, and v are all natural numbers). The target detection result is the matching result of the various devices for the targets.

[0040] In one implementation, the target detection result is obtained through steps S201-S204, such as... Figure 2 As shown.

[0041] Step S201: Based on the third target data and the fourth target data, obtain the first target matching result of the targets identified by the two devices.

[0042] In this embodiment, preliminary analysis and matching are performed on the data collected by millimeter-wave radar and lidar.

[0043] In one implementation, the target difference is calculated using Equation 3-1. The smaller the target difference, the greater the likelihood that the two targets belong to the same obstacle.

[0044] Equation 3-1:

[0045]

[0046] Where dif(ij) is the target difference degree, and i = 1,...,m, j = 1,...n represent the labels of the two devices for the identified target, respectively; ω d ,ω a ,ω s and μ d ,μ a ,μ s These represent the weighting coefficients and error thresholds for the target distance, target orientation, and size difference between the two devices. Here, subscript 'd' represents the target distance, subscript 'a' represents the target orientation, and subscript 's' represents the size difference.

[0047] When the difference is greater than the set threshold η, the difference degree of the judgment factor is set to 1, and finally the target difference degree matrix D is obtained. RL and similarity matrix S RL .

[0048]

[0049] To determine which targets have the highest similarity and thus may represent the same obstacle, a preferred implementation uses the Hungarian algorithm. This algorithm establishes an optimal match between tasks and resources while minimizing cost. In this implementation, it is used to establish an optimal match between the third and fourth target data.

[0050] Applying the Hungarian algorithm yields three main results: a list of successfully matched targets, a list of targets detected only by millimeter-wave radar but not matched by lidar, and a list of targets detected only by lidar but not matched by millimeter-wave radar. These three lists provide valuable input information for subsequent steps, making the entire target matching and recognition process more accurate and efficient.

[0051] Suppose that in a given time instance, millimeter-wave radar detects targets A, B, and C, while lidar detects targets X, Y, and Z. After calculating the difference and applying the Hungarian algorithm, the following matching results are obtained:

[0052] A matches X;

[0053] B matches Z;

[0054] C has no match;

[0055] Y has no match.

[0056] This means that targets A and X, as well as targets B and Z, are considered two different representations of the same object. Targets C and Y, however, do not match any other targets in this time instance, which could be due to various reasons such as detection errors, occlusion, or specific blind spots of the device.

[0057] Step S202: Based on the third target data and the fifth target data, obtain the second target matching result of the targets identified by the two devices.

[0058] In one embodiment, the three-dimensional third target data is first converted into two dimensions so that it can be compared with the data from the camera device in the same coordinate system. Preferably, in another embodiment, the third target data is obtained by perspective projection onto the camera device plane, thus obtaining its two-dimensional representation, i.e., the third two-dimensional target data.

[0059] The Generalized Intersection over Union (GIoU) ​​method was then used for comparison. GIoU is a method for calculating the ratio of the intersection to the union of two rectangles, but it also takes into account the smallest closed rectangle that contains both rectangles, providing a more comprehensive assessment of target overlap.

[0060] A diagram illustrating the Generalized Intersection over Union (GIoU) ​​shows how to compare the target region projected onto the radar image with the target region from the camera. Mathematically, GIoU is calculated using the following formula:

[0061]

[0062] Where A and B represent the target area projected by the radar into the image and the target area of ​​the camera, respectively; A∩B is their intersection; A∪B is their union; and A... c It is the area of ​​the smallest rectangle that contains both A and B.

[0063] The matching degree between the targets identified by the two devices can be determined by using the results calculated using GIoU.

[0064] Step S203: Based on the fourth target data and the fifth target data, obtain the third target matching result of the targets identified by the two devices.

[0065] In one implementation, this step is similar to step S202, where the fourth target data also needs to be transformed to generate fourth two-dimensional target data. Then, based on the fourth two-dimensional target data and the fifth target data, the third target matching result of the targets identified by the two devices is obtained. GIoU is also used for calculation, which will not be elaborated here.

[0066] Step S204: Based on the first target matching result, the second target matching result and the third target matching result, obtain the target detection result of the target detected by each device.

[0067] In this embodiment, the matching results from the preceding steps are combined to form a comprehensive target detection result for the targets detected by each device.

[0068] In one implementation, targets are categorized into three types: targets that are successfully matched by all three sensors, targets that are successfully matched by only two sensors, and targets that are detected by only a single sensor.

[0069] Step S30: Obtain the confidence level of the detection based on the first target data and the detection characteristics of each device.

[0070] In this embodiment, the confidence level includes a first confidence level corresponding to the millimeter-wave radar, a second confidence level corresponding to the lidar, and a third confidence level corresponding to the camera device. Confidence level refers to the accuracy of the identification result during the target identification process.

[0071] In one implementation, the confidence level of the detection is obtained through steps S301-S303, such as... Figure 3 As shown.

[0072] Step S301: Based on the third target data and the detection characteristics of the millimeter-wave radar, obtain the first level of confidence.

[0073] In this embodiment, the detection characteristics of the millimeter-wave radar include a first target range threshold. First target size threshold First target distance weight and the first target size weight These parameters reflect the characteristic that the detection accuracy of millimeter-wave radar decreases as the distance increases or the target becomes smaller.

[0074] In one implementation, the distance to a first target and the size of a first target are obtained based on third target data. Then, based on the first target distance and the first target size, a first target distance weight and a first target size weight are obtained. Considering the different importance of distance and size, it is generally believed that the identification of targets at closer range is more accurate. Therefore, in this implementation, the first target distance weight decreases as the distance increases. However, the size of the target may affect its detection probability; larger targets are easier to detect, so the first target size weight can be set according to the target's size.

[0075] Next, a first target distance threshold and a first target size threshold are obtained based on the first target size. The relationship between the first target distance threshold, the first target size, and the first target size threshold is as follows:

[0076]

[0077] Based on the first target distance, the first target size, the first target distance weight, the first target size weight, the first target distance threshold, and the first target size threshold, the first confidence level of the millimeter-wave radar for the target is obtained.

[0078] In this embodiment, the trust level is calculated using the following formula.

[0079]

[0080] Where S R This is the highest level of trust. R For the first target size, d R This is the distance to the first target.

[0081] Step S302: Based on the fourth target data and the detection characteristics of the lidar, obtain the second level of confidence.

[0082] In this embodiment, the detection characteristics of the lidar include a second target size threshold. Target distance attenuation factor α, second target distance weight and the second target size weight

[0083] In one implementation, the distance to the second target and the size of the second target are obtained based on the fourth target data. Then, based on the second target distance and the second target size, the weight of the second target distance and the weight of the second target size are obtained.

[0084] The system obtains a second target size threshold and a target distance attenuation factor based on the second target distance. When the distance is less than 60 meters, the weight assigned to it is larger due to the high detection accuracy of the lidar. However, when the target distance exceeds 60 meters, the weight gradually decreases according to the target distance attenuation factor. Specifically:

[0085]

[0086] Based on the second target distance, second target size, second target distance weight, second target size weight, second target size threshold, and target distance attenuation factor, the second confidence level of the lidar for the target is obtained.

[0087] In this embodiment, the calculation is performed using the following formula:

[0088]

[0089] Step S303: Based on the fifth target data and the detection characteristics of the camera device, obtain the third level of trust.

[0090] In one implementation, image processing techniques, such as edge detection, region growing, and deep learning, are used to extract key features of the target, such as contours, textures, colors, and motion characteristics.

[0091] Then, by using the camera's intrinsic and extrinsic parameters and other calibration information, combined with the target features in the image, the size and distance of the target are calculated using triangulation or monocular depth estimation techniques.

[0092] Based on the detection characteristics of the camera device, the weights for the trust score are determined. For example, under good lighting conditions, the weights of color and texture features may increase, while under low lighting conditions, motion characteristics and contours may be more critical. Combining the target features, the characteristics of the camera device, and the determined weights, a third trust score is calculated. The trust score reflects the confidence of the camera device in the current target recognition result.

[0093] In addition to the steps mentioned above, step S304 includes assigning a trust level value to the sensors that did not detect the target, as detailed below:

[0094] Step S304: Based on the target detection results, selectively assign a value to the trust level of the target for devices that did not detect the target.

[0095] In one implementation, in practical applications of target detection, sensors may miss some targets due to various reasons, such as environmental noise, occlusion, and equipment performance limitations. First, the target detection results are analyzed to determine which sensors failed to detect specific targets.

[0096] Based on the physical characteristics and measured data of each sensor, a minimum detection confidence level can be preset for each sensor. For example, lidar may be affected by weather conditions under certain circumstances, leading to missed detections; cameras may miss targets in low light or obstructed conditions. To address these issues, a minimum confidence level is set for each sensor.

[0097] In this embodiment, the assignment is performed through the following steps:

[0098] Based on the target detection results, targets whose detection results for the same target differ between devices are identified. That is, in a multi-device collaborative target detection system, the detection results for the same target may vary between devices. An initial comparison of the target detection results output by each device is performed to identify targets that are detected by some devices but not by others.

[0099] Next, for targets with inconsistent detection results, the first device that failed to detect the target is identified. From the known list of inconsistent targets, those devices that failed to detect the specific target are further identified. These devices may have experienced missed detections or false positives in target detection.

[0100] The minimum trust level preset for this device is used as the first device's trust level for the target. For the identified device, its preset minimum trust level is obtained. This preset trust level is usually derived based on the device's detection performance, real-world testing, or historical data.

[0101] In this embodiment, a specific minimum trust level is assigned. It can be seen that in this embodiment, the minimum trust level is related to both the target distance and the target size.

[0102]

[0103] A preset minimum level of trust is assigned to the target corresponding to the device. In this way, even if the device fails to detect the target, the system still holds a minimum level of trust in its detection result, ensuring that this part of the data is not completely ignored in the overall decision-making process.

[0104] In this embodiment, this strategy helps to enhance the robustness of the system, ensuring that even if a single device fails or misjudges, the overall decision-making will not be biased too much.

[0105] Step S40: Based on the target detection results and trust scores of each device, perform trust score fusion to obtain the confidence score.

[0106] In this embodiment, the DS evidence theory is used to fuse trust levels.

[0107] In one implementation, first, an identification frame Θ is determined, which contains all possible target states. In DS evidence theory, let Θ be the identification frame, where Θ is a non-empty finite set. If there exists a mapping function from Θ to Θ, satisfying the following condition:

[0108]

[0109] This is then referred to as the basic probability assignment or mass function on the recognition framework Θ. For each device, the basic probability assignment function based on its trust level is calculated. For A∈2 Θ m(A) is called the basic confidence level or mass value of proposition A, representing the degree of confidence that the evidence supports the truth of proposition A. In this embodiment, the confidence levels of each device are fused based on the Dempster combination rule to obtain the confidence level.

[0110] Let m1, m2, and m3 be the basic probability assignment functions for the three pieces of evidence under the recognition framework Θ. Dempster's combinatorial rule can be expressed as:

[0111]

[0112] in The conflict coefficient K represents the degree of inconsistency between two pieces of evidence. The smaller the value, the less conflict there is between the evidence.

[0113] Using DS evidence theory for multi-sensor confidence fusion can effectively address the problems of incomplete information and uncertainty, providing a more reliable and accurate confidence level than that of a single sensor.

[0114] For example, after the target matching process, it is shown that a certain target object is only detected by LiDAR and the vision system. At this time, the evidence theory fusion process is as follows:

[0115] The conflict coefficient K is calculated as follows:

[0116]

[0117] The confidence level is calculated as follows:

[0118]

[0119] Step S50: Selectively obtain the second target data after fusion based on the first target data, based on the confidence level.

[0120] In one implementation, if the confidence level is higher than a preset threshold, the third target data and the fourth target data are combined to obtain the second target data. Specifically, the confidence level obtained based on evidence theory is compared with the preset threshold.

[0121] If the confidence level exceeds a preset threshold, the system enters the target data fusion stage. At this point, the second target data is obtained using the third and fourth target data. This is based on the assumption that the presence of the target is greatly increased when two different sensors detect the same target. Since only millimeter-wave radar and lidar can calculate object motion information, these two sets of data are used for fusion.

[0122] The fusion formula is as follows:

[0123]

[0124] F(d,a,l,w) represents the fused target range, target azimuth, and target size information. R(d,a,l,w) and L(d,a,l,w) represent the target range, target azimuth, and target size information of the millimeter-wave radar and the lidar, respectively.

[0125] This section details the role of data fusion. Individual sensors have inherent limitations and blind spots. For example, millimeter-wave radar may be more reliable than cameras or lidar in adverse weather conditions (such as fog, rain, and snow), but it may be less accurate than lidar in identifying small or non-metallic objects. By fusing data from multiple sensors, the strengths of each sensor can be utilized while reducing or eliminating their weaknesses. Data fusion provides a more complete and accurate view of the target, thereby increasing confidence in the target's presence and attributes (such as position, velocity, and size).

[0126] Taking unmanned equipment as an example, when unmanned equipment relies on a single sensor for decision-making, misjudgment by any sensor can lead to incorrect decisions, which can be catastrophic in some cases. By fusing data from multiple sensors, the true state of the target can be determined more accurately, and decisions can be made accordingly.

[0127] It should be noted that although the steps in the above embodiments are described in a specific order, those skilled in the art will understand that in order to achieve the effects of the present invention, different steps do not necessarily have to be executed in such an order. They can be executed simultaneously (in parallel) or in other orders, and these variations are all within the scope of protection of the present invention.

[0128] Those skilled in the art will understand that all or part of the processes in the method of the above embodiment of the present invention can also be implemented by a computer program instructing related hardware. The computer program can be stored in a computer-readable storage medium, and when executed by a processor, it can implement the steps of the various method embodiments described above. The computer program includes computer programs, which can be in the form of source code, object code, executable files, or certain intermediate forms. The computer-readable storage medium can include any entity or device capable of carrying the computer program, a medium, a USB flash drive, a portable hard drive, a magnetic disk, an optical disk, a computer memory, a read-only memory, a random access memory, an electrical carrier signal, a telecommunication signal, and a software distribution medium, etc. It should be noted that the content included in the computer-readable storage medium can be appropriately added or removed according to the requirements of legislation and patent practice in the jurisdiction. For example, in some jurisdictions, according to legislation and patent practice, the computer-readable storage medium does not include electrical carrier signals and telecommunication signals.

[0129] Furthermore, the present invention also provides a control device. In one embodiment of the control device according to the present invention, the control device includes a processor and a storage device. The storage device can be configured to store a program for executing the information processing method of the multi-identification device described in the above-described method embodiments. The processor can be configured to execute the program in the storage device, which includes, but is not limited to, a program for executing the information processing method of the multi-identification device described in the above-described method embodiments. For ease of explanation, only the parts related to the embodiments of the present invention are shown; for specific technical details not disclosed, please refer to the method section of the embodiments of the present invention. This control device can be a control device device comprising various electronic devices.

[0130] Furthermore, the present invention also provides a computer-readable storage medium. In one embodiment of the computer-readable storage medium according to the present invention, the computer-readable storage medium can be configured to store a program for executing the information processing method of the multi-identification device described in the above-described method embodiments. This program can be loaded and run by a processor to implement the information processing method of the multi-identification device described above. For ease of explanation, only the parts related to the embodiments of the present invention are shown; for specific technical details not disclosed, please refer to the method section of the embodiments of the present invention. The computer-readable storage medium can be a storage device comprising various electronic devices. Optionally, in the embodiments of the present invention, the computer-readable storage medium is a non-transitory computer-readable storage medium.

[0131] Furthermore, it should be understood that since the various modules are only provided to illustrate the functional units of the device of the present invention, the physical devices corresponding to these modules may be the processor itself, or a part of the processor's software, hardware, or a combination of software and hardware. Therefore, the number of modules shown in the figures is merely illustrative.

[0132] Those skilled in the art will understand that the various modules in the device can be adaptively split or combined. Such splitting or combining of specific modules will not cause the technical solution to deviate from the principles of the present invention; therefore, the technical solutions after splitting or combining will fall within the protection scope of the present invention.

[0133] The technical solution of the present invention has been described above with reference to the preferred embodiments shown in the accompanying drawings. However, it will be readily understood by those skilled in the art that the scope of protection of the present invention is obviously not limited to these specific embodiments. Without departing from the principles of the present invention, those skilled in the art can make equivalent changes or substitutions to the relevant technical features, and the technical solutions after such changes or substitutions will all fall within the scope of protection of the present invention.

Claims

1. An information processing method for multiple identification devices, characterized in that, include: Acquire the first target data collected by each device, wherein the first target data includes at least one target data detected by a corresponding device; The target detection result is obtained based on the first target data; The confidence level of the detection is obtained based on the first target data and the detection characteristics of each device; The confidence level is obtained by fusing the target detection results and confidence levels of each device. The second target data is selectively obtained based on the confidence level after fusion with the first target data; The trust level mentioned therein includes a first trust level corresponding to the millimeter-wave radar setting and a second trust level corresponding to the lidar setting. "Obtaining the detection trust level based on the first target data and the detection characteristics of each device" includes the following steps: The distance and size of the first target are obtained based on the data of the third target. Based on the first target distance and the first target size, the first target distance weight and the first target size weight are obtained; Obtain the first target distance threshold and obtain the first target size threshold based on the first target size; Based on the first target distance, the first target size, the first target distance weight, the first target size weight, the first target distance threshold, and the first target size threshold, the first confidence level of the millimeter-wave radar for the target is obtained; The distance and size of the second target are obtained based on the data from the fourth target. Based on the second target distance and the second target size, the second target distance weight and the second target size weight are obtained; Obtain the second target size threshold and the target distance attenuation factor based on the second target distance; Based on the second target distance, second target size, second target distance weight, second target size weight, second target size threshold, and target distance attenuation factor, the second trust level of the lidar for the target is obtained.

2. The information processing method for multiple identification devices according to claim 1, characterized in that, The identification device includes millimeter-wave radar, lidar, and a camera device. The first target data includes: third target data collected by the millimeter-wave radar, fourth target data collected by the lidar, and fifth target data collected by the camera device. The step of "obtaining target detection results based on the first target data" includes: Based on the third target data and the fourth target data, the first target matching result of the targets identified by the two devices is obtained; Based on the third target data and the fifth target data, the second target matching result of the targets identified by the two devices is obtained; Based on the fourth and fifth target data, the third target matching result of the targets identified by the two devices is obtained; Based on the first target matching result, the second target matching result, and the third target matching result, the target detection result of each device is obtained.

3. The information processing method for multiple identification devices according to claim 2, characterized in that, The third and fourth target data are three-dimensional target data, and the fifth target data is two-dimensional target data. The method further includes: The coordinates of the third and fourth target data are transformed to obtain the corresponding third two-dimensional target data and fourth two-dimensional target data. "Based on the third and fifth target data, the second target matching result of the targets identified by the two devices is obtained," including: Based on the third and second-dimensional target data and the fifth target data, the second target matching result of the targets identified by the two devices is obtained; "Based on the fourth and fifth target data, the third target matching result of the targets identified by the two devices is obtained," including: Based on the fourth and fifth target data, the third target matching result of the targets identified by the two devices is obtained.

4. The information processing method for multiple identification devices according to claim 1, characterized in that, "Obtaining the trust level of the corresponding device based on the first target data and the detection characteristics of each device" also includes: Based on the target detection results, a value is selectively assigned to the trust level of the target for devices that did not detect the target.

5. The information processing method for multiple identification devices according to claim 4, characterized in that, "Selectively assigning trust values ​​to devices that did not detect targets based on target detection results" includes: Based on the target detection results, targets with inconsistent detection results for the same target among different devices are screened out; For targets with inconsistent detection results, obtain the first device that did not detect the target; The minimum trust level preset for this device is used as the first device's trust level for the target.

6. The information processing method for multiple identification devices according to claim 2, characterized in that, "Confidence level is obtained by fusing the target detection results and confidence levels of each device," including: The confidence level is obtained by fusing the trust levels of each device based on the Dempster combination rule.

7. The information processing method for a multi-identification device according to any one of claims 1-6, characterized in that, "Selective acquisition of second target data based on confidence level after fusion of first target data" includes: If the confidence level is higher than a preset threshold, then the third target data and the fourth target data are used to obtain the second target data.

8. A control device comprising a processor and a storage device, said storage device being adapted to store a plurality of computer programs, characterized in that, The computer program is adapted to be loaded and run by the processor to perform the information processing method for the multi-identification device according to any one of claims 1 to 7.

9. A computer-readable storage medium storing a plurality of computer programs, characterized in that, The computer program is adapted to be loaded and run by a processor to perform the information processing method for the multi-identification device according to any one of claims 1 to 7.

Images