Target identification method and device, electronic equipment and storage medium

Abstract

The embodiment of the invention discloses a target recognition method and device, electronic equipment and a storage medium, and relates to the technical field of target recognition, and the method comprises the steps: carrying out the region division of a target space, and determining a plurality of regions in the target space; obtaining at least one point cloud in the target space, and mapping the point clouds to a plurality of areas in the target space based on the first position information of the point clouds to obtain state information of the areas; the state information is used for indicating whether an interference source exists in the region or not; and identifying a target in the target space based on the state information of each region. According to the invention, the problem of low accuracy of target identification based on the millimeter wave radar in the prior art is solved.

Description

Technical Field

[0001] The present invention relates to the technical field of target recognition, and in particular, to a target recognition method, apparatus, electronic device, and storage medium. Background Art

[0002] Millimeter-wave radar has the advantages of no privacy infringement and no contact, and can be used for applications such as target presence, tracking, and recognition in the indoor home space environment of smart homes. However, existing millimeter-wave radars are prone to interference during target recognition, resulting in phenomena such as missed detection, loss of tracking, or false detection in the presence, tracking, and recognition of targets, and thus the accuracy of target recognition based on millimeter-wave radar is not high enough.

[0003] Therefore, there is an urgent need for a method capable of identifying interference sources to improve the accuracy of target recognition. Summary of the Invention

[0004] Embodiments of the present invention provide a target recognition method, apparatus, electronic device, and storage medium to solve the problem of low accuracy of target recognition based on millimeter-wave radar in related technologies. The technical solutions are as follows:

[0005] According to one aspect of the present invention, a target recognition method includes: dividing a target space into multiple regions, and determining the multiple regions in the target space; obtaining at least one point cloud in the target space, and based on the first position information of each point cloud, mapping each point cloud to the multiple regions in the target space respectively to obtain the state information of each region; the state information is used to indicate whether there is an interference source in the region; and recognizing a target in the target space based on the state information of each region.

[0006] In one embodiment, obtaining at least one point cloud in the target space, and based on the first position information of each point cloud, mapping each point cloud to the multiple regions in the target space respectively to obtain the state information of each region is implemented through the following steps: within a first time period, obtaining multiple frames of point clouds in the target space; the first time period refers to the time period for short-term learning whether there is an interference source in the multiple regions in the target space; based on the first position information of each frame of point cloud, calculating the score of each frame of point cloud in the corresponding region in the target space; the score represents the probability that there is an interference source in the region; and comparing the scores of each region with a score threshold to obtain the state information of each region corresponding to the current first time period.

[0007] In one embodiment, calculating the score of the corresponding region of each frame of the point cloud in the target space based on the first position information of each frame of the point cloud is achieved through the following steps: respectively converting the first position information of each frame of the point cloud in the target space into the second position information of each frame of the point cloud at the corresponding region; calculating the score of each region based on the second position information of each frame of the point cloud at the corresponding region.

[0008] In one embodiment, calculating the score of each region based on the second position information of each frame of the point cloud at the corresponding region is achieved through the following steps: searching for a first region in the target space based on the second position information of each point cloud at the corresponding region, and increasing the score of the first region by a first set score; the first region includes the region mapped by the point cloud; in the target space, searching for a second region located between the target detection device and the first region, and reducing the score of the second region by a second set score; determining the score of each region in the target space based on the score of the first region and the score of the second region.

[0009] In one embodiment, before identifying the target in the target space based on the status information of each region, the following steps are further included: within a second time period, obtaining multiple status information of each region; the second time period refers to the time period for long-term learning whether there are interference sources in multiple regions, and the second time period includes multiple first time periods; each status information corresponds to a first time period respectively; updating the status information of each region by filtering the obtained multiple status information.

[0010] In one embodiment, updating the status information of each region by filtering the obtained multiple status information is achieved through the following steps: filtering the multiple status information of each region within the second time period to obtain the processing result of each region; if the processing result of the region indicates that the multiple status information of the region within the second time period is inconsistent, modifying the status information of the region; if the processing result of the region indicates that the multiple status information of the region within the second time period is consistent, keeping the status information of the region unchanged.

[0011] In one embodiment, the recognition of the target in the target space based on the status information of each of the regions is achieved through the following steps: obtaining at least one first target point cloud in the target space; the first target point cloud is detected for the target in the target space; based on the status information of each of the regions, performing interference source filtering processing on each of the first target point clouds in the target space to obtain at least one second target point cloud in the target space; and using each of the second target point clouds in the target space to perform an upper-layer recognition task, where the upper-layer recognition task includes at least any one of target tracking, attitude recognition, and gesture recognition.

[0012] In one embodiment, the interference source filtering processing of each of the first target point clouds in the target space based on the status information of each of the regions to obtain at least one second target point cloud in the target space is achieved through the following steps: mapping each of the first target point clouds in the target space to a plurality of regions in the target space respectively; determining whether there is an interference source in the region mapped to based on the status information of the region mapped to by the first target point cloud; if there is an interference source in the region mapped to, then removing the first target point cloud mapped to the region, and using the first target point cloud after the removal as the second target point cloud.

[0013] In one embodiment, the method further includes the following steps: obtaining the point cloud information of each of the point clouds in the target space; calculating the distance, horizontal angle, and / or pitch angle of each of the point clouds according to the point cloud information; and calculating the first position information of each of the point clouds according to the distance, horizontal angle, and / or pitch angle of each of the point clouds.

[0014] According to one aspect of the present invention, a target recognition device includes: a region division module for dividing the target space into regions to determine a plurality of regions in the target space; a region mapping module for obtaining at least one point cloud in the target space and mapping each of the point clouds to a plurality of regions in the target space respectively based on the first position information of each point cloud to obtain the status information of each of the regions; the status information is used to indicate whether there is an interference source in the region; and a target recognition module for recognizing the target in the target space based on the status information of each of the regions.

[0015] According to one aspect of the present invention, an electronic device includes at least one processor and at least one memory, where a computer-readable instruction is stored on the memory; the computer-readable instruction is executed by one or more of the processors, so that the electronic device implements the target recognition method as described above.

[0016] According to one aspect of the present invention, a storage medium stores computer-readable instructions, which are executed by one or more processors to implement the target recognition method as described above.

[0017] The beneficial effects brought by the technical solution provided by the present invention are as follows:

[0018] In the above technical solution, first, the target space is divided into regions to determine multiple regions in the target space, so that the space can be further subdivided into subspaces that are convenient for determining interference sources. Then, at least one point cloud in the target space is obtained, and based on the first position information of each point cloud, each point cloud is mapped to multiple regions in the target space respectively to obtain the state information of each region. The state information is used to indicate whether there is an interference source in the region. Then, based on the point cloud information, the accurate situation of whether there is an interference source in each subdivided subspace can be obtained, realizing both the avoidance of false target detection caused by strong reflecting surface objects and the prevention of missed detection of stationary targets, thereby effectively solving the problem of low accuracy of target recognition based on millimeter-wave radar in the related art. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following will briefly introduce the drawings required for the description of the embodiments of the present invention. Obviously, the following described drawings are only some embodiments of the present invention, and those skilled in the art can obtain other drawings based on these drawings without creative efforts.

[0020] Figure 1 is a schematic diagram of the implementation environment related to the present invention;

[0021] Figure 2 is a flowchart of a target recognition method shown according to an exemplary embodiment;

[0022] Figure 3 is Figure 2 a schematic diagram corresponding to the region division in step 210 in the corresponding embodiment;

[0023] Figure 4 is a flowchart of the process of obtaining the state information of each region shown according to an exemplary embodiment;

[0024] Figure 5 is Figure 2 a flowchart corresponding to updating the state information of the region in the corresponding embodiment;

[0025] Figure 6 is a schematic flowchart of a target recognition method in an application scenario;

[0026] Figure 7It is a block diagram of an object recognition device shown according to an exemplary embodiment;

[0027] Figure 8 It is a hardware structure diagram of an electronic device shown according to an exemplary embodiment;

[0028] Figure 9 It is a block diagram of an electronic device shown according to an exemplary embodiment. Detailed implementation manners

[0029] The embodiments of the present invention will be described in detail below. The examples of the embodiments are shown in the drawings, where the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below by referring to the drawings are exemplary and are only used to explain the present invention and cannot be construed as a limitation to the present invention.

[0030] Those skilled in the art of the present technology can understand that, unless specifically stated otherwise, the singular forms "a", "an", "the" and "said" used herein may also include the plural forms. It should be further understood that the term "including" used in the specification of the present disclosure means the presence of the described features, integers, steps, operations, elements and / or components, but does not exclude the presence or addition of one or more other features, integers, steps, operations, elements, components and / or their groups. It should be understood that when we say an element is "connected" or "coupled" to another element, it can be directly connected or coupled to other elements, or there may also be intermediate elements. In addition, the "connection" or "coupling" used herein may include wireless connection or wireless coupling. The term "and / or" used herein includes all or any unit and all combinations of one or more related listed items.

[0031] In the process of object recognition based on a millimeter-wave radar, when the millimeter-wave radar detects a micro-moving or moving object, it will be affected by a stationary strong-reflecting surface object, resulting in low accuracy of object recognition.

[0032] For this reason, the prior art often suppresses the misdetection of objects by increasing the cfar (Constant False-Alarm Rate) threshold. However, an inappropriate cfar threshold will cause missed detection of stationary objects and still cannot ensure high accuracy of object recognition.

[0033] As can be seen from the above, the related art still has the defect of low accuracy in object recognition based on a millimeter-wave radar.

[0034] To this end, the present invention provides a target recognition method that can avoid misdetection of targets and also prevent omission of stationary targets. This target recognition method is applicable to a target recognition device, which can be a computer device configured with a von Neumann architecture. For example, the computer device includes a desktop computer, a laptop computer, a server, etc.; the electronic device can also be an electronic device with a central control function. For example, the electronic device includes a gateway, etc.; the electronic device can also refer to an intelligent device with the ability to collect and process point cloud information. For example, the intelligent device can be a millimeter-wave radar, etc. The target recognition method in the embodiments of the present invention can be applied to various scenarios, such as anti-theft monitoring, smart home, etc.

[0035] Figure 1 It is a schematic diagram of an implementation environment involved in a target recognition method. This implementation environment at least includes a user terminal 110, an intelligent device 130, a server side 170, and a network device. In Figure 1 it, the network device includes a gateway 150 and a router 190, but this is not a specific limitation here.

[0036] Among them, the user terminal 110, which can also be regarded as the user side or the terminal, can deploy (also understood as install) the client associated with the intelligent device 130. This user terminal 110 can be an electronic device such as a smart phone, a tablet computer, a laptop computer, a desktop computer, a smart control panel, and other devices with display and control functions, and is not limited here.

[0037] Among them, the client, which is associated with the intelligent device 130, actually means that the user registers an account in the client and configures the intelligent device 130 in the client. For example, this configuration includes adding a device identifier to the intelligent device 130, etc., so that when the client runs in the user terminal 110, it can provide functions such as device display and device control for the user. This client can be in the form of an application program or a web page. Correspondingly, the interface for the client to display the device can be in the form of a program window or a web page, and this is not limited here either.

[0038] The intelligent device 130 is deployed in the gateway 150 and communicates with the gateway 150 through its own configured communication module, and thus is controlled by the gateway 150. It should be understood that the intelligent device 130 generally refers to one of multiple intelligent devices 130. Only the intelligent device 130 is taken as an example in the embodiments of the present invention. That is, the number and device type of the intelligent devices deployed in the gateway 150 are not limited in the embodiments of the present invention. In an application scenario, the intelligent device 130 accesses the gateway 150 through a local area network, and thus is deployed in the gateway 150. The process by which the intelligent device 130 accesses the gateway 150 through a local area network includes: the gateway 150 first establishes a local area network, and the intelligent device 130 joins the local area network established by the gateway 150 by connecting to the gateway 150. This local area network includes but is not limited to: ZIGBEE or Bluetooth. Among them, the intelligent device 130 may be a millimeter wave radar or the like.

[0039] The interaction between the user terminal 110 and the intelligent device 130 can be realized through a local area network or a wide area network. In an application scenario, the user terminal 110 establishes a wired or wireless communication connection with the gateway 150 through the router 190. For example, the wired or wireless method includes but is not limited to WIFI, etc., so that the user terminal 110 and the gateway 150 are deployed in the same local area network, and thus the user terminal 110 can realize the interaction with the intelligent device 130 through the local area network path. In another application scenario, the user terminal 110 establishes a wired or wireless communication connection with the gateway 150 through the server side 170. For example, the wired or wireless method includes but is not limited to 2G, 3G, 4G, 5G, WIFI, etc., so that the user terminal 110 and the gateway 150 are deployed in the same wide area network, and thus the user terminal 110 can realize the interaction with the intelligent device 130 through the wide area network path.

[0040] Among them, the server side 170 can also be considered as the cloud, cloud platform, platform side, service side, etc. This server side 170 can be a single server, or a server cluster composed of multiple servers, or a cloud computing center composed of multiple servers, so as to better provide background services for a large number of user terminals 110. For example, the background services include target recognition services.

[0041] In an application scenario, after the intelligent device 130 acquires the point cloud information, it calls the target recognition service to divide the target space into regions, determine multiple regions in the target space, and then map each point cloud to the multiple regions in the target space based on the first position information of each point cloud obtained in the target space, obtaining the status information of each region. The status information is used to indicate whether there is an interference source target in the region. By accurately identifying the interference source in the target space, the problem of low accuracy of target recognition based on millimeter-wave radar existing in the related art is solved.

[0042] Of course, in other application scenarios, the above process of target recognition completed by the intelligent device 130 can also be implemented by the gateway 150 or the server 170. Taking the server 170 as an example, after the server 170 receives the point cloud information sent by the intelligent device 130, it can call the target recognition service to identify the interference source in the target space.

[0043] Please refer to Figure 2 , the embodiment of the present invention provides a target recognition method, which is applicable to an electronic device, and the electronic device can be Figure 1 the intelligent device 130 in the shown implementation environment, or can also be Figure 1 the gateway 150 or the server 170 in the shown implementation environment.

[0044] In the following method embodiments, for the convenience of description, it is described by taking the execution subject of each step of the method as an electronic device as an example, but this is not a specific limitation.

[0045] As Figure 2 shown, the method may include the following steps:

[0046] Step 210, divide the target space into regions and determine multiple regions in the target space.

[0047] Among them, region division is a process of converting continuous space information into discrete regions or grid data. This processing usually involves dividing geographical or spatial data into regular rectangular or polygonal regions, and then storing relevant information on each region.

[0048] Specifically, as Figure 3 shown, on the left is the three-dimensional display of the target space, and on the right is the three-dimensional display after the spatial region division of the target space on the left. It can be observed that through the spatial region division, the target space is divided into many regular rectangular regions, and each rectangular region is called a region.

[0049] In the above process, the information describing the space is the partitioned regions of the 3D space. However, the interference sources are usually local information within the regions. Therefore, it is necessary to consider describing the local information of the space. Thus, the 3D space region is partitioned, and the spatial information of the interference sources in the space is described by the state information of each region.

[0050] Specifically, the region size can be set by comprehensively considering two factors: the memory capacity and the expression of spatial information. The smaller the region size, the larger the memory capacity, and the more accurate the expression of spatial information. The region size can be set to 25 cm, but it is not limited here.

[0051] Step 230: Obtain at least one point cloud in the target space, and based on the first position information of each point cloud, map each point cloud to multiple regions in the target space to obtain the state information of each region.

[0052] Among them, the state information of each region is used to indicate whether there is an interference source in that region. For example, if there is an interference source in a certain region, the state information of that region is represented as 1; conversely, if there is no interference source in that region, the state information of that region is represented as 0. Of course, in other embodiments, the state information of the region is not limited to being represented by numbers, and can also be represented by letters, characters, etc., which is not specifically limited here.

[0053] In a possible implementation, step 230 may include: obtaining the point cloud information of each point cloud in the target space, calculating the distance, horizontal angle, and / or pitch angle of each point cloud according to the point cloud information, and calculating the first position information of each point cloud according to the distance, horizontal angle, and / or pitch angle of each point cloud.

[0054] First of all, it should be noted that in this embodiment, the first position information of each point cloud is obtained based on a millimeter-wave radar. Specifically, the millimeter-wave radar first emits millimeter-wave signals to the target space and obtains the echo signals reflected by various objects (which may be targets or interference sources) in the target space. Through analog-to-digital conversion, signal processing, etc., the point cloud information in the target space can be obtained. Then, according to the point cloud information, the distance, horizontal angle, pitch angle, etc. of each point cloud are calculated, and combined with the distance, horizontal angle, pitch angle, etc. of each point cloud, the first position information of each point cloud is calculated. It is worth mentioning that this first position information reflects the position of the point cloud in the target space and can be represented by coordinates.

[0055] Secondly, after obtaining the first position information of each point cloud in the target space, each point cloud can be partitioned into each sub-space (i.e., region) in the target space based on the position of each point cloud in the target space. That is to say, the purpose of mapping is to determine which region in the target space each point cloud is located in.

[0056] Taking millimeter-wave radar as an example, the inventors found that, assuming that point cloud A can be mapped to area B in the target space, there may be targets or interference sources in area B. However, for other areas in the target space, especially those areas B' between point cloud A and the millimeter-wave radar, there can be no interference sources. If there are interference sources in these areas B', point cloud A cannot be mapped to area B. Therefore, based on the mapping process of point clouds between the target space and each area, it is possible to determine whether there are interference sources in each area, that is, to obtain the status information of each area.

[0057] Step 250 , based on the status information of each area, target recognition in the target space is performed.

[0058] In a possible implementation, step 250 may include the following steps:

[0059] Step 251, obtaining at least one first target point cloud in the target space.

[0060] The first target point cloud is obtained by detecting the target in the target space.

[0061] Step 253 , based on the status information of each area, perform interference source filtering processing on each first target point cloud in the target space to obtain at least one second target point cloud in the target space.

[0062] In one possible implementation, step 253 may include: mapping each first target point cloud in the target space to multiple areas in the target space, respectively, and determining whether there is an interference source in the mapped area based on the status information of the area mapped to the first target point cloud; if there is an interference source in the mapped area, the first target point cloud mapped to the area is eliminated, and the first target point cloud after the elimination is used as the second target point cloud.

[0063] Step 255: Use each second target point cloud in the target space to perform an upper-level recognition task, where the upper-level recognition task includes at least one of target tracking, posture recognition, and gesture recognition.

[0064] In this embodiment, the status information of each area can be obtained from the short-term learning of interference sources. For example, if the first time period is set to 30s, and the first time period refers to the time period for short-term learning whether there are interference sources in multiple areas in the short-term learning target space, then the short-term learning result of the interference source can be the status information of each area obtained by short-term learning of the interference source from the first position information of each point cloud in each frame of point cloud information within 30s. It can also be obtained from the long-term learning of interference sources. For example, if the second time period is set to 3 hours, and the second time period refers to the time period for long-term learning whether there are interference sources in multiple areas, assuming that there are multiple short-term learnings within 3 hours, and each short-term learning corresponds to a short-term learning result of the interference source, then the long-term learning result can be the status information of each area obtained by long-term learning of the interference source from the multiple short-term learning results of the interference source.

[0065] In a possible implementation, the long-term learning process of the interference source may include: within the second time period, obtaining multiple status information of each area, where the second time period refers to the time period for long-term learning whether there are interference sources in multiple areas, and the second time period includes multiple first time periods; each status information corresponds to a first time period respectively, and by performing filtering processing on the obtained multiple status information, the status information of each area is updated.

[0066] Specifically, as Figure 4 shown, during the long-term learning process, the update process of the status information of each area may include the following steps:

[0067] Step 510, performing filtering processing on multiple status information of each area within the second time period to obtain the processing result of each area.

[0068] Step 530, if the processing result of the area indicates that the multiple status information of the area within the second time period is inconsistent, then modify the status information of the area.

[0069] Step 550, if the processing result of the area indicates that the multiple status information of the area within the second time period is consistent, then keep the status information of the area unchanged.

[0070] For example, assume that the second time period is 3 hours and the first time period is 30s. There are 5 short-term learning results of the interference source within the first time period, that is, there are 5 status information in each grid area. At this time, the AND operation can be performed on the 5 status information of each area. If the 5 status information of a certain area are all 1, then the status information of the area is 1, indicating that the long-term learning result of the area is that there is an interference source in the area; if one status information or more than 1 status information is 0, then the status information of the area is 0, indicating that the long-term learning result of the area is that there is no interference source in the area.

[0071] In a possible implementation, after successfully identifying the interference source, the filtered point cloud can be used for target tracking or other upper-layer applications, such as pose recognition, gesture recognition, etc.

[0072] Through the above process, in the embodiment of the present invention, by first dividing the target space into regions and determining multiple regions in the target space, the space can be further subdivided into subspaces that are convenient for determining the interference source. Then, based on the first position information of each point cloud obtained in the target space, each point cloud is mapped to multiple regions in the target space to obtain the state information of each region. The state information is used to indicate whether there is an interference source in the region. Then, it is possible to accurately obtain whether there is an interference source in each subdivided subspace for target recognition from the point cloud information, achieving both the avoidance of false target detection caused by strong reflecting surface objects and the prevention of missed detection of stationary targets, thereby effectively solving the problem of low accuracy of target recognition based on millimeter-wave radar in the related art.

[0073] In an exemplary embodiment, as Figure 5 shown, the process of obtaining the state information of each region may include the following steps:

[0074] Step 310, within a first time period, obtain multiple frames of point clouds in the target space.

[0075] Wherein, the first time period refers to the time period for short-term learning of whether there is an interference source in multiple regions in the target space.

[0076] Step 330, based on the first position information of each frame of point cloud, calculate the score of each frame of point cloud in the corresponding region in the target space.

[0077] Wherein, the score represents the probability that there is an interference source in the region. It should be understood that the higher the score, the greater the probability that there is an interference source in the region. Then, when the score of the region exceeds the score threshold, it is considered that there is an interference source in the region.

[0078] In a possible implementation, the first position information of each frame of point cloud in the target space is respectively converted into the second position information of each frame of point cloud at the corresponding region, and based on the second position information of each frame of point cloud at the corresponding region, the score of each region is calculated.

[0079] In a possible implementation, calculating the score of each region based on the second position information of each frame of point cloud at the corresponding region includes the following steps:

[0080] Step S1, based on the second position information of each point cloud at the corresponding region, search for a first region in the target space and increase the score of the first region by a first set score.

[0081] Wherein, the first region includes the region mapped by the point cloud.

[0082] Step S2, in the target space, search for a second region located between the target detection device and the first region, and reduce the score of the second region by a second set score.

[0083] Step S3, based on the scores of the first region and the second region, determine the scores of each region in the target space.

[0084] Among them, the second position information of each point cloud is used to indicate the position of each point cloud at the corresponding region. It can also be considered that the second position information of each point cloud reflects the position of each point cloud in the sub - subspace under the target space. Similarly to the representation method of the first position information, the second position information can also be represented by coordinates, which is not limited here.

[0085] As described above, through the mapping process between the point cloud in the target space and each region, it is possible to determine whether there is an interference source in each region. Specifically, the score calculation process of each region can include the following steps: based on the second position information of each point cloud at the corresponding region, search for the first region in the target space and increase the score of the first region by a first set score. The first region includes the region mapped by the point cloud; in the target space, search for a second region located between the target detection device and the first region, and reduce the score of the second region by a second set score. The second region includes the region in the target space located between the target detection device and the first region; based on the scores of the first region and the second region, determine the scores of each region in the target space. It is worth mentioning that the target detection device can be Figure 1 the intelligent device 130 in the shown implementation environment. For example, the intelligent device 130 can be a millimeter - wave radar.

[0086] For example, set the score range of each region to be from 0 to 100, and initialize the scores of all regions to 50. When a certain point cloud is mapped to this region, the score of this region +5, indicating that this region may be a target or an interference source; at the same time, subtract 10 from the scores of all regions on the straight line from this point cloud to the target detection device, indicating that this region cannot be an interference source, otherwise the point cloud cannot have a mapping. Supplementary note: when the region score is greater than or equal to 100, the region score is 100; when the region score is less than or equal to 0, the region score is 0.

[0087] In this way, the point cloud coordinates are mapped into the region space. By multiple mappings to confirm that this region is an interference source, so describing the probability that this region is an interference source with a probability region space can improve the accuracy of target recognition.

[0088] Step 350, compare the scores of each region with a score threshold to obtain the state information of each region corresponding to the current first time period.

[0089] For example, assume the score threshold is 70. Then, when the score of a region is greater than or equal to 70, there is an interference source in that region, and the corresponding status information is 1. When the score of the region is less than 70, there is no interference source in that region, and the corresponding status information is 0.

[0090] Among them, the score threshold can be set according to the actual requirements of the application scenario. For example, if the score is higher, the requirement for identifying interference sources is higher, and there are fewer missed detections in target recognition.

[0091] In this way, according to the required score threshold, it can better adapt to the target recognition requirements in different application scenarios. At the same time, it can also reduce the impact of different environments on target recognition, significantly improving the accuracy of target recognition.

[0092] Figure 6 It is a schematic flow diagram of a target recognition method in an application scenario.

[0093] Step S1, divide the 3D space into regions.

[0094] Among them, 0 is used to represent that there is no interference source in a certain region of the space, and 1 is used to represent that there is an interference source in a certain region of the space.

[0095] Step S2, short-term learning process.

[0096] First, create a probability region space, and use a scoring system to describe the probability that there is an interference source in this region. The score range of the region is from 0 to 100, and the initial score of all regions is 50. Learn the spatial information within 30 seconds, map the point clouds detected by each frame of radar data within 30 seconds into the region. When a certain point cloud is mapped into this region, the score of this region is increased by 5. After learning the spatial information within 30 seconds, cumulatively judge whether the region score is greater than 70. If the region score is greater than 70, mark the status of this region as 1.

[0097] Step S3, long-term learning process.

[0098] If the status of the fence is 1 for 5 consecutive times, there is an interference source in this region; otherwise, there is no interference source in this region.

[0099] Step S4, filter the point clouds detected in each frame, and filter out the point clouds mapped into this region.

[0100] The filtered point clouds are used for target tracking or other upper-layer applications, such as pose recognition, gesture recognition, etc. The spatial interference sources are learned and updated in real time according to steps S2 and S3, that is, after learning the short-term interference source information according to step S2 every 3 hours, judge the interference source together with the short-term interference source information learned in the previous 4 times according to the description in step S4.

[0101] Through the above process, in the embodiments of the present invention, the target space is first divided into regions to determine multiple regions in the target space, so as to subdivide the space to facilitate determining the position of the interference source. Then, based on the first position information of each point cloud obtained in the target space, each point cloud is mapped to multiple regions in the target space respectively to obtain the state information of each region. The state information is used to indicate whether there is an interference source in the region. Then, the accurate situation of each subdivided space can be obtained through the point cloud information. Finally, based on the state information of each region, interference source filtering processing is performed on each point cloud recognized for the target in the target space to obtain the point cloud information of the target. Through accurate target recognition, accurate target point cloud information is obtained, realizing target recognition that can avoid false detection and at the same time will not cause missed detection of stationary targets, thereby effectively solving the problem of low accuracy of target recognition based on millimeter-wave radar in the related art.

[0102] The following is an embodiment of the device of the present invention, which can be used to execute the target recognition method involved in the present invention. For details not disclosed in the embodiment of the device of the present invention, please refer to the method embodiment of the target recognition method involved in the present invention.

[0103] Please refer to Figure 7 , an object recognition device 800 is provided in the embodiments of the present invention.

[0104] The device 800 includes, but is not limited to: a region division module 810, a region mapping module 830, and an object recognition module 850.

[0105] Among them, the region division module 810 is configured to divide the target space into regions to determine multiple regions in the target space.

[0106] The region mapping module 830 is configured to obtain at least one point cloud in the target space, and based on the first position information of each point cloud, map each point cloud to multiple regions in the target space respectively to obtain the state information of each region.

[0107] The object recognition module 850 is configured to recognize the object in the target space based on the state information of each region.

[0108] In an exemplary embodiment, the region mapping module 830 is further configured to obtain multiple frames of point clouds in the target space within a first time period; the first time period refers to the time period for short-term learning whether there is an interference source in multiple regions in the target space; calculate the scores of each frame of the point clouds in the corresponding regions in the target space based on the first position information of each frame of the point clouds; the score represents the probability that the region has an interference source; compare the scores of each region with a score threshold to obtain the state information of each region corresponding to the current first time period.

[0109] In an exemplary embodiment, the area mapping module 830 is further configured to respectively convert the first position information of each frame of the point cloud in the target space into the second position information of each frame of the point cloud at the corresponding area; and calculate the scores of each of the areas based on the second position information of each frame of the point cloud at the corresponding area.

[0110] In an exemplary embodiment, the area mapping module 830 is further configured to search for a first area in the target space based on the second position information of each point cloud at the corresponding area, and increase the score of the first area by a first set score; the first area includes the area mapped by the point cloud; search for a second area located between the target detection device and the first area in the target space, and decrease the score of the second area by a second set score; and determine the scores of each of the areas in the target space based on the score of the first area and the score of the second area.

[0111] In an exemplary embodiment, the device 800 is further configured to obtain a plurality of status information of each of the areas within a second time period; the second time period refers to the time period for long-term learning whether there are interference sources in multiple areas, and the second time period includes a plurality of first time periods; each status information corresponds to a first time period respectively; and update the status information of each of the areas by performing filtering processing on the obtained plurality of status information.

[0112] In an exemplary embodiment, the device 800 is further configured to perform filtering processing on the plurality of status information of each of the areas within the second time period to obtain a processing result of each of the areas; if the processing result of the area indicates that the plurality of status information of the area within the second time period is inconsistent, modify the status information of the area; if the processing result of the area indicates that the plurality of status information of the area within the second time period is consistent, keep the status information of the area unchanged.

[0113] In an exemplary embodiment, the target recognition module 850 is further configured to obtain at least one first target point cloud in the target space; the first target point cloud is obtained by detecting the target in the target space; perform interference source filtering processing on each first target point cloud in the target space based on the status information of each of the areas to obtain at least one second target point cloud in the target space; and use each second target point cloud in the target space to perform an upper-level recognition task, where the upper-level recognition task includes at least any one of target tracking, pose recognition, and gesture recognition.

[0114] In an exemplary embodiment, the target recognition module 850 is further configured to map each of the first target point clouds in the target space to multiple regions in the target space; determine whether there is an interference source in the region mapped by the first target point cloud based on the status information of the region mapped by the first target point cloud; if there is an interference source in the mapped region, remove the first target point cloud mapped to the region, and use the first target point cloud after the removal as the second target point cloud.

[0115] In an exemplary embodiment, the apparatus 800 is further configured to obtain the point cloud information of each of the point clouds in the target space; calculate the distance, horizontal angle, and / or pitch angle of each of the point clouds according to the point cloud information; and calculate the first position information of each of the point clouds according to the distance, horizontal angle, and / or pitch angle of each of the point clouds.

[0116] It should be noted that when performing target recognition in the above embodiments, only the above-mentioned division of each functional module is used for illustration. In actual applications, the above functions can be allocated to different functional modules according to needs, that is, the internal structure of the target recognition device will be divided into different functional modules to complete all or part of the functions described above.

[0117] In addition, the target recognition device provided in the above embodiments and the embodiments of the target recognition method belong to the same concept. The specific ways in which each module performs operations have been described in detail in the method embodiments, and will not be repeated here.

[0118] Figure 8 The structural schematic of an electronic device shown according to an exemplary embodiment. This electronic device is applicable to Figure 1 the intelligent device 130 in the shown implementation environment. For example, the intelligent device 130 can be a millimeter-wave radar, and is also applicable to Figure 1 the gateway 150 or the server side 170 in the shown implementation environment.

[0119] It should be noted that this electronic device is only an example adapted to the present invention, and cannot be considered as providing any limitation to the scope of use of the present invention. This electronic device cannot be interpreted as requiring dependence on or necessarily having Figure 8 one or more components in the shown exemplary electronic device 2000.

[0120] The hardware structure of the electronic device 2000 may vary greatly due to different configurations or performances. As Figure 8 shown, the electronic device 2000 includes: a power supply 210, an interface 230, at least one memory 250, and at least one central processing unit (CPU) 270.

[0121] Specifically, the power supply 210 is used to provide operating voltage for each hardware device on the electronic device 2000.

[0122] The interface 230 includes at least one wired or wireless network interface 231 for interacting with external devices. For example, for the interaction Figure 1 between the smart device 130 and the gateway 150 in the illustrated implementation environment.

[0123] Of course, in other examples adapted to the present invention, the interface 230 may further include at least one serial-to-parallel conversion interface 233, at least one input / output interface 235, and at least one USB interface 237, etc., as Figure 8 shown, and specific limitations are not imposed herein.

[0124] The memory 250, as a carrier for resource storage, can be a read-only memory, a random access memory, a magnetic disk, an optical disk, etc. The resources stored thereon include an operating system 251, application programs 253, and data 255, etc. The storage method can be temporary storage or permanent storage.

[0125] Among them, the operating system 251 is used to manage and control each hardware device and application program 253 on the electronic device 2000 to enable the central processing unit 270 to perform operations and processing on the massive data 255 in the memory 250. It can be Windows ServerTM, Mac OS XTM, UnixTM, LinuxTM, FreeBSD TM, etc.

[0126] The application program 253 is a computer-readable instruction that completes at least one specific task based on the operating system 251. It may include at least one module ( Figure 8 (not shown), and each module can separately contain computer-readable instructions for the electronic device 2000. For example, the target recognition device can be regarded as an application program 253 deployed on the electronic device 2000.

[0127] The data 255 can be photos, pictures, etc. stored in the magnetic disk, or can also be point cloud information, etc., and is stored in the memory 250.

[0128] The central processing unit 270 may include one or more than one processors and is configured to communicate with the memory 250 through at least one communication bus to read the computer-readable instructions stored in the memory 250, and then perform operations and processing on the massive data 255 in the memory 250. For example, the target recognition method is completed in the form of reading a series of computer-readable instructions stored in the memory 250 by the central processing unit 270.

[0129] In addition, the present invention can also be implemented by a hardware circuit or a combination of a hardware circuit and software. Therefore, the implementation of the present invention is not limited to any specific hardware circuit, software, or the combination of the two.

[0130] Please refer to Figure 9 , in an embodiment of the present invention, an electronic device 4000 is provided, and the electronic device 400 may include: a millimeter-wave radar with the ability to collect and process point cloud information, a desktop computer, a laptop computer, a server, etc. with the ability to process point cloud information.

[0131] In Figure 9 , the electronic device 4000 includes at least one processor 4001 and at least one memory 4003.

[0132] Among them, the data interaction between the processor 4001 and the memory 4003 can be realized through at least one communication bus 4002. The communication bus 4002 may include a path for transmitting data between the processor 4001 and the memory 4003. The communication bus 4002 may be a PCI (Peripheral Component Interconnect) bus or an EISA (Extended Industry Standard Architecture) bus, etc. The communication bus 4002 can be divided into an address bus, a data bus, a control bus, etc. For the sake of convenience of representation, Figure 9 only a thick line is shown in

[0133] but it does not mean that there is only one bus or one type of bus. Optionally, the electronic device 4000 may further include a transceiver 4004, and the transceiver 4004 can be used for data interaction between the electronic device and other electronic devices, such as data sending and / or data receiving, etc. It should be noted that in practical applications, the transceiver 4004 is not limited to one, and the structure of the electronic device 4000 does not constitute a limitation to the embodiments of the present invention.

[0134] The processor 4001 may be a CPU (Central Processing Unit), a general-purpose processor, a DSP (Digital Signal Processor), an ASIC (Application Specific Integrated Circuit), an FPGA (Field Programmable Gate Array), or other programmable logic devices, transistor logic devices, hardware components, or any combination thereof. It can implement or execute various exemplary logic blocks, modules, and circuits described in connection with the disclosure of the present invention. The processor 4001 may also be a combination that implements computing functions, such as a combination of one or more microprocessors, a combination of a DSP and a microprocessor, etc.

[0135] The memory 4003 may be a ROM (Read Only Memory) or other type of static storage device that can store static information and instructions, a RAM (Random Access Memory), or other type of dynamic storage device that can store information and instructions. It may also be an EEPROM (Electrically Erasable Programmable Read Only Memory), a CD-ROM (Compact Disc Read Only Memory), or other optical disc storage, optical disc storage (including compact discs, laser discs, optical discs, digital versatile discs, Blu-ray discs, etc.), magnetic disk storage media, or other magnetic storage devices, or any other medium that can be used to carry or store the desired program instructions or code in the form of instruction or data structures and can be accessed by the electronic device 400, but is not limited thereto.

[0136] Computer-readable instructions are stored on the memory 4003, and the processor 4001 can read the computer-readable instructions stored in the memory 4003 through the communication bus 4002.

[0137] The computer-readable instructions are executed by one or more processors 4001 to implement the target recognition method in the above embodiments.

[0138] In addition, an embodiment of the present invention provides a storage medium on which computer-readable instructions are stored, and the computer-readable instructions are executed by one or more processors to implement the target recognition method as described above.

[0139] In an embodiment of the present invention, a computer program product is provided. The computer program product includes computer-readable instructions. The computer-readable instructions are stored in a storage medium. One or more processors of an electronic device read the computer-readable instructions from the storage medium, load and execute the computer-readable instructions, so that the electronic device implements the target recognition method as described above.

[0140] Compared with the related art, the beneficial effects of the present invention are as follows:

[0141] 1. By first dividing the target space into regions and determining multiple regions in the target space, the present invention can subdivide the space to facilitate determining the position of the interference source. Then, based on the first position information of each point cloud obtained in the target space, each point cloud is mapped to multiple regions in the target space to obtain the state information of each region. The state information is used to indicate whether there is an interference source in the region. Then, the accurate situation of each subdivided space can be obtained through the point cloud information. Finally, based on the state information of each region, interference source filtering processing is performed on each point cloud recognized for the target in the target space to obtain the point cloud information of the target. Through accurate target recognition, accurate target point cloud information is obtained, realizing target recognition that can avoid false detection and at the same time will not cause missed detection of stationary targets, thereby effectively solving the problem of low accuracy of target recognition based on millimeter-wave radar in the related art.

[0142] 2. The present invention proposes a method for self-learning of indoor space interference sources based on millimeter-wave radar, which can filter out false detections of strong reflection surface objects and will not cause missed detections of stationary humans.

[0143] 3. The present invention can improve the target recognition performance. By timely identifying and counteracting interference sources, the radar system can reduce the impact of interference on the target recognition performance, including reducing the false alarm rate, that is, false reporting of targets, and the missed detection rate, that is, missed reporting of targets, thereby improving the accuracy and reliability of target recognition.

[0144] 4. The present invention can improve the survival ability of the radar system. Enemy interference sources may try to interfere with the normal operation of the radar system, making it unable to detect targets. By identifying these interference sources, the radar system can take countermeasures to improve its survival ability and ensure continuous task execution.

[0145] 5. The present invention can reduce false alarms and unnecessary responses. Identifying interference sources helps reduce false alarm situations, thereby reducing unnecessary resource allocation and emergency responses, improving the system efficiency, and reducing unnecessary costs.

[0146] 6. The present invention can improve the awareness of the electromagnetic environment. Radar interference source identification helps improve the understanding of the electromagnetic environment, including identifying the type, frequency, intensity, and direction of interference, which can help the radar system better adapt to different electromagnetic environments.

[0147] 7. The present invention can improve target tracking. Identifying and countering interference sources helps maintain the continuity of target tracking. If the interference sources are identified and processed in a timely manner, the system can better track the target without being affected by the interference.

[0148] 8. The present invention can enhance security. In military applications, timely identification of interference sources can help protect critical facilities and military resources, improving the security of military forces.

[0149] It should be understood that although the steps in the flowchart of the accompanying drawings are shown sequentially in the direction of the arrows, these steps are not necessarily executed sequentially in the order indicated by the arrows. Unless otherwise clearly stated in this document, there is no strict order restriction for the execution of these steps, and they can be executed in other orders. Moreover, at least some of the steps in the flowchart of the accompanying drawings may include multiple sub-steps or multiple stages. These sub-steps or stages are not necessarily executed at the same time, but can be executed at different times, and their execution order is not necessarily sequential, but can be executed alternately or in turn with at least a part of other steps or sub-steps or stages of other steps.

[0150] The above are only some embodiments of the present invention. It should be noted that for those of ordinary skill in the art in this technical field, without departing from the principle of the present invention, several improvements and refinements can be made, and these improvements and refinements should also be regarded as the protection scope of the present invention.

Claims

1. A target recognition method, characterized in that, The method includes: Dividing the target space into regions to determine a plurality of regions in the target space; Obtaining at least one point cloud in the target space, and based on the first position information of each point cloud, mapping each point cloud to a corresponding one of the plurality of regions in the target space to obtain the state information of each region; the state information is used to indicate whether there is an interference source in the region; Identifying a target in the target space based on the state information of each region.

2. The method according to claim 1, characterized in that, The step of obtaining at least one point cloud in the target space, and based on the first position information of each point cloud, mapping each point cloud to a corresponding one of the plurality of regions in the target space to obtain the state information of each region includes: Obtaining multiple frames of point clouds in the target space within a first time period; the first time period refers to a time period for short-term learning whether there is an interference source in the plurality of regions in the target space; Calculating the score of each frame of point cloud in the corresponding region in the target space based on the first position information of each frame of point cloud; the score represents the probability that there is an interference source in the region; Comparing the scores of each region with a score threshold to obtain the state information of each region corresponding to the current first time period.

3. The method according to claim 2, characterized in that, The step of calculating the score of each frame of point cloud in the corresponding region in the target space based on the first position information of each frame of point cloud includes: Converting the first position information of each frame of point cloud in the target space into the second position information of each frame of point cloud at the corresponding region; Calculating the score of each region based on the second position information of each frame of point cloud at the corresponding region.

4. The method according to claim 3, characterized in that, The step of calculating the score of each region based on the second position information of each frame of point cloud at the corresponding region includes: Searching for a first region in the target space based on the second position information of each point cloud at the corresponding region, and increasing the score of the first region by a first set score; the first region includes the region mapped by the point cloud; Searching for a second region in the target space that is between the target detection device and the first region, and decreasing the score of the second region by a second set score; Determining the scores of all regions in the target space based on the score of the first region and the score of the second region.

5. The method according to claim 1, wherein Before identifying a target in the target space based on the state information of each region, the method further includes: Obtaining multiple state information of each region within a second time period; the second time period refers to a time period for long-term learning whether there is an interference source in the multiple regions, and the second time period includes multiple first time periods; each state information corresponds to a first time period respectively; Updating the state information of each region by performing filtering processing on the obtained multiple state information.

6. The method according to claim 5, wherein The step of updating the state information of each region by performing filtering processing on the obtained multiple state information includes: Performing filtering processing on the multiple state information of each region within the second time period to obtain the processing result of each region; If the processing result of the area indicates that multiple status information of the area in the second time period is inconsistent, modify the status information of the area; If the processing result of the area indicates that multiple status information of the area in the second time period is consistent, keep the status information of the area unchanged.

7. The method according to claim 1, wherein The identifying the target in the target space based on the status information of each area includes: Obtaining at least one first target point cloud in the target space; the first target point cloud is detected for the target in the target space; Based on the status information of each area, performing interference source filtering processing on each first target point cloud in the target space to obtain at least one second target point cloud in the target space; Using each second target point cloud in the target space to perform an upper-level identification task, where the upper-level identification task includes at least any one of target tracking, pose recognition, and gesture recognition.

8. The method according to claim 7, characterized in that The performing interference source filtering processing on each first target point cloud in the target space based on the status information of each area to obtain at least one second target point cloud in the target space includes: Mapping each of the first target point clouds in the target space to multiple areas in the target space respectively; Based on the status information of the area mapped by the first target point cloud, determining whether there is an interference source in the mapped area; if there is an interference source in the mapped area, removing the first target point cloud mapped to the area, and using the first target point cloud after the removal as the second target point cloud.

9. The method according to any one of claims 1 to 8, characterized in that The method further includes: Obtaining the point cloud information of each point cloud in the target space; Calculating the distance, horizontal angle, and / or pitch angle of each point cloud according to the point cloud information; Calculating the first position information of each point cloud according to the distance, horizontal angle, and / or pitch angle of each point cloud.

10. An object recognition device, characterized in that, The device includes: An area division module, configured to divide the target space into areas and determine multiple areas in the target space; An area mapping module, configured to obtain at least one point cloud in the target space and map each point cloud to multiple areas in the target space respectively based on the first position information of each point cloud to obtain the status information of each area; the status information is used to indicate whether there is an interference source in the area; A target identification module, configured to identify the target in the target space based on the status information of each area.

11. An electronic device, characterized in that, including: At least one processor and at least one memory, where The memory stores computer-readable instructions; The computer-readable instructions are executed by one or more of the processors, so that the electronic device implements the target identification method according to any one of claims 1 to 9.

12. A storage medium having computer-readable instructions stored thereon, characterized in that, The computer-readable instructions are executed by one or more processors to implement the target identification method according to any one of claims 1 to 9.

Images