Target object recognition method, device, and storage medium

Abstract

This application provides a method, apparatus, device, and storage medium for identifying a target object, relating to the field of video processing technology. The method includes: when identifying a target object, first acquiring video frame images of the target object captured by a vision device, and simultaneously detecting the position of the target object by a radar device; and based on the position of the target object, determining the target image region where the target object is located from the video frame images; based on the target image region, extracting a portion of the video frame images containing the target object; and then identifying the target object in the portion of the video frame images to obtain the identification result of the target object. Using the technical solution of this application, the accuracy of the identification result can be improved without being limited by computing power.

Description

Technical Field

[0001] This application relates to the field of video processing technology, and in particular to a method, apparatus, device, and storage medium for identifying target objects. Background Technology

[0002] In some video surveillance scenarios, visual devices, such as cameras, are typically used to capture video frame images of the monitored scene and identify objects in the captured video frame images to ensure the security of the monitored scene.

[0003] Typically, the recognition of video frame images is limited by the computing power of the device. If the computing power is insufficient, it is necessary to reduce the resolution of the video frame images to meet the computing power requirements, and then recognize the objects in the video frame images with reduced resolution.

[0004] However, using the above method results in low accuracy when identifying objects in video frame images, especially distant objects. Summary of the Invention

[0005] This application provides a method, apparatus, device, and storage medium for identifying target objects, which improves the accuracy of identification results without being limited by computing power.

[0006] This application provides a method for recognizing a target object, applied to a vision device, the method comprising: The system acquires video frame images of the target object collected by the vision device, and the position of the target object detected by the radar device at the same time. Based on the location of the target object, the target image region where the target object is located is determined from the video frame image; Based on the target image region, a portion of the video frame image including the target object is extracted from the video frame image; The target object in the aforementioned video frame images is identified to obtain the identification result of the target object.

[0007] According to the target object identification method provided in this application, determining the target image region where the target object is located from the video frame image based on the position of the target object includes: The position of the target object is transformed based on the target position transformation matrix to obtain the target position of the target object in the video frame image; wherein, the target position transformation matrix is ​​the position transformation matrix between the radar device and the vision device; The target image region is determined from the video frame image based on the location of the target object and the target location.

[0008] According to the target object identification method provided in this application, determining the target image region from the video frame image based on the location of the target object and the target location includes: Based on the location of the target object, determine the target distance between the target object and the radar device; The target image region is determined from the video frame image based on the target location and the target distance.

[0009] According to the target object identification method provided in this application, the step of determining the target image region from the video frame image based on the target location and the target distance includes: Based on the target location and the target distance, a bounding box including the target object is determined in the video frame image; The image region corresponding to the recognition box is determined as the target image region.

[0010] According to the target object identification method provided in this application, determining the target image region where the target object is located from the video frame image based on the position of the target object includes: Based on the position of the target object, determine the target magnification corresponding to the vision device; If the target magnification does not meet the number of pixels required to identify the target object, the target image region is determined from the video frame image based on the position of the target object detected by the radar device.

[0011] According to the target object identification method provided in this application, the method further includes: When the target magnification meets the number of pixels required to identify the target object, video frame recognition technology is used to determine the position of the target object in the video frame image; The target image region is determined from the video frame image based on the position of the target object in the video frame image.

[0012] This application also provides a target object recognition device, applied to a vision device, the device comprising: The acquisition unit is used to acquire video frame images of the target object collected by the vision device, and the position of the target object detected by the radar device at the same time. The determining unit is configured to determine the target image region where the target object is located from video frame images of the target object acquired by the vision device based on the position of the target object; The cropping unit is used to crop a portion of the video frame image that includes the target object from the video frame image based on the target image region; The recognition unit is used to recognize the target object in the partial video frame image and obtain the recognition result of the target object.

[0013] This application also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the target object identification method as described above.

[0014] This application also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the target object identification method as described above.

[0015] This application also provides a computer program product, including a computer program that, when executed by a processor, implements the target object identification method as described above.

[0016] The target object identification method, apparatus, device, and storage medium provided in this application can first acquire video frame images of the target object collected by a vision device, and the position of the target object monitored by a radar device at the same time, when identifying the target object; and determine the target image region where the target object is located from the video frame images based on the position of the target object; extract a portion of the video frame images containing the target object based on the target image region; and then identify the target object in the portion of the video frame images to obtain the identification result of the target object. By combining the location of the target object detected by the radar equipment at the same time, a portion of the video frame image including the target object can be accurately extracted from the video frame image acquired by the vision equipment. This portion of the video frame image is then used to identify the target object. Firstly, since the computational power required to process the extracted portion of the video frame image is lower than that required for the entire video frame image, the computational power requirements on the equipment can be reduced to a certain extent. When computational power is limited, this effectively avoids the reduction of pixels in the video frame image caused by reducing the resolution of the entire video frame image, thus improving the identification capability and distance for the target object. When computational power is sufficient, reducing the number of pixels increases the speed of target object identification. Secondly, since the portion of the video frame image already meets the image size that the video analysis equipment can process without reducing the resolution of the portion of the video frame image, compared with existing technologies that rely on video frame images with reduced resolution for identification, the accuracy of the identification results can be effectively improved. Attached Figure Description

[0017] To more clearly illustrate the technical solutions in this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0018] Figure 1 This is a flowchart illustrating a target object identification method provided in an embodiment of this application.

[0019] Figure 2 This is a schematic flowchart illustrating a method for determining a target image region from a video frame image based on the location of a target object, as provided in an embodiment of this application.

[0020] Figure 3 This is a schematic diagram of the framework of a target object identification method provided in an embodiment of this application.

[0021] Figure 4 This is a schematic diagram of the framework of another target object identification method provided in the embodiments of this application.

[0022] Figure 5 This is a schematic diagram of the structure of a target object identification device provided in an embodiment of this application.

[0023] Figure 6 This is a schematic diagram of the physical structure of an electronic device provided in an embodiment of this application. Detailed Implementation

[0024] To make the objectives, technical solutions, and advantages of this application clearer, the technical solutions of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0025] In the embodiments of this application, "at least one" refers to one or more, and "more than one" refers to two or more. "And / or" describes the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, and B existing alone, where A and B can be singular or plural. In the textual description of this application, the character " / " generally indicates that the preceding and following related objects have an "or" relationship.

[0026] The technical solutions provided in this application can be applied to scenarios such as autonomous driving, intelligent transportation, and security monitoring. Taking video surveillance as an example, with the continuous development of high-resolution video, the resolution of video frame images captured by visual devices in surveillance scenarios can often reach several million or even more than 10MP. That is, the resolution of the captured video frame images is high. If objects in high-resolution video frame images can be identified, accurate identification results can be obtained to ensure the security of the surveillance scenario.

[0027] However, considering the limited computing power of video analysis equipment, the size of the video frame images it can process is often smaller than the size of the video frame images acquired by the vision device, generally only 2MP. Obviously, its computing power is insufficient to directly identify objects in high-resolution video frame images. In this case, it is necessary to reduce the resolution of the video frame images and identify objects in the reduced-resolution video frame images.

[0028] However, using the above method results in low accuracy when identifying objects in video frame images, especially distant objects.

[0029] In order to improve the accuracy of recognition results without being limited by the computing power of the device, this application provides a method for recognizing a target object. The execution subject can be a vision device or an electronic device such as a server, or a target object recognition device in an electronic device. The target object recognition device can be implemented by software, hardware or a combination of both.

[0030] For example, in the embodiments of this application, the visual device can be a camera or the like, and can be specifically configured according to actual needs.

[0031] The method for identifying the target object provided in this application will be described in detail below through several specific embodiments. It is understood that these specific embodiments can be combined with each other, and the same or similar concepts or processes may not be described again in some embodiments.

[0032] Figure 1 This application provides a flowchart illustrating a method for identifying a target object, applied to a vision device. For example, please refer to... Figure 1 As shown, the method for identifying the target object may include: S101. Acquire video frame images of the target object collected by the vision device, and the position of the target object detected by the radar device at the same time.

[0033] For example, when radar monitors the position of a target object, it can detect the position of the target object by sending and receiving signals such as radio waves or lasers. The implementation principle can include: the radar equipment emits a beam of radio waves or laser signals; when the signal encounters the target object, it is reflected back and received by the radar equipment; the radar equipment processes the received reflected signal, and by measuring information such as the signal's flight time and phase change, it calculates relevant parameters such as the distance and speed of the target object, thereby determining the position of the target object.

[0034] It is understandable that when radar equipment detects the position of a target object within its measurement range, the accuracy of its monitoring results is not limited by the distance of the target object. That is, it can accurately detect the position of the target object whether it is a close-range target object or a distant target object.

[0035] For example, when obtaining the location of a target object detected by a radar device at the same time, since the radar device may detect more than one target object, different strategies can be adopted in this case. For example, the target objects can be polled in chronological order, or the target objects can be identified based on the priority of regional importance. The target objects can be determined from multiple objects detected by the radar device in turn, thereby obtaining the location of the target objects detected by the radar device at that time, so that the vision device can obtain the location of the target objects at the same time.

[0036] After the radar device detects the exact location of the target object, it can send the location of the target object to the vision device so that the vision device can obtain the location of the target object detected by the radar device. In this way, the vision device can combine the location of the target object detected by the radar device to identify the target object in the video frame image of the target object it has collected, that is, to execute the following S102-S103.

[0037] S102. Based on the location of the target object, determine the target image region where the target object is located from the video frame image.

[0038] The target image region is the area where the target object is located in the video frame image.

[0039] Since the accuracy of the radar equipment's monitoring results is not limited by the distance of the target object when it detects the position of the target object within its measurement range, and the position of the target object can be accurately determined, the target image area where the target object is located can be accurately determined from the video frame image acquired by the vision equipment by combining the position of the target object detected by the radar equipment. After determining the target image area where the target object is located, the following S103 is executed.

[0040] S103. Based on the target image region, extract a portion of the video frame image that includes the target object from the video frame image.

[0041] In this case, the number of pixels in a partial video frame is less than the number of pixels in the entire video frame. Therefore, the computing power required to process a partial video frame is less than the computing power required to process the entire video frame.

[0042] By combining the location of the target object detected by radar equipment, a portion of the video frame image containing the target object can be accurately extracted from the video frame image acquired by the vision equipment. When identifying the target object from this portion of the video frame image, firstly, since the computational power required to process the extracted portion is lower than that required for the entire video frame image, the computational power requirements on the equipment can be reduced to a certain extent. When computational power is limited, this effectively avoids the reduction of pixels in the video frame image caused by reducing the resolution of the entire video frame image, thus improving the identification capability and distance for the target object. When computational power is sufficient, reducing the number of pixels increases the identification speed of the target object. Secondly, since the portion of the video frame image already meets the image size that the video analysis equipment can process without reducing its resolution, subsequent identification of the target object based on the portion of the video frame image can effectively improve the accuracy of the identification results compared to existing technologies that identify targets based on video frame images with reduced resolution.

[0043] S104. Identify the target object in a portion of the video frame images and obtain the identification result of the target object.

[0044] For example, in the embodiments of this application, when identifying target objects in some video frame images, traditional image recognition methods, such as frame difference, background subtraction, or optical flow, can be used to identify target objects in some video frame images; alternatively, object recognition models based on deep learning algorithms, such as convolutional neural network (CNN) models or recurrent neural network (RNN) models, can be used to identify target objects in some video frame images. The specific settings can be configured according to actual needs.

[0045] As can be seen, in the embodiments of this application, when identifying a target object, the video frame image of the target object collected by the vision device and the position of the target object detected by the radar device at the same time can be obtained first; and based on the position of the target object, the target image region where the target object is located is determined from the video frame image; based on the target image region, a portion of the video frame image including the target object is extracted from the video frame image; and then the target object in the portion of the video frame image is identified to obtain the identification result of the target object. By combining the location of the target object detected by radar equipment, a portion of the video frame image containing the target object can be accurately extracted from the video frame image acquired by the vision equipment. This portion of the video frame image is then used to identify the target object. Firstly, since the computational power required to process the extracted video frame image is lower than that required for the entire video frame image, the computational power requirements on the equipment can be reduced to a certain extent. When computational power is limited, this effectively avoids the reduction of pixels in the video frame image due to lowering the resolution of the entire video frame image, thus improving the identification capability and distance for the target object. When computational power is sufficient, reducing the number of pixels increases the speed of target object identification. Secondly, since the portion of the video frame image already meets the image size that the video analysis equipment can process without reducing its resolution, compared with existing technologies that rely on video frame images with reduced resolution for identification, the accuracy of the identification results can be effectively improved.

[0046] Based on the above Figure 1 The illustrated embodiment, in order to facilitate understanding of how, in S102 above, the target image region where the target object is located is determined from the video frame image of the target object acquired by the vision device based on the position of the target object, will be explained below. Figure 2 The embodiments shown are described in detail below.

[0047] Figure 2 This application provides a flowchart illustrating a method for determining a target image region from a video frame image based on the location of a target object, as illustrated in the embodiments of this application. For example, please refer to [link to relevant documentation]. Figure 2 As shown, the method may include: S201. Transform the position of the target object based on the target position transformation matrix to obtain the target position of the target image in the video frame image.

[0048] The target position transformation matrix is ​​the position transformation matrix between the radar equipment and the vision equipment.

[0049] For example, when transforming the position of a target object detected by a radar device based on the target position transformation matrix, the transformation operation performed on the target position can include transformation operations such as translation, rotation, or scaling, to obtain the position of the target object in the video frame image, which can be denoted as the target position.

[0050] S202. Based on the location of the target object and the target position, determine the target image region from the video frame image.

[0051] For example, in the embodiments of this application, when determining the target image region including the target object from the video frame image based on the position of the target object and the target location, the target distance between the target object and the radar device can be determined first based on the position of the target object; and the target image region including the target object can be determined from the video frame image based on the target location and the target distance.

[0052] For example, in the embodiments of this application, when determining the target image region including the target object from the video frame image based on the target location and target distance, a recognition box including the target object can be determined in the video frame image based on the target location and target distance; and the image region corresponding to the recognition box can be determined as the target image region where the target object is located.

[0053] For example, when determining a bounding box containing a target object in a video frame image based on the target location and target distance, since the target object's location detected by the radar device is a point, after transforming this point through a target location transformation matrix, the target location of the target object in the video frame image is also a point, which can be understood as the center point of the target object in the video frame image. Then, combined with the target distance detected by the radar device, by expanding the target object's center point in the video frame image outward by the size of the target object, a bounding box containing the target object can be determined in the video frame image.

[0054] For example, when expanding the size of a target object outward from its center point in a video frame image, the type of the target object can be considered. For instance, on a lane, the target object should be a vehicle. The size of the target object in the video frame image can be estimated by combining the normal size of the target object. Based on the estimated size of the target object in the video frame image, the bounding box including the target object can be obtained by expanding outward from the center point of the target object in the video frame image.

[0055] As can be seen, in this embodiment, when determining the target image region where the target object is located from the video frame image acquired by the vision device, the position of the target object monitored by the radar device can be transformed based on the target position transformation matrix to obtain the target position of the target image in the video frame image. Based on the position of the target object and the target position, the target image region including the target object can be accurately determined from the video frame image. This allows for the accurate extraction of a portion of the video frame image including the target object from the video frame image acquired by the vision device based on the target image region where the target object is located. This extracted portion of the video frame image can then be used as the basis for target object identification. On the one hand, since the computational power required for processing the extracted portion of the video frame image is lower than that required for the entire video frame image, the computational power requirement on the device can be reduced to a certain extent. On the other hand, since the portion of the video frame image already meets the image size that the video analysis device can process without reducing the resolution of the portion of the video frame image, compared with the prior art of identification based on video frame images with reduced resolution, the accuracy of the identification result can be effectively improved.

[0056] Based on any of the above embodiments, when the vision device is a variable magnification vision device, considering that the magnification of the vision device is variable, for example, before performing the above-described S102 to determine the target image region where the target object is located from the video frame image based on the position of the target object detected by the radar device, the target magnification corresponding to the vision device can be determined based on the position of the target object detected by the radar device, and it can be determined whether the target magnification meets the number of pixels required to identify the target object; if the target magnification does not meet the number of pixels required to identify the target object, the target image region is determined from the video frame image based on the position of the target object detected by the radar device, so as to execute the target object identification method provided in this application embodiment. In this way, by determining whether the target magnification meets the number of pixels required to identify the target object, and when the target magnification does not meet the number of pixels required to identify the target object, the target image region is determined from the video frame image in a targeted manner in combination with the position of the target object detected by the radar device, thereby reducing the additional data processing amount generated by determining the target image region from the video frame image based on the position of the target object when the number of pixels required to identify the target object at the target magnification is reduced.

[0057] For example, when determining the target magnification corresponding to the vision device based on the location of the target object, the target distance between the target object and the radar device can be determined first based on the location of the target object. Then, based on the target distance and the preset mapping relationship between distance and magnification, the magnification corresponding to the target distance can be determined as the target magnification corresponding to the vision device. Here, the target magnification can be understood as the maximum magnification corresponding to the target distance.

[0058] When the target magnification meets the number of pixels required to identify the target object, video frame recognition technology can be directly used to determine the position of the target object in the video frame image. Based on the position of the target object in the video frame image, the target image region can be accurately determined from the video frame image. Based on the target image region, a portion of the video frame image including the target object can be extracted. This allows for subsequent identification of the target object from the portion of the video frame image. On the one hand, since the computing power required to process the extracted portion of the video frame image is lower than that required for the entire video frame image, the computing power requirements on the device can be reduced to a certain extent. On the other hand, since the portion of the video frame image already meets the image size that the video analysis device can process without reducing the resolution of the portion of the video frame image, the accuracy of the identification results can be effectively improved compared with the existing technology of identification based on video frame images with reduced resolution.

[0059] To facilitate understanding of the target object identification method provided in the embodiments of this application, the following will explain... Figure 3 and Figure 4 The embodiments shown are described in detail below.

[0060] Figure 3 This is a schematic diagram illustrating the framework of a target object recognition method provided in this application embodiment. When recognizing a target object, a vision device acquires a video frame image of the target object and obtains the position of the target object monitored by a radar device at the same time. Since the position coordinate system corresponding to the radar device is different from that corresponding to the vision device, the position of the target object monitored by the radar device can be transformed based on the target position transformation matrix between the radar device and the vision device to obtain the target position of the target image in the video frame image. Based on the position of the target object, the target distance between the target object and the radar device is determined. Then, based on the target position and target distance, a recognition box including the target object is determined in the video frame image, and the recognition box is mapped to... The image region is determined as the target image region. By combining this with the location of the target object detected by the radar equipment, a portion of the video frame image containing the target object can be accurately extracted from the video frame image acquired by the vision equipment. This portion of the video frame image is then used to identify the target object. On the one hand, since the computational power required to process the extracted portion of the video frame image is lower than that required for the entire video frame image, the computational power requirements on the equipment can be reduced to some extent. On the other hand, since the portion of the video frame image already meets the image size that the video analysis equipment can process without reducing the resolution of the portion of the video frame image, compared with the existing technology of identifying based on video frame images with reduced resolution, the accuracy of the identification results can be effectively improved.

[0061] Figure 4This is a schematic diagram illustrating the framework of another target object recognition method provided in this application embodiment. When the vision device is a variable magnification vision device, considering the variable magnification of the vision device, the target distance between the target object and the radar device is determined based on the position of the target object detected by the radar device. Based on the target distance and the preset mapping relationship between distance and magnification, the magnification corresponding to the target distance is determined as the target magnification corresponding to the vision device. It is then determined whether the target magnification meets the number of pixels required to recognize the target object. If the target magnification does not meet the number of pixels required to recognize the target object, it indicates that the zoom operation cannot meet the recognition requirements. Therefore, the above-described method can be executed. Figure 3 The related operations shown involve transforming the position of the target object monitored by the radar device based on the target position transformation matrix between the radar and vision devices to obtain the target position in the video frame image; determining the target distance between the target object and the radar device based on the target position; and then determining a bounding box containing the target object in the video frame image based on the target position and target distance, and defining the image area corresponding to the bounding box as the target image area. By combining the position of the target object monitored by the radar device, a portion of the video frame image containing the target object can be accurately extracted from the video frame image acquired by the vision device. This partial video frame image is then used to identify the target object. On the one hand, since the computational power required to process the extracted partial video frame image is lower than that required for the entire video frame image, the computational power requirements on the device can be reduced to some extent. On the other hand, since the partial video frame image already meets the image size that the video analysis device can process without reducing the resolution of the partial video frame image, compared with the existing technology of recognition based on video frame images with reduced resolution, the accuracy of the recognition results can be effectively improved.

[0062] If the target magnification meets the number of pixels required to identify the target object, it indicates that the magnification operation satisfies the recognition requirements. Therefore, video frame recognition technology can be directly used to determine the position of the target object in the video frame image. Based on the position of the target object in the video frame image, the target image region can be accurately determined from the video frame image. Based on the target image region, a portion of the video frame image including the target object can be extracted. This allows for subsequent recognition of the target object from the portion of the video frame image. On the one hand, since the computational power required to process the extracted portion of the video frame image is lower than that required for the entire video frame image, the computational power requirements on the device can be reduced to a certain extent. On the other hand, since the portion of the video frame image already meets the image size that the video analysis device can process without reducing the resolution of the portion of the video frame image, the accuracy of the recognition results can be effectively improved compared with the existing technology of recognition based on video frame images with reduced resolution.

[0063] The target object identification device provided in this application is described below. The target object identification device described below and the target object identification method described above can be referred to in correspondence.

[0064] Figure 5 This is a schematic diagram of the structure of a target object recognition device provided in an embodiment of this application, applied to a vision device. For example, please refer to [link to relevant documentation]. Figure 5 As shown, the target object identification device 50 may include: The acquisition unit 501 is used to acquire video frame images of the target object collected by the vision device, and the position of the target object detected by the radar device at the same time. The determining unit 502 is used to determine the target image region where the target object is located from the video frame image based on the position of the target object; The cropping unit 503 is used to crop a portion of the video frame image that includes the target object from the video frame image based on the target image region; The recognition unit 504 is used to recognize the target object in the partial video frame image and obtain the recognition result of the target object.

[0065] For example, in an embodiment of this application, the cropping unit 503 is used to determine the target image region where the target object is located from a video frame image based on the position of the target object, including: The position of the target object is transformed based on the target position transformation matrix to obtain the target position of the target object in the video frame image; wherein, the target position transformation matrix is ​​the position transformation matrix between the radar device and the vision device; The target image region is determined from the video frame image based on the location of the target object and the target location.

[0066] For example, in an embodiment of this application, the cropping unit 503 is used to determine the target image region from the video frame image based on the position of the target object and the target position, including: Based on the location of the target object, determine the target distance between the target object and the radar device; The target image region is determined from the video frame image based on the target location and the target distance.

[0067] For example, in an embodiment of this application, the cropping unit 503 is used to determine the target image region from the video frame image based on the target location and the target distance, including: Based on the target location and the target distance, a bounding box including the target object is determined in the video frame image; The image region corresponding to the recognition box is determined as the target image region.

[0068] For example, in an embodiment of this application, the determining unit 502 is used to determine the target image region where the target object is located from the video frame image based on the position of the target object, including: Based on the position of the target object, determine the target magnification corresponding to the vision device; If the target magnification does not meet the number of pixels required to identify the target object, the target image region is determined from the video frame image based on the position of the target object detected by the radar device.

[0069] For example, in an embodiment of this application, the determining unit 502 is further configured to: When the target magnification meets the number of pixels required to identify the target object, video frame recognition technology is used to determine the position of the target object in the video frame image; The target image region is determined from the video frame image based on the position of the target object in the video frame image.

[0070] The target object identification device 50 provided in this application embodiment can execute the technical solution of the target object identification method in any of the above embodiments. Its implementation principle and beneficial effects are similar to those of the target object identification method. Please refer to the implementation principle and beneficial effects of the target object identification method. It will not be repeated here.

[0071] Figure 6 This is a schematic diagram of the physical structure of an electronic device provided in an embodiment of this application, such as... Figure 6As shown, the electronic device may include a processor 610, a communications interface 620, a memory 630, and a communication bus 640, wherein the processor 610, communications interface 620, and memory 630 communicate with each other via the communication bus 640. The processor 610 can call logical instructions in the memory 630 to execute the aforementioned target object recognition method. This method includes: acquiring video frame images of the target object collected by the vision device, and the position of the target object detected by the radar device at the same time; determining the target image region where the target object is located from the video frame images based on the position of the target object; extracting a portion of the video frame images including the target object from the video frame images based on the target image region; and recognizing the target object in the portion of the video frame images to obtain the recognition result of the target object.

[0072] Furthermore, the logical instructions in the aforementioned memory 630 can be implemented as software functional units and, when sold or used as independent products, can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0073] On the other hand, this application also provides a computer program product, which includes a computer program that can be stored on a computer-readable storage medium. When the computer program is executed by a processor, the computer is able to execute the target object recognition method provided by the above methods. The method includes: acquiring a video frame image of the target object collected by the vision device, and the position of the target object detected by the radar device at the same time; determining the target image region where the target object is located from the video frame image based on the position of the target object; extracting a portion of the video frame image including the target object from the video frame image based on the target image region; and recognizing the target object in the portion of the video frame image to obtain the recognition result of the target object.

[0074] In another aspect, this application also provides a computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements the target object recognition method provided by the methods described above. The method includes: acquiring video frame images of a target object collected by the vision device, and the position of the target object detected by a radar device at the same time; determining a target image region where the target object is located from the video frame images based on the position of the target object; extracting a portion of the video frame images including the target object from the video frame images based on the target image region; and recognizing the target object in the portion of the video frame images to obtain a recognition result of the target object.

[0075] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without any creative effort.

[0076] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0077] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application.

Claims

1. A method for identifying a target object, characterized in that, Applied to vision devices, the method includes: The system acquires video frame images of the target object collected by the vision device, and the position of the target object detected by the radar device at the same time. Based on the location of the target object, the target image region where the target object is located is determined from the video frame image; Based on the target image region, a portion of the video frame image including the target object is extracted from the video frame image; The target object in the aforementioned video frame images is identified to obtain the identification result of the target object.

2. The target object identification method according to claim 1, characterized in that, Determining the target image region where the target object is located from the video frame image based on the position of the target object includes: The position of the target object is transformed based on the target position transformation matrix to obtain the target position of the target object in the video frame image; wherein, the target position transformation matrix is ​​the position transformation matrix between the radar device and the vision device; The target image region is determined from the video frame image based on the location of the target object and the target location.

3. The target object identification method according to claim 2, characterized in that, Determining the target image region from the video frame image based on the location of the target object and the target location includes: Based on the location of the target object, determine the target distance between the target object and the radar device; The target image region is determined from the video frame image based on the target location and the target distance.

4. The target object identification method according to claim 3, characterized in that, Determining the target image region from the video frame image based on the target location and the target distance includes: Based on the target location and the target distance, a bounding box including the target object is determined in the video frame image; The image region corresponding to the recognition box is determined as the target image region.

5. The method for identifying a target object according to any one of claims 1-4, characterized in that, Determining the target image region where the target object is located from the video frame image based on the position of the target object includes: Based on the position of the target object, determine the target magnification corresponding to the vision device; If the target magnification does not meet the number of pixels required to identify the target object, the target image region is determined from the video frame image based on the position of the target object detected by the radar device.

6. The target object identification method according to claim 5, characterized in that, The method further includes: When the target magnification meets the number of pixels required to identify the target object, video frame recognition technology is used to determine the position of the target object in the video frame image; The target image region is determined from the video frame image based on the position of the target object in the video frame image.

7. A target object identification device, characterized in that, Applied to vision devices, the device includes: The acquisition unit is used to acquire video frame images of the target object collected by the vision device, and the position of the target object detected by the radar device at the same time. The determining unit is configured to determine the target image region where the target object is located from the video frame image based on the position of the target object; The cropping unit is used to crop a portion of the video frame image that includes the target object from the video frame image based on the target image region; The recognition unit is used to recognize the target object in the partial video frame image and obtain the recognition result of the target object.

8. An electronic device comprising a memory, a processor, and a computer program stored in the memory and running on the processor, characterized in that, When the processor executes the computer program, it implements the target object identification method as described in any one of claims 1 to 6.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the target object identification method as described in any one of claims 1 to 6.

10. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by the processor, it implements the target object identification method as described in any one of claims 1 to 6.

Images