Target object detection method, device, electronic device and readable storage medium

Abstract

The present disclosure provides a target object detection method that can be applied to the fields of artificial intelligence, the Internet of Vehicles, and smart cities. The method includes: determining a predicted overlapping distribution area based on the respective distribution positions of N target objects in a detected image, where N is a positive integer; using the predicted overlapping distribution area to filter out a second detection frame corresponding to the predicted overlapping distribution area from a first detection frame set, where the first detection frame set is the detection frame obtained after target object detection is performed on the detected image; using a preset area algorithm to process the distribution positions of the second detection frame and the target object to facilitate determining the object overlapping image area in the detected image; and using the first detection frame screening algorithm to process the first detection frame in the object overlapping image area to obtain a first detection result of the target object in the detected image. The present disclosure also provides a target object detection device, equipment, storage medium, and program product.

Description

Technical Field

[0001] The present disclosure relates to the fields of artificial intelligence, Internet of Vehicles, and smart cities, and specifically to a target object detection method, device, electronic device, readable storage medium, and program product. Background Art

[0002] With the rapid development of artificial intelligence (AI) technology, object detection methods based on AI are being widely used in scenarios such as automated assisted driving. For example, in these scenarios, AI-based image detection models can be used to process captured image information, enabling rapid detection of traffic signs, traffic lights, and other such objects within the image. This allows the automated assisted driving system to perform its assisted driving functions based on the detected objects.

[0003] In the process of realizing the concept of the present disclosure, the inventors found that there are at least the following problems in the related technology: the accuracy of the detection results for the target object in the image is low, especially when the content contained in the collected image information is relatively complex, it is difficult for the detection accuracy of the target object to meet actual needs. Summary of the Invention

[0004] In view of this, the present disclosure provides a target object detection method, apparatus, electronic device, readable storage medium, and program product.

[0005] One aspect of the present disclosure provides a target object detection method, comprising:

[0006] Determine the predicted overlapping distribution area according to the respective distribution positions of N target objects in the detected image, where N is a positive integer;

[0007] Using the predicted overlapping distribution area, a second detection frame corresponding to the predicted overlapping distribution area is selected from the first detection frame set, wherein the first detection frame set is a detection frame obtained after performing target object detection on the detected image;

[0008] Processing the distribution positions of the second detection frame and the target object using a preset area algorithm to determine an object overlapping image area in the detected image; and

[0009] A first detection frame screening algorithm is used to process the first detection frame in the object overlapping image area to obtain a first detection result of the target object in the detected image.

[0010] According to an embodiment of the present disclosure, determining the predicted overlapping distribution area based on respective distribution positions of N target objects in the detected image includes:

[0011] According to the respective distribution positions of the N target objects in the detected image, M first target objects are screened out from the N target objects according to a preset position screening rule;

[0012] According to the first distribution positions of the respective first target objects, cluster analysis is performed on the M first target objects to obtain the predicted overlapping distribution area.

[0013] According to an embodiment of the present disclosure, using the predicted overlapping distribution area to filter out a second detection frame corresponding to the predicted overlapping distribution area from the first detection frame set includes:

[0014] Expand the predicted overlapping distribution area according to a preset boundary distance to obtain a detection frame screening area; and

[0015] According to the first detection frame positions of the respective first detection frames in the first detection frame set, the first detection frames located in the detection frame screening area are filtered out from the first detection frame set to obtain a second detection frame corresponding to the predicted overlapping distribution area.

[0016] According to an embodiment of the present disclosure, the predicted overlapping distribution area includes a second target object, and the second target object is a target object having the same clustering attribute among the M first target objects;

[0017] Processing the distribution position of the second detection frame and the target object using a preset area algorithm to determine an overlapping image area of ​​the object in the detected image includes:

[0018] The second detection frame and the second distribution position of the second target object corresponding to the second detection frame are processed using the preset area algorithm to obtain an object overlapping image area corresponding to the predicted overlapping distribution area.

[0019] According to an embodiment of the present disclosure, the first detection frame in the object overlapping image region includes a plurality of;

[0020] Before processing the first detection frame in the object overlapping image area using the first detection frame screening algorithm, the target object detection method further includes:

[0021] classifying the plurality of first detection frames according to their respective detection frame categories to obtain one or more detection frame sets, wherein the first detection frames in the same detection frame set have the same detection frame category;

[0022] For the same detection frame set, the detection frame score of the target detection frame in the detection frame set is updated using the following method, wherein the detection frame score of the target detection frame is greater than the detection frame scores of other detection frames in the detection frame set. The detection frame score is used to represent the comprehensive detection result of the detection frame position and detection frame category of the detection frame. The above method includes:

[0023] Calculate the intersection-union ratio of the target detection frame and the other detection frames in the detection frame set;

[0024] For each of the above detection frame intersection-over-union ratios, when the above detection frame intersection-over-union ratio is greater than a preset attribute threshold, the current detection frame score of the above target detection frame is iteratively optimized using the position parameter to obtain the target detection frame score after iteration of the above target detection frame;

[0025] According to the target detection frame score after the iteration of the above target detection frame, the first detection frame score of the first detection frame corresponding to the above target detection frame in the above object overlapping image area is updated to obtain the target first detection frame score of the first detection frame corresponding to the above target detection frame.

[0026] According to an embodiment of the present disclosure, iteratively optimizing the current detection frame score of the target detection frame using position parameters includes:

[0027] Iteratively calculate the sum of the above position parameters and the current detection frame score of the above target detection frame;

[0028] The above-mentioned position parameter includes: the product of the number of detection frames in the above-mentioned detection frame set and the preset position parameter and the minimum value of the above-mentioned preset position parameters.

[0029] According to an embodiment of the present disclosure, the target object detection method further includes:

[0030] Using the object overlapping image area, screening out the object non-overlapping area in the detected image; and

[0031] The first detection frame in the non-overlapping image area of ​​the object is processed using a second detection frame screening algorithm to obtain a second detection result of the target object in the detected image.

[0032] According to an embodiment of the present disclosure, the target object detection method further includes:

[0033] A target object detection result for the detected image is determined according to the first target object detection result and the second target object detection result.

[0034] According to an embodiment of the present disclosure, the first detection frame screening algorithm includes a non-maximum suppression algorithm; and / or

[0035] The second detection frame screening algorithm includes a smoothing-non-maximum suppression algorithm.

[0036] According to an embodiment of the present disclosure, the above-mentioned preset area algorithm includes at least one of the following:

[0037] Triangulation algorithm, minimum region box algorithm.

[0038] According to an embodiment of the present disclosure, the detected image is acquired by an image acquisition device, and the N target objects are pre-arranged in an image acquisition space corresponding to the detected image;

[0039] The target object detection method further includes:

[0040] According to the position calibration relationship between the image acquisition position of the above-mentioned image acquisition device and the respective preset positions of the N above-mentioned target objects, the respective preset positions of the N above-mentioned target objects are converted to obtain the respective distribution positions of the N above-mentioned target objects in the detected image.

[0041] Another aspect of the present disclosure provides a target object detection device, comprising:

[0042] A first determination module is configured to determine a predicted overlapping distribution area based on respective distribution positions of N target objects in the detected image, where N is a positive integer;

[0043] a first screening module configured to screen out, from a first set of detection frames, second detection frames corresponding to the predicted overlapping distribution area using the predicted overlapping distribution area, wherein the first set of detection frames is detection frames obtained after performing target object detection on the detected image;

[0044] a second determining module, configured to process the distribution positions of the second detection frame and the target object using a preset area algorithm, so as to determine an object overlapping image area in the detected image; and

[0045] The first detection module is used to process the first detection frame in the above-mentioned object overlapping image area using a first detection frame screening algorithm to obtain a first detection result of the target object in the above-mentioned detected image.

[0046] Another aspect of the present disclosure provides an electronic device, comprising:

[0047] one or more processors;

[0048] a memory for storing one or more programs,

[0049] When the one or more programs are executed by the one or more processors, the one or more processors implement the target object detection method as described above.

[0050] Another aspect of the present disclosure provides a computer-readable storage medium storing computer-executable instructions, which are used to implement the target object detection method described above when executed.

[0051] Another aspect of the present disclosure provides a computer program product, which includes computer-executable instructions. When the instructions are executed, the computer program product is used to implement the target object detection method described above.

[0052] According to an embodiment of the present disclosure, through the respective distribution positions of N target objects in the detected image, a predicted overlapping distribution area in which overlapping distribution may exist in the detected image can be preliminarily predicted, and then the predicted overlapping distribution area is used to filter out a second detection frame corresponding to the predicted overlapping distribution area from the unfiltered first detection frame set, and then the preset area algorithm is used to process the distribution position of the second detection frame and the target object, so that the object overlapping image area where the target object overlapping scene exists can be effectively determined from the detected image, and then the first detection frame screening algorithm is used to process the first detection frame in the object overlapping image area, which can at least partially solve the technical problem in the related art that it is difficult to distinguishably filter the first detection frame in the target object overlapping scene area, and can improve the accuracy of the first detection frame screening algorithm for the first detection frame in the object overlapping image area, so that the first detection result of the target object obtained can achieve the technical effect of improving the accuracy of target object detection. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] The above and other objects, features and advantages of the present disclosure will become more apparent through the following description of the embodiments of the present disclosure with reference to the accompanying drawings, in which:

[0054] Figure 1 Schematically illustrates an exemplary system architecture to which the target object detection method and apparatus according to an embodiment of the present disclosure can be applied;

[0055] Figure 2 The following schematically shows a flow chart of a target object detection method according to an embodiment of the present disclosure;

[0056] Figure 3 Schematically illustrates a flow chart for determining a predicted overlapping distribution area based on respective distribution positions of N target objects in a detected image according to an embodiment of the present disclosure;

[0057] Figure 4A flowchart of filtering out a second detection frame corresponding to the predicted overlapping distribution area from a set of first detection frames using the predicted overlapping distribution area according to an embodiment of the present disclosure is schematically shown;

[0058] Figure 5 The following schematically illustrates an application scenario of a target object detection method according to an embodiment of the present disclosure;

[0059] Figure 6 A block diagram schematically illustrates a target object detection device according to an embodiment of the present disclosure; and

[0060] Figure 7 A block diagram of an electronic device suitable for implementing a target object detection method according to an embodiment of the present disclosure is schematically shown. DETAILED DESCRIPTION

[0061] Hereinafter, embodiments of the present disclosure will be described with reference to the accompanying drawings. However, it should be understood that these descriptions are merely exemplary and are not intended to limit the scope of the present disclosure. In the detailed description below, for ease of explanation, many specific details are set forth to provide a comprehensive understanding of the embodiments of the present disclosure. However, it is apparent that one or more embodiments may also be implemented without these specific details. In addition, in the following description, descriptions of well-known structures and technologies are omitted to avoid unnecessary confusion of the concepts of the present disclosure.

[0062] The terms used herein are only for describing specific embodiments and are not intended to limit the present disclosure. The terms "comprise," "include," etc. used herein indicate the presence of the features, steps, operations, and / or components, but do not exclude the presence or addition of one or more other features, steps, operations, or components.

[0063] All terms used herein (including technical and scientific terms) have the meanings commonly understood by those skilled in the art unless otherwise defined. It should be noted that the terms used herein should be interpreted as having a meaning consistent with the context of this specification and should not be interpreted in an idealized or overly rigid manner.

[0064] When expressions such as "at least one of A, B and C, etc." are used, they should generally be interpreted in accordance with the meaning of the expression commonly understood by those skilled in the art (for example, "a system having at least one of A, B and C" should include but is not limited to a system having A alone, B alone, C alone, A and B, A and C, B and C, and / or A, B, C, etc.).

[0065] In the relevant technologies of target object detection, target object detection can usually be performed on the detected image based on the target object recognition model, and the generated initial detection results generate an initial detection frame used to represent the target object. This initial detection frame can preliminarily predict the classification and distribution location of the target object, but the initial detection frame usually contains a large number of redundant detection frames. In actual application scenarios, it is not expected to detect too many detection frames for a single target, so the NMS algorithm, i.e., the Non-Maximum Suppression (NMS) algorithm, is generally used in actual applications to filter out other redundant detection frames.

[0066] However, due to its direct filtering characteristics, the NMS algorithm not only relies on the setting of the IOU (Intersection over Union) threshold for its filtering results, but also may mistakenly filter out other target objects in scenarios where target objects overlap, affecting the final target object detection results.

[0067] To address the direct filtering characteristics of the NMS algorithm, the SOFT-NMS algorithm (SOFT-Non-Maximum Suppression) was proposed in related art. The core of the SOFT-NMS algorithm is to build on the NMS algorithm and, instead of directly filtering the detection boxes using the detection box score (i.e., not clearing the SCORE), use SCORE (detection box score) decay to smooth the detection box filtering problem. This, to a certain extent, addresses the problem of abnormal filtering in scenes with overlapping objects. However, the SOFT-NMS algorithm introduces new problems, such as under-filtering for single objects. Furthermore, because the SOFT-NMS algorithm uses SCORE as a decision factor, it relies more heavily on the detection box score of the initial detection box in the preliminary detection results than the NMS algorithm. However, in object detection tasks, the detection box score of an object is generally the overall score of the detection location and the detection box classification, making it difficult to effectively distinguish between the detection box classification result and the detection location result. Therefore, how to better apply the detection box score metric when using the SOFT-NMS algorithm is a challenge for the more effective application of the SOFT-NMS algorithm.

[0068] Based on the above technical problems, the embodiments of the present disclosure provide a target object detection method, device, electronic device and readable storage medium. The target object detection method includes:

[0069] According to the respective distribution positions of N target objects in the detected image, a predicted overlapping distribution area is determined, where N is a positive integer; a second detection frame corresponding to the predicted overlapping distribution area is screened out from a first detection frame set using the predicted overlapping distribution area, where the first detection frame set is the detection frame obtained after target object detection is performed on the detected image; the distribution positions of the second detection frame and the target object are processed using a preset area algorithm to facilitate determination of the object overlapping image area in the detected image; and the first detection frame screening algorithm is used to process the first detection frame in the object overlapping image area to obtain a first detection result of the target object in the detected image.

[0070] According to an embodiment of the present disclosure, through the respective distribution positions of N target objects in the detected image, a predicted overlapping distribution area in which overlapping distribution may exist in the detected image can be preliminarily predicted, and then the predicted overlapping distribution area is used to filter out a second detection frame corresponding to the predicted overlapping distribution area from the unfiltered first detection frame set, and then the preset area algorithm is used to process the distribution position of the second detection frame and the target object, so that the object overlapping image area where the target object overlapping scene exists can be effectively determined from the detected image, and then the first detection frame screening algorithm is used to process the first detection frame in the object overlapping image area, which can at least partially solve the technical problem in the related art that it is difficult to distinguishably filter the first detection frame in the target object overlapping scene area, and can improve the accuracy of the first detection frame screening algorithm for the first detection frame in the object overlapping image area, so that the first detection result of the target object obtained can achieve the technical effect of improving the accuracy of target object detection.

[0071] It should be noted that the target object detection method provided by the embodiments of the present disclosure can be applied to multiple application scenarios such as self-assisted driving and urban smart transportation. Accordingly, the target object detection method provided by the embodiments of the present disclosure can be applied to the vehicle network field, but is not limited to this. The target object detection method provided by the embodiments of the present disclosure can also be applied to other fields, for example, it can be applied to smart security, smart urban transportation and other fields. The embodiments of the present disclosure do not limit the application fields of the target object detection method.

[0072] In the technical solution disclosed herein, the user's authorization or consent is obtained before obtaining or collecting the user's personal information.

[0073] In the technical solution disclosed herein, the collection, storage, use, processing, transmission, provision, disclosure and application of user personal information involved comply with the provisions of relevant laws and regulations, take necessary confidentiality measures, and do not violate public order and good morals.

[0074] Figure 1The following schematically illustrates an exemplary system architecture to which the target object detection method and apparatus according to an embodiment of the present disclosure can be applied. Figure 1 The examples shown are merely examples of system architectures to which the embodiments of the present disclosure may be applied, to help those skilled in the art understand the technical content of the present disclosure, but do not mean that the embodiments of the present disclosure may not be used in other devices, systems, environments or scenarios.

[0075] like Figure 1 As shown, the system architecture 100 according to this embodiment may include terminal devices 101, 102, 103, a network 104, and a server 105. The network 104 is used as a medium for providing communication links between the terminal devices 101, 102, 103 and the server 105. The network 104 may include various connection types, such as wired and / or wireless communication links, etc.

[0076] Users can use terminal devices 101, 102, and 103 to interact with server 105 via network 104 to receive or send messages, etc. Various communication client applications can be installed on terminal devices 101, 102, and 103, such as shopping applications, web browser applications, search applications, instant messaging tools, email clients, and / or social platform software (for example only).

[0077] The terminal devices 101 , 102 , and 103 may be various electronic devices having a display screen and supporting web browsing, including but not limited to smart phones, tablet computers, laptop computers, and desktop computers.

[0078] The server 105 may be a server that provides various services, such as a background management server (for example only) that supports websites browsed by users using the terminal devices 101, 102, and 103. The background management server may analyze and process received data such as user requests, and feed back processing results (such as web pages, information, or data obtained or generated according to user requests) to the terminal device.

[0079] It should be noted that the target object detection method provided in the embodiment of the present disclosure can generally be executed by the server 105. Accordingly, the target object detection device provided in the embodiment of the present disclosure can generally be set in the server 105. The target object detection method provided in the embodiment of the present disclosure can also be executed by a server or server cluster that is different from the server 105 and can communicate with the terminal devices 101, 102, 103 and / or the server 105. Accordingly, the target object detection device provided in the embodiment of the present disclosure can also be set in a server or server cluster that is different from the server 105 and can communicate with the terminal devices 101, 102, 103 and / or the server 105. Alternatively, the target object detection method provided in the embodiment of the present disclosure can also be executed by the terminal devices 101, 102, or 103, or can also be executed by other terminal devices different from the terminal devices 101, 102, or 103. Accordingly, the target object detection device provided in the embodiment of the present disclosure may also be set in the terminal device 101, 102, or 103, or in other terminal devices different from the terminal device 101, 102, or 103.

[0080] For example, the distribution positions of the detected image and the target object can be originally stored in any one of the terminal devices 101, 102, or 103 (for example, the terminal device 101, but not limited thereto), or stored on an external storage device and can be imported into the terminal device 101. Then, the terminal device 101 can locally execute the target object detection method provided by the embodiment of the present disclosure, or send the distribution positions of the detected image and the target object to other terminal devices, servers, or server clusters, and the target object detection method provided by the embodiment of the present disclosure can be executed by the other terminal devices, servers, or server clusters that receive the distribution positions of the detected image and the target object.

[0081] It should be understood that Figure 1 The number of terminal devices, networks and servers in the embodiment is merely illustrative. Any number of terminal devices, networks and servers may be provided as required.

[0082] Figure 2 The flowchart of the target object detection method according to an embodiment of the present disclosure is schematically shown.

[0083] like Figure 2 As shown, the target object detection method 200 includes operations S210 to S240.

[0084] In operation S210 , a predicted overlapping distribution area is determined according to respective distribution positions of N target objects in the detected image, where N is a positive integer.

[0085] According to an embodiment of the present disclosure, the distribution position of the target object in the detected image may include the coordinate position of the target object in the detected image, and the coordinate position may be obtained in a pre-set manner, or may be calculated based on a relevant prediction algorithm. The embodiment of the present disclosure does not limit the method for obtaining the distribution position of the target object.

[0086] According to an embodiment of the present disclosure, the predicted overlapping distribution area may include an area where target objects in the detected image overlap with respect to a near view position and a far view position.

[0087] In operation S220 , the predicted overlapping distribution area is used to filter out a second detection frame corresponding to the predicted overlapping distribution area from the first detection frame set, where the first detection frame set is the detection frame obtained after target object detection is performed on the detected image.

[0088] According to an embodiment of the present disclosure, a method for performing target object detection on a detected image may include a method for processing the detected image based on a network model constructed based on a neural network. For example, the detected image may be processed by a target detection model constructed based on an RPN (Region Proposal Network) network to obtain a first detection frame set.

[0089] It should be noted that the first detection frame in the first detection frame set is an initial detection frame that has not been screened or filtered. The second detection frame may include the first detection frame that overlaps at least partially with the predicted overlapping distribution area. The predicted overlapping distribution area is more likely to overlap with the target objects, so the second detection frame further reflects the overlap of the target objects in the detected image.

[0090] In operation S230, the distribution positions of the second detection frame and the target object are processed using a preset region algorithm to determine an object overlapping image region in the detected image.

[0091] According to an embodiment of the present disclosure, the preset region algorithm may include an algorithm for generating region information based on location information, such as a triangulation algorithm, etc. The embodiment of the present disclosure does not limit the prediction region algorithm, and those skilled in the art may select one according to actual needs.

[0092] In operation S240 , a first detection frame screening algorithm is used to process the first detection frame in the object overlapping image area to obtain a first detection result of the target object in the detected image.

[0093] According to an embodiment of the present disclosure, the first detection frame screening algorithm may include a detection frame filtering algorithm in the related technology. The first detection frame screening algorithm may filter the first detection frame in the object overlapping image area. The detection frame obtained after filtering may have a detection frame classification result and a detection frame position, thereby realizing the use of the first detection result of the target object to characterize the position and classification of the target object in the detected image.

[0094] According to an embodiment of the present disclosure, through the respective distribution positions of N target objects in the detected image, a predicted overlapping distribution area in which overlapping distribution may exist in the detected image can be preliminarily predicted, and then the predicted overlapping distribution area is used to filter out a second detection frame corresponding to the predicted overlapping distribution area from the unfiltered first detection frame set, and then the preset area algorithm is used to process the distribution position of the second detection frame and the target object, so that the object overlapping image area where the target object overlapping scene exists can be effectively determined from the detected image, and then the first detection frame screening algorithm is used to process the first detection frame in the object overlapping image area, which can at least partially solve the technical problem in the related art that it is difficult to distinguishably filter the first detection frame in the target object overlapping scene area, and can improve the accuracy of the first detection frame screening algorithm for the first detection frame in the object overlapping image area, so that the first detection result of the target object obtained can achieve the technical effect of improving the accuracy of target object detection.

[0095] According to an embodiment of the present disclosure, the detected image is acquired by an image acquisition device, and N target objects are pre-arranged in an image acquisition space corresponding to the detected image.

[0096] The target object detection method may further include the following operations:

[0097] According to the position calibration relationship between the image acquisition position of the image acquisition device and the respective preset positions of the N target objects, the respective preset positions of the N target objects are positionally transformed to obtain the respective distribution positions of the N target objects in the detected image.

[0098] According to an embodiment of the present disclosure, the target object pre-arranged in the image acquisition space corresponding to the detected image can be, for example, a traffic sign, billboard, crosswalk, etc., located in the image acquisition space. The preset position of the target object can include the coordinate position of the target object in space. The calibration relationship between the preset position and the image acquisition position can be represented by a coordinate transformation matrix in related technologies. By performing a position transformation on the preset position, the target object can be mapped to the image coordinate system of the detected image, thereby obtaining the distribution position of the target object in the spare part image.

[0099] Figure 3The flowchart of determining the predicted overlapping distribution area according to the respective distribution positions of N target objects in the detected image according to an embodiment of the present disclosure is schematically shown.

[0100] like Figure 3 As shown, operation S210, determining the predicted overlapping distribution area according to the respective distribution positions of the N target objects in the detected image, includes operations S310 to S320.

[0101] In operation S310 , M first target objects are screened out from the N target objects according to respective distribution positions of the N target objects in the detected image and a preset position screening rule.

[0102] In operation S320 , cluster analysis is performed on the M first target objects according to the first distribution positions of the first target objects to obtain predicted overlapping distribution areas.

[0103] According to an embodiment of the present disclosure, the distribution position of the target object in the detected image may be represented by the position of the target object coordinate point, or may be represented by the target object detection frame of the target object in the detected image.

[0104] According to an embodiment of the present disclosure, the M first target objects screened out according to the preset position screening rules can represent target objects with overlapping coordinates in the detected image, such as first target objects with different depth information and close image plane coordinate positions.

[0105] The calculation process of selecting M first target objects from N target objects can be expressed by formula (1).

[0106]

[0107] In formula (1), P i and P j Represents the target object, iou() represents the intersection of the detection boxes of two different target objects, distance() represents the Euclidean distance between two different target objects, depth() represents the depth information of the target object in the detected image, μ IOU Represents the preset first intersection-over-union threshold, μ Distance Indicates the preset distance threshold.

[0108] According to formula (1), the distribution positions represented by the detection box and the distribution positions represented by the target object coordinate points can be filtered respectively. By traversing each target object in the N target objects through formula (1), M first target objects can be filtered out.

[0109] It should be understood that the detection box intersection-over-union (IOU) ratio may be the ratio of the intersection area to the union area of ​​two different detection boxes.

[0110] According to embodiments of the present disclosure, M first target objects can be processed using a clustering algorithm known in related art to obtain one or more clusters. The first target objects in the same cluster can constitute a cluster subset. The cluster subset can form a corresponding predicted overlapping distribution area in the detected image.

[0111] In some embodiments of the present disclosure, the predicted overlapping distribution area may include the area where the clusters are actually distributed in the detected image, or may also include a circular area formed with a first target object at the center of the cluster as the center and a preset radius. Alternatively, it may be a circular area formed with a first target object at the center of the cluster as the center and a preset radius.

[0112] It should be noted that any clustering algorithm in the relevant technology can be used for cluster analysis, such as the k-means algorithm, DBSCAN algorithm, etc. The embodiments of the present disclosure do not limit the specific algorithm type of cluster analysis, and those skilled in the art can design it according to actual needs.

[0113] Figure 4 The flowchart of using the predicted overlapping distribution area to filter out the second detection frame corresponding to the predicted overlapping distribution area from the first detection frame set according to an embodiment of the present disclosure is schematically shown.

[0114] like Figure 4 As shown, operation S220 of filtering out a second detection frame corresponding to the predicted overlapping distribution area from the first detection frame set using the predicted overlapping distribution area may include operations S410 to S420.

[0115] In operation S410 , the predicted overlapping distribution area is expanded according to a preset boundary distance to obtain a detection frame screening area.

[0116] In operation S420 , first detection frames located in the detection frame screening area are screened out from the first detection frame set according to respective first detection frame positions of the first detection frames in the first detection frame set, to obtain second detection frames corresponding to the predicted overlapping distribution area.

[0117] According to the embodiments of the present disclosure, the preset boundary distance can be designed based on actual needs. By expanding the predicted overlapping distribution area according to the preset boundary distance, the screening range of the first detection frame that may overlap can be adaptively expanded. By predicting the regional space of the overlapping distribution area in the detected image and the detection frame position of the first detection frame in the detected image, the second detection frame can be effectively screened out.

[0118] In one embodiment of the present disclosure, the position of the detection frame may be represented by the coordinate position of the center point of the detection frame in the detected image.

[0119] According to an embodiment of the present disclosure, it is predicted that the overlapping distribution area includes a second target object, and the second target object is a target object having the same clustering attribute among the M first target objects.

[0120] Operation S230, processing the distribution position of the second detection frame and the target object using a preset area algorithm to determine the object overlapping image area in the detected image, may include the following operations:

[0121] The second detection frame and the second distribution position of the second target object corresponding to the second detection frame are processed using a preset area algorithm to obtain an object overlapping image area corresponding to the predicted overlapping distribution area.

[0122] According to an embodiment of the present disclosure, the preset area algorithm includes at least one of the following:

[0123] Triangulation algorithm, minimum region box algorithm.

[0124] According to an embodiment of the present disclosure, the minimum area frame algorithm can be formed based on the function minAreaRect(). When the preset area algorithm is the minimum area frame algorithm, the object overlapping image area can be determined based on formula (2).

[0125] R Overlap =minAreaRect(Distribute(0) d +τ+{B i}) (2)

[0126] In formula (1), Distribute(O) d represents the second distribution position of the second target object in the predicted overlapping distribution area, τ represents the preset boundary distance, {B i} represents the second detection box corresponding to the predicted overlapping distribution area, R overlap Indicates that objects overlap image areas.

[0127] According to an embodiment of the present disclosure, the triangulation algorithm may include a Delaunay triangulation algorithm in related technologies, and the like.

[0128] According to the embodiments of the present disclosure, by determining the object overlapping image area, the scene area where the target object overlaps can be conveniently determined in the detected image, thereby laying the foundation for further first detection frame screening for the object overlapping image area.

[0129] According to an embodiment of the present disclosure, the first detection frame in the object overlapping image region includes a plurality of first detection frames.

[0130] Before processing the first detection frame in the object overlapping image area using the first detection frame screening algorithm in operation S240, the target object detection method further includes the following operations:

[0131] According to the respective detection frame categories of the multiple first detection frames, the multiple first detection frames are classified to obtain one or more detection frame sets, wherein the first detection frames in the same detection frame set have the same detection frame category; and for the same detection frame set, the detection frame score of the target detection frame in the detection frame set is updated in the following manner, wherein the detection frame score of the target detection frame is greater than the detection frame scores of other detection frames in the detection frame set, and the detection frame score is used to represent the comprehensive detection result of the detection frame position and the detection frame category of the detection frame.

[0132] This method includes the following operations:

[0133] Calculate the detection frame intersection-over-union ratio of the target detection frame and other detection frames in the detection frame set; for each detection frame intersection-over-union ratio, when the detection frame intersection-over-union ratio is greater than a preset attribute threshold, use the position parameter to iteratively optimize the current detection frame score of the target detection frame to obtain the target detection frame score after the iteration; and based on the target detection frame score after the iteration, update the first detection frame score of the first detection frame corresponding to the target detection frame in the object overlapping image area to obtain the target first detection frame score of the first detection frame corresponding to the target detection frame.

[0134] It should be understood that the first detection frame may have detection frame attribute information such as detection frame category and detection frame score. The detection frame category can be used to characterize the classification result of predicting the target object in the detection frame, and the detection frame score can be used to characterize the comprehensive scoring result of the detection frame category and the detection frame position.

[0135] According to an embodiment of the present disclosure, in the related art, the detection frame screening algorithm for scene areas with target overlap usually relies on the overall score of the detection frame, that is, the detection frame is screened based on the comprehensive score result of the detection frame category and the detection frame position, which results in the relevant detection algorithm being unable to accurately describe the detection frame position of the detection frame.

[0136] By updating the detection frame score of the target detection frame of the same detection frame category, the updated target detection frame score can at least partially enhance the detection frame position score of the target detection frame, thereby effectively improving the accuracy of the detection frame position of the detection frame in the object overlapping image area, and further improving the accuracy of the subsequent first detection result of the target object in the object overlapping image area.

[0137] According to an embodiment of the present disclosure, iteratively optimizing the current detection frame score of the target detection frame using position parameters may include the following operations:

[0138] Iteratively calculate the sum of the position parameter and the current detection frame score of the target detection frame;

[0139] The position parameter includes: the product of the number of detection frames in the detection frame set and the preset position parameter and the minimum value of the preset position parameters.

[0140] According to an embodiment of the present disclosure, the detection frame scores of target detection frames in the same detection frame set can be updated using formula (3).

[0141]

[0142] In formula (3), SCORE i Represents the detection frame score of the target detection frame in the detection frame set, iou() represents the intersection-union ratio of the target detection frame to any other detection frame in the same detection frame set, num(L)×δ represents the product of the number of detection frames in the detection frame set and the preset position parameter, δ represents the preset position parameter, and L represents the number of detection frames in the detection frame set. Indicates the preset attribute threshold.

[0143] In the detection frame set, the target detection frame and other detection frames in the detection frame set can be traversed using formula (3), so that when the intersection and union ratio of the target detection frame and other detection frames in the detection frame set is greater than the preset attribute threshold, the current detection frame score of the target detection frame is updated once, so that after traversing all L detection frames in the detection frame set, the detection frame score of the target detection frame can be updated, thereby obtaining the iterated target detection frame score.

[0144] Based on the association between the target detection frame and the first detection frame in the detected image, the iterated target detection frame score can be used to update the detection frame score of the first detection frame corresponding to the target detection frame, so that the target first detection frame score can at least partially enhance the detection frame position score of the first detection frame, thereby effectively improving the accuracy of the detection frame position of the detection frame in the object overlapping image area, and thus achieving the improvement of the accuracy of the subsequent first detection results of the target object in the object overlapping image area.

[0145] According to an embodiment of the present disclosure, the target object detection method may further include the following operations:

[0146] The object overlapping image area is used to screen out the object non-overlapping area in the detected image; and the first detection frame in the object non-overlapping image area is processed by a second detection frame screening algorithm to obtain a second detection result of the target object in the detected image.

[0147] According to an embodiment of the present disclosure, after determining the object overlapping image area, other image areas in the detected image except the object overlapping image area can also be used as the object non-overlapping area, so that different detection frame screening algorithms can be used to process the first detection frame in the corresponding image area, so as to improve the adaptability of the detection frame screening algorithm to scene areas with different attributes in the detected image, and thereby improve the accuracy of the second detection result for the target object.

[0148] According to an embodiment of the present disclosure, the first detection frame screening algorithm includes a non-maximum suppression algorithm; and / or the second detection frame screening algorithm includes a smoothing-non-maximum suppression algorithm.

[0149] According to an embodiment of the present disclosure, the non-maximum suppression algorithm may include a non-maximum suppression (NMS) algorithm in the related art. The smoothing-non-maximum suppression algorithm may include a SOFT-Non-Maximum Suppression (also known as a SOFT-NMS algorithm).

[0150] According to the embodiments of the present disclosure, by determining the object overlapping image area and the object non-overlapping area in the detected image, and respectively using the NMS algorithm and the SOFT-NMS algorithm to process the object overlapping image area and the object non-overlapping area, the problem of mis-screening of the detection frame by the NMS algorithm in the image scene area where multiple target objects overlap can be at least partially solved, and the problem of under-screening (i.e., under-filtering problem) of the detection frame by the SOFT-NMS algorithm in the image scene area where a single target object exists can also be at least partially solved, thereby achieving the technical effect of improving the screening accuracy for the first detection frame and improving the accuracy of target object detection.

[0151] According to an embodiment of the present disclosure, the target object detection method may further include the following operations:

[0152] A target object detection result for the detected image is determined according to the first target object detection result and the second target object detection result.

[0153] According to an embodiment of the present disclosure, the following formulas (4) to (7) may be used to determine the target object detection result of the detected image.

[0154]

[0155] FRsoft-nms =SOFTNMS({B i |B i in R overlap}); (5)

[0156] FR nms =NMS({B i |B i in R normal}); (6)

[0157] Result = FR soft-nms UFR nms ; (7)

[0158] In formula (4) to formula (7), SCORE i represents the detection box score of the first detection box in the object overlapping image area, B i and B j Represents different first detection boxes, R overlap Represents the overlapping area of ​​objects in the detected image, R normal Indicates the non-overlapping area of ​​the object in the detected image, FR soft-nms The first detection result set representing the first detection result of the target object, FR nms The second detection result set represents a second detection result of the target object, and Result represents the target object detection result of the detected image.

[0159] According to an embodiment of the present disclosure, the first detection result set and / or the second detection result set may be a detection frame for the target object in the detected image obtained after filtering the first detection frame, and the filtered detection frame may have a category prediction result for the target object and the position of the detection frame.

[0160] It should be noted that, when the first detection frame corresponds to the target detection frame, SCORE i It may be the target first detection frame score obtained by updating the detection frame score of the first detection frame based on the method in the above embodiment.

[0161] It should be understood that formula (4) can be used to represent the specific calculation process of the SOFT-NMS algorithm in the related art.

[0162] Figure 5 The following schematically illustrates an application scenario of a target object detection method according to an embodiment of the present disclosure.

[0163] like Figure 5 As shown, the application scenario includes a detected image 500, and the detected image 500 can be captured by a camera installed on a vehicle C510.

[0164] The camera device 511 in the detected image 500 is used as a target object pre-arranged in the image acquisition space. At the same time, when multiple marker barrels 512 are arranged on the road according to their respective preset positions, the multiple marker barrels can also be used as other target objects.

[0165] According to the target object detection method provided by the above embodiment, the object overlapping image area 520 can be determined in the detected image 500, and correspondingly, other image areas in the detected image 500 except the object overlapping image area 520 are determined as object non-overlapping image areas.

[0166] Then, the SOFT-NMS algorithm can be used to process the first detection frame in the object overlapping image region 520, and the NMS algorithm can be used to process the first detection frame in the object non-overlapping image region in the detected image 500. Then, based on the first target object detection result and the second target object detection result obtained, a target object detection result for the detected image 500 is determined. For example, a detection frame for a target object such as a vehicle, a tree, a sign barrel, or a pedestrian in the detected image 500 can be obtained.

[0167] Figure 6 The block diagram schematically shows a target object detection device according to an embodiment of the present disclosure.

[0168] like Figure 6 As shown, the target object detection device 600 includes a first determination module 610 , a first screening module 620 , a second determination module 630 and a first detection module 640 .

[0169] The first determination module 610 is configured to determine a predicted overlapping distribution area according to respective distribution positions of N target objects in the detected image, where N is a positive integer.

[0170] The first screening module 620 is used to screen out second detection frames corresponding to the predicted overlapping distribution area from the first detection frame set using the predicted overlapping distribution area, wherein the first detection frame set is the detection frame obtained after target object detection is performed on the detected image.

[0171] The second determination module 630 is configured to process the distribution positions of the second detection frame and the target object using a preset area algorithm, so as to determine the object overlapping image area in the detected image.

[0172] The first detection module 640 is configured to process the first detection frame in the object overlapping image region by using a first detection frame screening algorithm to obtain a first detection result of the target object in the detected image.

[0173] According to an embodiment of the present disclosure, the first determination module includes: a first screening unit and a first clustering unit.

[0174] The first screening unit is used to screen out M first target objects from the N target objects according to respective distribution positions of the N target objects in the detected image and a preset position screening rule.

[0175] The first clustering unit is configured to perform cluster analysis on the M first target objects according to the first distribution positions of the respective first target objects to obtain predicted overlapping distribution areas.

[0176] According to an embodiment of the present disclosure, the first screening module includes: an area expansion unit and a second screening unit.

[0177] The region expansion unit is used to expand the predicted overlapping distribution area according to a preset boundary distance to obtain a detection frame screening area.

[0178] The second screening unit is used to screen out the first detection frames located in the detection frame screening area from the first detection frame set according to the first detection frame positions of the first detection frames in the first detection frame set, and obtain a second detection frame corresponding to the predicted overlapping distribution area.

[0179] According to an embodiment of the present disclosure, it is predicted that the overlapping distribution area includes a second target object, and the second target object is a target object having the same clustering attribute among the M first target objects.

[0180] The second determination module includes: a regional algorithm processing unit.

[0181] The regional algorithm processing unit is used to use a preset regional algorithm to process the second detection frame and the second distribution position of the second target object corresponding to the second detection frame to obtain an object overlapping image area corresponding to the predicted overlapping distribution area.

[0182] According to an embodiment of the present disclosure, the first detection frame in the object overlapping image region includes a plurality of first detection frames.

[0183] Before using the first detection frame screening algorithm to process the first detection frame in the object overlapping image area, the target object detection device may further include: a first classification module and a detection frame score updating module.

[0184] The first classification module is used to classify the multiple first detection frames according to their respective detection frame categories to obtain one or more detection frame sets, wherein the first detection frames in the same detection frame set have the same detection frame category.

[0185] The detection frame score update module is used to update the detection frame score of the target detection frame in the same detection frame set in the following manner, wherein the detection frame score of the target detection frame is greater than the detection frame scores of other detection frames in the detection frame set. The detection frame score is used to represent the comprehensive detection result of the detection frame position and detection frame category of the detection frame. The methods include:

[0186] Calculate the intersection-union ratio of the target detection frame and the other detection frames in the detection frame set;

[0187] For each detection frame intersection-over-union ratio, when the detection frame intersection-over-union ratio is greater than the preset attribute threshold, the position parameter is used to iteratively optimize the current detection frame score of the target detection frame to obtain the target detection frame score after iteration;

[0188] According to the target detection frame score after the iteration of the target detection frame, the first detection frame score of the first detection frame corresponding to the target detection frame in the object overlapping image area is updated to obtain the target first detection frame score of the first detection frame corresponding to the target detection frame.

[0189] According to an embodiment of the present disclosure, iteratively optimizing the current detection frame score of the target detection frame using position parameters includes:

[0190] Iteratively calculate the sum of the position parameter and the current detection frame score of the target detection frame;

[0191] The position parameter includes: the product of the number of detection frames in the detection frame set and the preset position parameter and the minimum value of the preset position parameters.

[0192] According to an embodiment of the present disclosure, the target object detection device further includes: a second screening module and a second detection module.

[0193] The second screening module is used to screen out non-overlapping areas of objects in the detected image by using the overlapping image areas of the objects.

[0194] The second detection module is used to process the first detection frame in the non-overlapping image area of ​​the object by using a second detection frame screening algorithm to obtain a second detection result for the target object in the detected image.

[0195] According to an embodiment of the present disclosure, the target object detection device further includes: a target object detection result determination module.

[0196] The target object detection result determination module is used to determine the target object detection result for the detected image according to the first target object detection result and the second target object detection result.

[0197] According to an embodiment of the present disclosure, the first detection frame screening algorithm includes a non-maximum suppression algorithm; and / or the second detection frame screening algorithm includes a smoothing-non-maximum suppression algorithm.

[0198] According to an embodiment of the present disclosure, the preset area algorithm includes at least one of the following:

[0199] Triangulation algorithm, minimum region box algorithm.

[0200] According to an embodiment of the present disclosure, the detected image is acquired by an image acquisition device, and N target objects are pre-arranged in an image acquisition space corresponding to the detected image.

[0201] The target object detection device further includes: a position conversion module.

[0202] The position conversion module is used to convert the preset positions of the N target objects according to the position calibration relationship between the image acquisition position of the image acquisition device and the preset positions of the N target objects, so as to obtain the distribution positions of the N target objects in the detected image.

[0203] According to the modules, submodules, units, and subunits of the embodiments of the present invention, any multiple or at least part of the functions of any multiple thereof can be implemented in one module. According to the modules, units, and any one or more of the embodiments of the present invention, any one or more can be split into multiple modules for implementation. According to the modules, submodules, units, and subunits of the embodiments of the present invention, any one or more can be at least partially implemented as a hardware circuit, such as a field programmable gate array (FPGA), a programmable logic array (PLA), a system on a chip, a system on a substrate, a system on a package, an application specific integrated circuit (ASIC), or can be implemented by hardware or firmware of any other reasonable way of integrating or packaging the circuit, or implemented in any one of the three implementation modes of software, hardware, and firmware or in an appropriate combination of any of several thereof. Alternatively, according to the modules, submodules, units, and subunits of the embodiments of the present invention, one or more can be at least partially implemented as a computer program module, which can perform the corresponding function when the computer program module is run.

[0204] For example, any multiple of the first determination module 610, the first screening module 620, the second determination module 630 and the first detection module 640 can be combined into one module / unit for implementation, or any one of the modules / units can be split into multiple modules / units / sub-units. Alternatively, at least part of the functions of one or more of these modules / units can be combined with at least part of the functions of other modules / units and implemented in one module / unit. According to an embodiment of the present disclosure, at least one of the first determination module 610, the first screening module 620, the second determination module 630 and the first detection module 640 can be at least partially implemented as a hardware circuit, such as a field programmable gate array (FPGA), a programmable logic array (PLA), a system on a chip, a system on a substrate, a system on a package, an application specific integrated circuit (ASIC), or can be implemented by hardware or firmware such as any other reasonable way of integrating or packaging the circuit, or implemented in any one of the three implementation methods of software, hardware and firmware or in an appropriate combination of any of them. Alternatively, at least one of the first determination module 610 , the first screening module 620 , the second determination module 630 and the first detection module 640 may be at least partially implemented as a computer program module, which may perform corresponding functions when executed.

[0205] It should be noted that the target object detection device part in the embodiment of the present disclosure corresponds to the target object detection method part in the embodiment of the present disclosure. The description of the target object detection device part specifically refers to the target object detection method part and will not be repeated here.

[0206] Figure 7 A block diagram of an electronic device suitable for implementing a target object detection method according to an embodiment of the present disclosure is schematically shown. Figure 7 The electronic device shown is only an example and should not limit the functions and scope of use of the embodiments of the present disclosure.

[0207] like Figure 7 As shown, the electronic device 700 according to an embodiment of the present disclosure includes a processor 701, which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 702 or a program loaded from a storage part 708 into a random access memory (RAM) 703. The processor 701 may include, for example, a general-purpose microprocessor (e.g., a CPU), an instruction set processor and / or a related chipset and / or a special-purpose microprocessor (e.g., an application-specific integrated circuit (ASIC)), etc. The processor 701 may also include an onboard memory for caching purposes. The processor 701 may include a single processing unit or multiple processing units for performing different actions of the method flow according to an embodiment of the present disclosure.

[0208] Various programs and data required for the operation of the electronic device 700 are stored in the RAM 703. The processor 701, ROM 702, and RAM 703 are connected to each other via a bus 704. The processor 701 executes the various operations of the method flow according to the embodiment of the present disclosure by executing the programs in the ROM 702 and / or RAM 703. It should be noted that the programs may also be stored in one or more memories other than the ROM 702 and the RAM 703. The processor 701 may also execute the various operations of the method flow according to the embodiment of the present disclosure by executing the programs stored in the one or more memories.

[0209] According to an embodiment of the present disclosure, electronic device 700 may further include an input / output (I / O) interface 705, which is also connected to bus 704. System 700 may also include one or more of the following components connected to I / O interface 705: an input section 706 including a keyboard, a mouse, etc.; an output section 707 including devices such as a cathode ray tube (CRT), a liquid crystal display (LCD), and speakers; a storage section 708 including a hard disk; and a communication section 709 including a network interface card such as a LAN card or a modem. Communication section 709 performs communication processing via a network such as the Internet. Drive 710 is also connected to I / O interface 705 as needed. Removable media 711, such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, etc., is installed in drive 710 as needed, so that computer programs read therefrom can be installed into storage section 708 as needed.

[0210] According to an embodiment of the present disclosure, the method flow according to an embodiment of the present disclosure can be implemented as a computer software program. For example, an embodiment of the present disclosure includes a computer program product, which includes a computer program carried on a computer-readable storage medium, and the computer program includes a program code for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from the network through the communication part 709, and / or installed from the removable medium 711. When the computer program is executed by the processor 701, the above-mentioned functions defined in the system of the embodiment of the present disclosure are executed. According to an embodiment of the present disclosure, the system, equipment, device, module, unit, etc. described above can be implemented by a computer program module.

[0211] The present disclosure also provides a computer-readable storage medium, which may be included in the device / apparatus / system described in the above embodiments, or may exist independently and not be incorporated into the device / apparatus / system. The computer-readable storage medium carries one or more programs, and when executed, implements the method according to the embodiments of the present disclosure.

[0212] According to an embodiment of the present disclosure, a computer-readable storage medium may be a non-volatile computer-readable storage medium. For example, it may include, but is not limited to: a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination thereof. In the present disclosure, a computer-readable storage medium may be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, apparatus, or device.

[0213] For example, according to an embodiment of the present disclosure, the computer-readable storage medium may include the ROM 702 and / or the RAM 703 described above and / or one or more memories other than the ROM 702 and the RAM 703 .

[0214] An embodiment of the present disclosure also includes a computer program product, which includes a computer program, which contains program code for executing the method provided by the embodiment of the present disclosure. When the computer program product is run on an electronic device, the program code is used to enable the electronic device to implement the target object detection method provided by the embodiment of the present disclosure.

[0215] When the computer program is executed by the processor 701, the above functions defined in the system / device of the embodiment of the present disclosure are performed. According to the embodiment of the present disclosure, the system, device, module, unit, etc. described above can be implemented by a computer program module.

[0216] In one embodiment, the computer program may be stored on a tangible storage medium such as an optical storage device or a magnetic storage device. In another embodiment, the computer program may be transmitted and distributed in the form of a signal on a network medium, downloaded and installed via the communication portion 709, and / or installed from a removable medium 711. The program code contained in the computer program may be transmitted using any appropriate network medium, including but not limited to wireless, wired, or any suitable combination thereof.

[0217] According to an embodiment of the present disclosure, the program code for executing the computer program provided by the embodiment of the present disclosure can be written in any combination of one or more programming languages. Specifically, these computer programs can be implemented using high-level procedural and / or object-oriented programming languages, and / or assembly / machine languages. Programming languages ​​include, but are not limited to, languages ​​such as Java, C++, python, "C" or similar programming languages. The program code can be executed entirely on the user computing device, partially on the user device, partially on a remote computing device, or entirely on a remote computing device or server. In cases involving a remote computing device, the remote computing device can be connected to the user computing device through any type of network, including a local area network (LAN) or a wide area network (WAN), or can be connected to an external computing device (for example, using an Internet service provider to connect via the Internet).

[0218] The flowcharts and block diagrams in the accompanying drawings illustrate the possible architectures, functions, and operations of the systems, methods, and computer program products according to various embodiments of the present disclosure. In this regard, each box in the flowchart or block diagram may represent a module, program segment, or portion of code, which contains one or more executable instructions for implementing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the boxes may occur in an order different from that marked in the accompanying drawings. For example, two boxes shown in succession may actually be executed substantially in parallel, or they may sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each box in the block diagram or flowchart, as well as the combination of boxes in the block diagram or flowchart, may be implemented using a dedicated hardware-based system that performs the specified function or operation, or may be implemented using a combination of dedicated hardware and computer instructions. It will be understood by those skilled in the art that the features described in the various embodiments and / or claims of the present disclosure may be combined and / or coupled in various ways, even if such combinations or couplings are not explicitly described in the present disclosure. In particular, without departing from the spirit and teachings of the present disclosure, the features described in the various embodiments and / or claims of the present disclosure may be combined and / or coupled in various ways, and all such combinations and / or couplings fall within the scope of the present disclosure.

[0219] The embodiments of the present disclosure are described above. However, these embodiments are for illustrative purposes only and are not intended to limit the scope of the present disclosure. Although each embodiment has been described separately above, this does not mean that the measures in each embodiment cannot be used in combination to advantage. The scope of the present disclosure is defined by the appended claims and their equivalents. Without departing from the scope of the present disclosure, those skilled in the art may make various substitutions and modifications, which should all fall within the scope of the present disclosure.

Claims

1. A target object detection method, comprising: Determine the predicted overlapping distribution area according to the respective distribution positions of N target objects in the detected image, where N is a positive integer; Using the predicted overlapping distribution area, filter out a second detection frame corresponding to the predicted overlapping distribution area from a first detection frame set, wherein the first detection frame set is a detection frame obtained after performing target object detection on the detected image; Processing the distribution positions of the second detection frame and the target object using a preset area algorithm to facilitate determining an object overlapping image area in the detected image; and A first detection frame screening algorithm is used to process the first detection frame in the object overlapping image area to obtain a first detection result of the target object in the detected image.

2. The method according to claim 1, wherein determining the predicted overlapping distribution area based on the respective distribution positions of the N target objects in the detected image comprises: According to the respective distribution positions of the N target objects in the detected image, M first target objects are screened out from the N target objects according to a preset position screening rule; A cluster analysis is performed on the M first target objects according to the first distribution positions of the respective first target objects to obtain the predicted overlapping distribution area.

3. The method according to claim 2, wherein: Filtering a second detection frame corresponding to the predicted overlapping distribution area from the first detection frame set using the predicted overlapping distribution area includes: Expanding the predicted overlapping distribution area according to a preset boundary distance to obtain a detection frame screening area; and According to the first detection frame positions of the respective first detection frames in the first detection frame set, the first detection frames located in the detection frame screening area are screened out from the first detection frame set to obtain a second detection frame corresponding to the predicted overlapping distribution area.

4. The method according to claim 2, wherein: The predicted overlapping distribution area includes a second target object, where the second target object is a target object having the same clustering attribute among the M first target objects; Processing the distribution position of the second detection frame and the target object using a preset area algorithm to determine an object overlapping image area in the detected image includes: The second detection frame and the second distribution position of the second target object corresponding to the second detection frame are processed using the preset area algorithm to obtain an object overlapping image area corresponding to the predicted overlapping distribution area.

5. The method according to claim 1, wherein The first detection frame in the object overlapping image area includes a plurality of; Before processing the first detection frame in the object overlapping image area using the first detection frame screening algorithm, the target object detection method further includes: classifying the plurality of first detection frames according to their respective detection frame categories to obtain one or more detection frame sets, wherein the first detection frames in the same detection frame set have the same detection frame category; For the same detection frame set, updating the detection frame score of the target detection frame in the detection frame set in the following manner, wherein the detection frame score of the target detection frame is greater than the detection frame scores of other detection frames in the detection frame set, and the detection frame score is used to represent the comprehensive detection result of the detection frame position and detection frame category of the detection frame, the manner comprising: Calculating the detection frame intersection-union ratio between the target detection frame and other detection frames in the detection frame set; For each of the detection frame intersection-over-union ratios, when the detection frame intersection-over-union ratio is greater than a preset attribute threshold, iteratively optimizing the current detection frame score of the target detection frame using the position parameter to obtain the target detection frame score after iteration; According to the target detection frame score after iteration of the target detection frame, the first detection frame score of the first detection frame corresponding to the target detection frame in the object overlapping image area is updated to obtain the target first detection frame score of the first detection frame corresponding to the target detection frame.

6. The method according to claim 5, wherein iteratively optimizing the current detection frame score of the target detection frame using position parameters comprises: Iteratively calculating the sum of the position parameter and the current detection frame score of the target detection frame; The position parameter includes: the product of the number of detection frames in the detection frame set and the preset position parameter and the minimum value of the preset position parameters.

7. The method according to any one of claims 1 to 6, further comprising: Filtering out object non-overlapping areas in the detected image using the object overlapping image areas; as well as The first detection frame in the non-overlapping image area of ​​the object is processed using a second detection frame screening algorithm to obtain a second detection result of the target object in the detected image.

8. The method according to claim 7, further comprising: A target object detection result for the detected image is determined according to the first target object detection result and the second target object detection result.

9. The method according to claim 7, wherein: The first detection frame screening algorithm includes a non-maximum suppression algorithm; and / or The second detection frame screening algorithm includes a smoothing-non-maximum suppression algorithm.

10. The method according to claim 1, wherein The preset area algorithm includes at least one of the following: Triangulation algorithm, minimum region box algorithm.

11. The method according to claim 1, wherein The detected image is acquired by an image acquisition device, and the N target objects are pre-arranged in an image acquisition space corresponding to the detected image; The target object detection method further includes: According to the position calibration relationship between the image acquisition position of the image acquisition device and the respective preset positions of the N target objects, the respective preset positions of the N target objects are converted to obtain the respective distribution positions of the N target objects in the detected image.

12. A target object detection device, comprising: A first determination module is configured to determine a predicted overlapping distribution area based on respective distribution positions of N target objects in the detected image, where N is a positive integer; a first screening module, configured to screen out second detection frames corresponding to the predicted overlapping distribution area from a first detection frame set using the predicted overlapping distribution area, wherein the first detection frame set is detection frames obtained after performing target object detection on the detected image; a second determining module, configured to process the distribution positions of the second detection frame and the target object using a preset area algorithm, so as to determine an object overlapping image area in the detected image; and The first detection module is configured to process a first detection frame in the object overlapping image area by using a first detection frame screening algorithm to obtain a first detection result of the target object in the detected image.

13. An electronic device comprising: one or more processors; a memory for storing one or more programs, When the one or more programs are executed by the one or more processors, the one or more processors are enabled to implement the method according to any one of claims 1 to 11.

14. A computer-readable storage medium having executable instructions stored thereon, which, when executed by a processor, causes the processor to implement the method according to any one of claims 1 to 11.

15. A computer program product comprising a computer program, which, when executed by a processor, implements the method according to any one of claims 1 to 11.

Images