Abnormal object processing method and device, equipment, medium and product

Abstract

The invention discloses an abnormal object processing method and device, equipment, a medium and a product. The method is applied to a target ship, and comprises the following steps: obtaining a to-be-processed image comprising a to-be-recognized object: processing the to-be-processed image based on an abnormal object recognition model to obtain an object recognition result corresponding to the to-be-recognized object; the object recognition result comprises a recognition category and bounding box information corresponding to the to-be-recognized object; determining a target processing device and a target device position corresponding to the abnormal object according to the identification category and bounding box information corresponding to the abnormal object under the condition that the to-be-processed image is determined to comprise the abnormal object based on the identification category; and setting a target processing device according to the target device position, and processing the corresponding abnormal object based on the set target processing device. According to the technical scheme, the effect of quickly recognizing the abnormal object associated with the ship based on the neural network model is achieved, and the recognition and processing efficiency of the abnormal object is improved.

Description

Technical Field

[0001] The present invention relates to the field of ship safety technology, and in particular to an abnormal object processing method, device, equipment, medium and product. Background Art

[0002] Abnormal objects on the ocean bring huge economic losses to ship owners, and escort warships are one of the important measures to prevent abnormal objects from boarding the ship. However, how to prevent abnormal objects from boarding the ship before the escort warships arrive is the difficulty of preventing abnormal objects.

[0003] In the related art, during the navigation process of a ship, abnormal objects are usually identified manually, and when abnormal objects are found, high-pressure water guns are manually operated to prevent abnormal objects from boarding the ship. Such object handling methods may be time-consuming and labor-intensive, with poor object recognition accuracy, and may even affect the personal safety of the crew. Summary of the invention

[0004] The present invention provides an abnormal object processing method, device, equipment, medium and product to achieve the effect of quickly identifying abnormal objects associated with ships based on a neural network model, thereby improving the recognition efficiency of abnormal objects and further improving the processing efficiency of abnormal objects.

[0005] According to one aspect of the present invention, a method for processing an abnormal object is provided, the method being applied to a target ship, comprising:

[0006] Get the image to be processed that includes the object to be recognized:

[0007] The image to be processed is processed based on a pre-trained abnormal object recognition model to obtain an object recognition result corresponding to the object to be recognized; wherein the abnormal object recognition model is used to recognize abnormal objects associated with the target ship; the abnormal objects are objects that need to be prevented from approaching or boarding the target ship; the object recognition result includes recognition categories and bounding box information corresponding to the object to be recognized;

[0008] In the case where it is determined based on the recognition category that the image to be processed includes an abnormal object, a target processing device and a target device position corresponding to the abnormal object are determined according to the recognition category and the bounding box information corresponding to the abnormal object; wherein the target processing device is used to process the abnormal object so that the abnormal object is away from the target ship;

[0009] The target processing device is set according to the position of the target device, and the corresponding abnormal object is processed based on the set target processing device.

[0010] According to another aspect of the present invention, there is provided an abnormal object processing device, which is configured in a target ship and comprises:

[0011] An image acquisition module is used to acquire an image to be processed including an object to be identified:

[0012] An object recognition module is used to process the image to be processed based on a pre-trained abnormal object recognition model to obtain an object recognition result corresponding to the object to be recognized; wherein the abnormal object recognition model is used to recognize abnormal objects associated with the target ship; the abnormal objects are objects that need to be prevented from approaching or boarding the target ship; and the object recognition result includes recognition categories and bounding box information corresponding to the object to be recognized;

[0013] a processing device determination module, configured to determine a target processing device and a target device position corresponding to the abnormal object according to the identification category and the bounding box information corresponding to the abnormal object, when it is determined that the image to be processed includes an abnormal object based on the identification category; wherein the target processing device is configured to process the abnormal object so that the abnormal object is away from the target ship;

[0014] The object processing module is used to set the target processing device according to the position of the target device, and process the corresponding abnormal object based on the set target processing device.

[0015] According to another aspect of the present invention, there is provided an electronic device, the electronic device comprising:

[0016] at least one processor; and

[0017] a memory communicatively connected to the at least one processor; wherein,

[0018] The memory stores a computer program executable by the at least one processor, and the computer program is executed by the at least one processor so that the at least one processor can execute the abnormal object processing method described in any embodiment of the present invention.

[0019] According to another aspect of the present invention, a computer-readable storage medium is provided, wherein the computer-readable storage medium stores computer instructions, and the computer instructions are used to enable a processor to implement the abnormal object processing method described in any embodiment of the present invention when executed.

[0020] According to another aspect of the present invention, a computer program product is provided. The computer program product includes a computer program. When the computer program is executed by a processor, the abnormal object processing method according to any embodiment of the present invention is implemented.

[0021] The technical solution of the embodiment of the present invention obtains an image to be processed including an object to be identified, and further processes the image to be processed based on a pre-trained abnormal object recognition model to obtain an object recognition result corresponding to the object to be identified; wherein the object recognition result includes a recognition category and bounding box information corresponding to the object to be identified, and further, when it is determined based on the recognition category that the image to be processed includes an abnormal object, the target processing device and the target device position corresponding to the abnormal object are determined according to the recognition category and bounding box information corresponding to the abnormal object, and further, the target processing device is set according to the target device position, and the corresponding abnormal object is processed based on the set target processing device, so as to solve the problem that the abnormal object processing method in the related art is time-consuming and labor-intensive, the object recognition accuracy is poor, and it may even affect the personal safety of the crew, and realizes the effect of quickly identifying abnormal objects associated with the ship based on the neural network model, improves the recognition efficiency of abnormal objects, and further improves the processing efficiency of abnormal objects, and realizes the effect of effectively processing abnormal objects without human intervention, and further improves the navigation safety of the ship.

[0022] It should be understood that the contents described in this section are not intended to identify the key or important features of the embodiments of the present invention, nor are they intended to limit the scope of the present invention. Other features of the present invention will become easily understood through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0024] Figure 1 is a flowchart of a method for processing an abnormal object provided according to Embodiment 1 of the present invention;

[0025] Figure 2 is a structural diagram of an abnormal object processing system applicable to Embodiment 1 of the present invention;

[0026] Figure 3 is a flowchart of a method for processing an abnormal object provided according to Embodiment 2 of the present invention;

[0027] Figure 4 is a flowchart of a method for processing an abnormal object provided according to Embodiment 3 of the present invention;

[0028] Figure 5 is a schematic diagram of the structure of an object processing device provided according to a fourth embodiment of the present invention;

[0029] Figure 6 The invention is a schematic diagram of the structure of an electronic device for implementing the abnormal object processing method according to an embodiment of the present invention. DETAILED DESCRIPTION

[0030] In order to enable those skilled in the art to better understand the scheme of the present invention, the technical scheme in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present invention.

[0031] It should be noted that the terms "first", "second", etc. in the specification and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments of the present invention described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0032] Embodiment 1

[0033] Figure 1 This is a flowchart of an abnormal object processing method provided by the first embodiment of the present invention. This embodiment can be applied to the case where an abnormal object close to or about to board a ship is processed during the navigation of the ship. The method can be applied to the target ship and executed by an abnormal object processing device. The abnormal object processing device can be implemented in the form of hardware and / or software, and the abnormal object processing device can be configured in a terminal and / or a server. Figure 1 As shown, the method includes:

[0034] S110: Acquire an image to be processed including an object to be identified.

[0035] The target ship may be any ship. The image to be processed may be an image captured in real time by a camera installed on the target ship, or may be a stored image retrieved from a relevant database by a server installed on the target ship. The image types of the image to be processed include color images and / or infrared images. A color image refers to an image in which each pixel is composed of red (R), green (G), and blue (B) components. An infrared image, also known as a thermal image, is an image formed by a thermal infrared scanner receiving and recording the thermal radiation energy reflected by a target object. In this embodiment, the color image including the object to be identified may be an image captured during the day; the infrared image including the object to be identified may be an image captured at night. Optionally, during the navigation of the target ship in the time period corresponding to the day, the camera installed on the target ship may capture images of the surrounding sea area of ​​the target ship, and use the captured images including the object to be identified as color images; during the navigation of the target ship in the time period corresponding to the night, the camera installed on the target ship may capture images of the surrounding sea area of ​​the target ship, and use the captured images including the object to be identified as infrared images. At the same time, the image to be processed includes one or more objects. It can be understood that the object in the image to be processed is the object to be identified. Based on the neural network model provided in the embodiment of the present invention, the identification result corresponding to the object to be identified can be determined. The object to be identified can be an object to be identified in the image to be processed. The object to be identified can be any object, and optionally, it can be other ships other than the target ship and / or devices hooked on the target ship and configured on the non-target ship.

[0036] It should be noted that in a specific application scenario, the image to be processed can be acquired in real time or periodically, or when it is detected that a user has uploaded an image, the image can be acquired and used as the image to be processed. This embodiment does not specifically limit this.

[0037] Exemplarily, when the target ship is sailing at sea, the surrounding area of ​​the target ship can be photographed based on the camera device set on the target ship, and the photographed image is used as the image to be processed. Further, the image to be processed can be identified based on the relevant algorithm. Then, it is determined that the object in the image to be processed is the object to be identified. The number of objects to be identified included in the image to be processed can be one or more, and this embodiment does not specifically limit this. It should be noted that, when the number of objects to be identified is multiple, all objects in the image can be used as objects to be identified; when the number of objects to be identified is one, the objects to be identified in the image can be calibrated in advance, and the calibrated image can be uploaded to the camera device or server, so that the camera device or server stores the characteristic attributes of the object to be identified. Then, after acquiring multiple images to be processed, when it is detected that at least one of the images to be processed includes the characteristic attributes of the object to be identified, the object in the image to be processed can be determined as the object to be identified, and the image to be processed can be processed.

[0038] S120, processing the image to be processed based on the abnormal object recognition model obtained by pre-training to obtain an object recognition result corresponding to the object to be recognized; wherein the object recognition result includes the recognition category and bounding box information corresponding to the object to be recognized.

[0039] In this embodiment, the abnormal object recognition model can be deployed in a camera device for collecting images to be processed, or can be deployed in a server, which is not specifically limited in this embodiment. Among them, the abnormal object recognition model can be a neural network model that uses an image including an object to be recognized as an input object to perform abnormal object recognition on the object to be recognized in the image, that is, a neural network model for identifying whether the object to be recognized is an abnormal object. The abnormal object recognition model can be used to identify abnormal objects associated with the target ship. An abnormal object is an object that needs to be prevented from approaching or boarding the target ship. The abnormal object can include any object, optionally including an abnormal ship close to the target ship and / or an abnormal hook hooked on the target ship. In other words, the abnormal object recognition model can be used to identify abnormal ships close to the target ship and / or abnormal hooks hooked on the target ship. The abnormal ship can be a ship that threatens the target ship. The abnormal hook can be a tool used by the crew on the abnormal ship to board the target ship. Generally, during the navigation of the target ship, there may be a situation where a ship that threatens the target ship approaches the target ship, and the ship close to the target ship can be regarded as an abnormal ship. Furthermore, in order to board the target ship, the crew on the abnormal ship will hook the hook on the side of the target ship, and the hook can be used as an abnormal hook. The abnormal object recognition model can be a deep learning model. More specifically, the abnormal object recognition model can be a target detection model. In this embodiment, the abnormal object recognition model can be set according to actual needs, for example, it can include a YOLO-V8 network.

[0040] In this embodiment, when an image to be processed including an object to be identified is obtained, the image to be processed can be input into an abnormal object recognition model for processing. Then, the model can output an object recognition result corresponding to the object to be identified. The object recognition result can be information used to characterize the attributes of the object to be identified, and can determine the object category of the object to be identified. The object recognition result can include multiple pieces of information used to characterize the object attributes of the object to be identified, optionally including at least two identification categories and bounding box information corresponding to the object to be identified and the category probability corresponding to each identification category. The identification category can be understood as the abnormal object category to which the object to be identified belongs based on the abnormal object recognition model. Optionally, the identification category includes abnormal ships, abnormal hooks, unidentifiable or not included. The identification category of unidentifiable can represent that the object to be identified in the bounding box is neither an abnormal ship nor a hook. The identification category of not included can represent that the bounding box does not include an abnormal ship or a hook. The bounding box information can be information associated with the predicted bounding box obtained after positioning the display position of the object to be identified in the image to be processed. Optionally, the bounding box information includes the center point coordinates of the bounding box and the width and height information of the bounding box. For example, assuming that the object to be identified in the image to be processed is a ship, after the image to be processed is processed based on the abnormal object recognition model, the object recognition result output by the model may be the recognition category corresponding to the ship as an abnormal ship and the center point coordinates and width and height information of the bounding box of the ship in the image to be processed.

[0041] As an optional implementation of this embodiment, during the navigation of the target ship, the surrounding sea area of ​​the target ship can be photographed by a camera device installed on the target ship, and the image including the object to be identified obtained by the photographing is used as the image to be processed. Further, the image to be processed can be input into an abnormal object recognition model pre-deployed in the camera device to process the image to be processed based on the abnormal object recognition model. Then, the object recognition result corresponding to the object to be identified can be output.

[0042] S130. When it is determined based on the recognition category that the image to be processed includes an abnormal object, determine a target processing device and a target device position corresponding to the abnormal object according to the recognition category and bounding box information corresponding to the abnormal object.

[0043] Among them, the target processing device may be a device for processing an abnormal object so that the abnormal object is away from the target ship. The device category corresponding to the target processing device may be corresponding to the identification category of the abnormal object. That is, for different identification categories of abnormal objects, object processing devices of different device categories can be selected as the target processing device. Optionally, the processing device category of the target processing device includes a high-pressure water gun or a decoupling device. Exemplarily, in the case where the identification category corresponding to the abnormal object is an abnormal ship, the processing device category of the target processing device corresponding to the abnormal object is a high-pressure water gun; in the case where the identification category corresponding to the abnormal object is an abnormal hook hooked on the target ship, the processing device category of the target processing device corresponding to the abnormal object is a decoupling device. The target device position may characterize the specific location of the target processing device on the target ship. The target device position may be the optimal setting position of the target processing device, at which the target processing device is set to be able to process the abnormal object to the greatest extent. The target device position may include device installation position information and position information to which the candidate processing device is moved, etc.

[0044] It should be noted that the number of objects to be identified included in the image to be processed may be multiple, and further, the number of abnormal objects included in the image to be processed may also be one or more. In the case where the number of objects to be identified included in the image to be processed is multiple, object recognition results corresponding to each object to be identified can be obtained respectively. Further, for multiple objects to be identified, it can be determined whether the object to be identified is an abnormal object based on the object category corresponding to the object to be identified. Further, in the case where multiple abnormal objects are determined based on the object category corresponding to the multiple objects to be identified, the target processing device corresponding to each abnormal object and its position with respect to the corresponding target device can be determined respectively. Furthermore, the corresponding abnormal object can be processed based on the target processing device at the corresponding target device position. That is, the technical solution provided in the embodiment of the present invention can process multiple abnormal objects at the same time.

[0045] It should also be noted that the bounding box information includes the coordinates of the center point of the bounding box, and the coordinates of the center point can be used as the object position of the object to be identified included in the bounding box. Furthermore, the object to be identified can be located based on the object position.

[0046] In this embodiment, when an object recognition result corresponding to the object to be recognized is obtained, the recognition category corresponding to the object to be recognized in the object recognition result can be detected. Further, when the object to be recognized is determined to be an abnormal object based on the recognition category, a target processing device for processing the object to be recognized can be determined based on the recognition category and the bounding box information. Furthermore, the target device position of the target processing device can be determined based on the bounding box information.

[0047] In this embodiment, the target processing device may be a candidate processing device determined from a plurality of pre-set candidate processing devices; or, the target processing device may also be a processing device set on the target ship for processing the determined abnormal object when the object to be identified is determined to be an abnormal object. Furthermore, the target processing device may correspond to at least two determination methods, and the two determination methods may be described below respectively.

[0048] As an optional implementation of this embodiment, when the object to be identified is determined to be an abnormal object based on the identification category, the category of the processing device corresponding to the object to be identified can be determined based on the identification category. Further, the device location information of at least one candidate processing device belonging to the category of the processing device pre-set on the target ship is obtained. Further, for at least one candidate processing device, the distance between the candidate processing device and the abnormal object can be determined based on the device location and bounding box information of the candidate processing device. Then, at least one distance is obtained. Further, the distance corresponding to the minimum distance value can be determined from at least one distance, and the candidate processing device corresponding to the distance can be used as the target processing device. Further, the device location information corresponding to the target processing device can be determined based on the bounding box information, the device location of the target processing device, and the device movement range.

[0049] As another optional implementation of this embodiment, when the object to be identified is determined to be an abnormal object according to the identification category, the processing device category corresponding to the object to be identified can be determined according to the identification category, and the object processing device belonging to the processing device category is used as the target processing device. Further, the candidate setting position closest to the abnormal object can be determined from multiple candidate setting positions pre-determined on the target ship according to the bounding box information, and the candidate setting position is used as the target device position.

[0050] S140: Setting a target processing device according to the location of the target device, and processing the abnormal object based on the set target processing device.

[0051] In this embodiment, when the target processing device for processing the abnormal object and the target device position of the target processing device are determined, the target processing device can be set according to the target device position to set the target processing device at the target device position on the target ship. Further, the set target processing device can be controlled to process the abnormal object.

[0052] Optionally, setting the target processing device according to the target device position includes: moving the determined target processing device pre-set on the target ship to the target device position; or, installing the target processing device at the target device position, etc.

[0053] Exemplarily, assuming that the abnormal object included in the image to be processed includes an abnormal ship, the target processing device is determined to be a target high-pressure water gun pre-set on the target ship and closest to the abnormal ship. Further, the target high-pressure water gun can be moved from the current device position to the target device position. Thereafter, when the target high-pressure water gun moves to the target device position, the target high-pressure water gun can be controlled to spray the abnormal ship according to a predetermined spray angle and spray pressure to prevent the abnormal ship from approaching the target ship.

[0054] Exemplarily, assuming that the abnormal object included in the image to be processed includes an abnormal hook, the target processing device is determined to be a target unhooking device pre-set on the target ship and closest to the abnormal hook. Further, the target unhooking device can be moved from the current device position to the target device position. Thereafter, when the target unhooking device moves to the target device position, the target unhooking device can be controlled to perform an unhooking operation on the abnormal hook so that the abnormal hook is no longer hooked on the target ship.

[0055] For example, Figure 2 FIG. 1 is a schematic diagram of a structure of an abnormal object processing system applicable to an embodiment of the present invention. Figure 2 As shown, the object processing system is set on the target ship, including a camera device 21, a computer 22, a decoupling device 23, a high-pressure water gun device 24, an inner guide rail 25 and an outer guide rail 26. Among them, the number of camera devices 21 is 4, which are respectively deployed at the bow, stern and left and right sides of the ship; each camera device 21 is deployed with an abnormal object recognition model. The inner guide rail 25 and the outer guide rail 26 are set around the ship's side. There are 7 decoupling devices 23, which are respectively deployed on the outer guide rails 26 at the bow, stern and left and right sides of the ship. There are 7 high-pressure water gun devices 24, which are respectively deployed on the inner guide rails 25 at the bow, stern and left and right sides of the ship. During the navigation of the target ship, the camera device 21 can be used to collect images of the surrounding sea area of ​​the target ship, and the collected images including the objects to be identified are used as images to be processed. Further, the image to be processed can be processed based on the abnormal object recognition model deployed in the camera device 21, and an object recognition result corresponding to the object to be recognized can be obtained, and the object recognition result includes the recognition category and bounding box information corresponding to the object to be recognized. Further, in the case where it is determined based on the recognition category that the image to be processed includes an abnormal object, the bounding box information and the recognition category of the abnormal object are sent to the computer 22. Further, the target processing device (such as the unhooking device 23 and / or the high-pressure water gun device 24) and the target device position for processing the abnormal object can be determined by the computer 22 based on the bounding box information and the recognition category of the abnormal object. Further, the target processing device can be controlled based on the computer 22 to move to the target device position along the inner guide rail 25 and / or the outer guide rail 26, and the target processing device can be started to process the abnormal object.

[0056] The technical solution of the embodiment of the present invention obtains an image to be processed including an object to be identified, and further processes the image to be processed based on a pre-trained abnormal object recognition model to obtain an object recognition result corresponding to the object to be identified; wherein the object recognition result includes a recognition category and bounding box information corresponding to the object to be identified, and further, when it is determined based on the recognition category that the image to be processed includes an abnormal object, the target processing device and the target device position corresponding to the abnormal object are determined according to the recognition category and bounding box information corresponding to the abnormal object, and further, the target processing device is set according to the target device position, and the corresponding abnormal object is processed based on the set target processing device, so as to solve the problem that the abnormal object processing method in the related art is time-consuming and labor-intensive, the object recognition accuracy is poor, and it may even affect the personal safety of the crew, and realizes the effect of quickly identifying abnormal objects associated with the ship based on the neural network model, improves the recognition efficiency of abnormal objects, and further improves the processing efficiency of abnormal objects, and realizes the effect of effectively processing abnormal objects without human intervention, and further improves the navigation safety of the ship.

[0057] Embodiment 2

[0058] Figure 3 It is a flowchart of an abnormal object processing method provided by the second embodiment of the present invention. On the basis of the above-mentioned embodiment, the method for determining the target processing device and the target device position is refined. Optionally, the target processing device and the target device position corresponding to the abnormal object are determined according to the identification category and bounding box information corresponding to the abnormal object, including: determining the target device category corresponding to the abnormal object according to the identification category corresponding to the abnormal object and the preset object device corresponding information; determining the target processing device according to the bounding box information and the target device category corresponding to the abnormal object, and determining the target device position according to the bounding box information corresponding to the abnormal object and the device association parameters corresponding to the target processing device. The specific implementation method can be referred to the technical solution of this embodiment. Among them, the technical terms that are the same or similar to those in the above-mentioned embodiments are not repeated here.

[0059] like Figure 3 As shown, the method includes:

[0060] S210: Acquire an image to be processed including an object to be identified.

[0061] S220, processing the image to be processed based on the pre-trained abnormal object recognition model to obtain an object recognition result corresponding to the object to be recognized; wherein the object recognition result includes a recognition category and bounding box information corresponding to the object to be recognized.

[0062] S230: When it is determined based on the recognition category that the image to be processed includes an abnormal object, determine the target device category corresponding to the abnormal object according to the recognition category corresponding to the abnormal object and preset object device corresponding information.

[0063] Among them, the object device corresponding information is used to indicate the correspondence between the identification category of the abnormal object and the device category corresponding to the object processing device for processing the abnormal object. In this embodiment, the identification categories of multiple abnormal objects can be predetermined, and the device category corresponding to the identification category of each abnormal object can be determined respectively. Further, the object device corresponding information can be determined based on the identification categories of multiple abnormal objects and their corresponding device categories. The target device category can be the category to which the object processing device for processing the abnormal object belongs. Optionally, the processing device category includes a high-pressure water gun and / or a decoupling device.

[0064] As an optional implementation of this embodiment, when it is determined that the image to be processed includes an abnormal object based on the recognition category, the predetermined object device corresponding information can be obtained. Further, the object device corresponding information can be queried according to the recognition category corresponding to the abnormal object to determine the device category corresponding to the recognition category, and the determined device category is used as the target device category corresponding to the abnormal object.

[0065] S240: Determine the target processing device according to the bounding box information corresponding to the abnormal object and the target device category, and determine the target device location according to the bounding box information corresponding to the abnormal object and the device association parameters corresponding to the target processing device.

[0066] In this embodiment, when the target device category corresponding to the abnormal object is determined, the abnormal object can be located according to the bounding box information corresponding to the abnormal object. Furthermore, the target processing device for processing the abnormal object can be determined according to the positioning information and the target device category of the abnormal object. Among them, the target processing device is an object processing device belonging to the target device category corresponding to the abnormal object. The device-associated parameter can be understood as a parameter for characterizing the device performance of the target processing device. The device-associated parameter may include multiple parameters associated with the target processing device, optionally including the movable range of the device.

[0067] In this embodiment, the method of determining the target processing device based on the bounding box information and the target device category corresponding to the abnormal object can correspond to at least two situations, namely, determining the target processing device from at least one candidate processing device of the target device category pre-set on the target ship; and, in the case of determining the target device position based on the bounding box information of the abnormal object, setting an object processing device of the target device category at the target device position, and using the set object processing device as the target processing device.

[0068] It should be noted that, in order to improve the processing efficiency of abnormal objects, preferably, the target processing device can be determined by determining the target processing device from at least one candidate processing device of the target device category pre-set on the target ship.

[0069] Optionally, the target processing device is determined based on the bounding box information and the target device category corresponding to the abnormal object, including: obtaining the current device position of at least one candidate processing device of the target device category pre-set on the target ship; for at least one candidate processing device, determining the distance between the candidate processing device and the abnormal object based on the current device position and the bounding box information; determining a minimum distance from at least one distance, taking the minimum distance as the target distance, and taking the candidate processing device corresponding to the target distance as the target processing device.

[0070] The candidate processing device may be an object processing device pre-installed on the target ship. The current device position may be device position information corresponding to the candidate processing device at the current moment. The current device position may be used to characterize the specific position of the candidate processing device on the target ship at the current moment.

[0071] As an optional implementation of this embodiment, for at least one candidate processing device of the target device category pre-set on the target ship, the current device position of the candidate processing device can be obtained. Further, the current device position of the candidate processing device and the bounding box information of the abnormal object can be processed according to a preset distance determination algorithm to determine the distance between the candidate processing device and the abnormal object. Then, the distance corresponding to at least one candidate processing device can be obtained. Further, the distance corresponding to the minimum distance value can be determined from at least one distance, and the distance can be used as the target distance. Further, the candidate processing device corresponding to the target distance can be used as the target processing device.

[0072] In this embodiment, when the target processing device is determined, the location of the target device may be determined according to the bounding box information corresponding to the abnormal object and the device association parameters corresponding to the target processing device.

[0073] Optionally, the device-associated parameters include a movable range of the device; determining the target device position based on bounding box information corresponding to the abnormal object and the device-associated parameters corresponding to the target processing device, including: determining the target device position based on the current device position of the target processing device, the movable range of the device, and the bounding box information corresponding to the abnormal object.

[0074] The movable range of the device is the range within which the target processing device can be freely moved on the target ship.

[0075] As an optional implementation of this embodiment, when the target processing device is determined, the current device position of the target processing device, the device movable range, and the bounding box information corresponding to the abnormal object can be processed according to a preset position determination algorithm to output the position information, and the output position information is used as the target device position.

[0076] S250: Setting a target processing device according to the location of the target device, and processing a corresponding abnormal object based on the set target processing device.

[0077] In this embodiment, when the target device position and the target processing device are determined, the target processing device can be set according to the target device position.

[0078] Optionally, setting the target processing device according to the target device position includes: moving the target processing device from a current device position to the target device position.

[0079] As an optional implementation of this embodiment, when the target device position and the target processing device are determined, the target processing device can be moved from the current device position to the target device position. Furthermore, the target processing device located at the target device position can be used as the set target processing device. Furthermore, the set target processing device can be controlled to process the corresponding abnormal object.

[0080] The technical solution of the embodiment of the present invention determines the target device category corresponding to the abnormal object according to the identification category corresponding to the abnormal object and the preset object device corresponding information when it is determined that the image to be processed includes an abnormal object based on the identification category; further, determines the target processing device according to the bounding box information corresponding to the abnormal object and the target device category, and determines the target device position according to the bounding box information corresponding to the abnormal object and the device association parameters corresponding to the target processing device, thereby achieving the effect of quickly determining the target processing device for processing the abnormal object and the device setting position of the target processing device according to the object recognition result corresponding to the abnormal object when it is determined that the abnormal object exists, thereby improving the processing efficiency of the abnormal object.

[0081] Embodiment 3

[0082] Figure 4 This is a flowchart of an abnormal object processing method provided by Embodiment 3 of the present invention. Based on the above embodiments, before applying the abnormal object recognition model, the abnormal object recognition model can be pre-trained to recognize the object to be recognized in the image based on the abnormal object recognition model, so as to determine the object recognition result corresponding to the object to be recognized. The specific implementation method can refer to the technical solution of this embodiment. Among them, the technical terms that are the same or similar to the above embodiments are not repeated here.

[0083] S310: Train and obtain an abnormal object recognition model.

[0084] It should be noted that before applying the abnormal object recognition model provided in this embodiment, the pre-constructed deep learning model can be trained in a supervised or unsupervised manner. Before training the deep learning model, multiple training samples can be constructed to train the model based on multiple training samples. In order to improve the recognition accuracy of the abnormal object recognition model, as many and rich training samples as possible can be constructed.

[0085] Optionally, training to obtain an abnormal object recognition model includes: obtaining multiple training samples; and training a pre-built deep learning model based on the multiple training samples to obtain the abnormal object recognition model.

[0086] The training samples include: sample images, the sample images include abnormal objects, and theoretical annotation results corresponding to the sample images, and the theoretical annotation results include theoretical categories and theoretical bounding box information.

[0087] The sample image may be an image taken by a camera, or an image reconstructed by an image reconstruction model, or a pre-stored image retrieved from a storage space. The image types of the sample image include color images and / or infrared images. Abnormal objects may be objects included in the sample image that need to be prevented from approaching or boarding the ship. Optionally, the abnormal objects include abnormal ships and / or abnormal hooks. The theoretical category may be the actual category to which each object to be identified included in the sample image belongs. The theoretical bounding box information may be the actual bounding box information that each object to be identified in the sample image should correspond. The pre-built deep learning model may be a neural network model whose model parameters are default values ​​or initial values.

[0088] In this embodiment, in order to construct a large number of rich training samples, a plurality of sample images including abnormal objects can be obtained. Furthermore, each sample image can be annotated according to the actual category of the abnormal object and the position information in the sample image. Then, the theoretical category and theoretical bounding box information corresponding to the abnormal object can be obtained, and the theoretical category and theoretical bounding box information can be used as the theoretical annotation results corresponding to the abnormal object. Further, training samples can be constructed based on the sample images and their corresponding theoretical annotation results, and multiple training samples can be obtained. Further, a pre-constructed deep learning model can be trained based on multiple training samples to obtain a trained deep learning model, and the trained deep learning model can be used as an abnormal object recognition model.

[0089] Optionally, a pre-constructed deep learning model is trained based on multiple training samples to obtain an abnormal object recognition model, including: for multiple training samples, inputting sample images in the training samples into a pre-constructed deep learning model to obtain actual output results corresponding to the sample images; determining a loss value based on the actual output results and the theoretical labeling results, and correcting model parameters in the deep learning model based on the loss value, taking the convergence of the loss function in the deep learning model as a training goal, and obtaining an abnormal object recognition model.

[0090] The actual output result includes the actual category and actual bounding box information of the abnormal object. The actual category is the recognition category output after the sample image is input into the deep learning model. The actual bounding box information is the bounding box information including the abnormal object output after the sample image is input into the deep learning model. The loss value can be a numerical value representing the degree of difference between the theoretical output and the actual output. The loss function can be a function representing the degree of difference between the theoretical output and the actual output. The loss function can be any loss function, and optionally, it is a cross entropy loss function and / or a smooth L1 loss function.

[0091] As an optional implementation of this embodiment, for multiple training samples, the sample images in the training samples can be input into a pre-built deep learning model to detect and identify abnormal objects in the sample images based on the deep learning model, and output the actual output results corresponding to the sample images. Further, the actual output results can be compared with the theoretical annotation results in the training samples to determine the loss value. Further, the model parameters in the deep learning model can be corrected based on the loss value, and then the training error of the loss function in the deep learning model, that is, the loss parameter, can be used as a condition for detecting whether the current loss function has reached convergence, for example, the training error is less than the preset error or the error change trend is stable, or whether the current model iteration number is equal to the preset number of times. If the detection reaches the convergence condition, for example, the training error of the loss function is less than the preset error or the error change tends to be stable, it indicates that the deep learning model training is completed, and at this time, the iterative training can be stopped. If it is detected that the convergence condition is not reached at present, other training samples can be further obtained to train the deep learning model until the training error of the loss function is within the preset range. When the training error of the loss function reaches convergence, the trained deep learning model can be used as an abnormal object recognition model.

[0092] It should be noted that in order to improve the model performance of the abnormal object recognition model during the model application process and improve the accuracy of the target ship in processing abnormal objects, the abnormal object recognition model can also be optimized in the process of applying the abnormal object recognition model based on the target ship. Optionally, the method also includes: deploying the abnormal object recognition model in the target ship, and updating the model parameters of the abnormal object recognition model based on the image to be processed, the object recognition result and the response result.

[0093] The response result may be user feedback on the object recognition result. In general, the accuracy of the object recognition result output by the abnormal object recognition model may be determined based on the response result. In other words, when it is determined that the image to be processed includes an abnormal object based on the object recognition result, it may be determined whether the abnormal object included in the image to be processed is a real abnormal object based on the response result.

[0094] As an optional implementation of this embodiment, when an abnormal object recognition model is obtained, the abnormal object recognition model can be deployed in the target ship. Further, when an image to be processed is obtained based on the target ship, the image to be processed can be processed based on the abnormal object recognition model, the object recognition result corresponding to the image to be processed can be determined, and the response result of the user to the object recognition result feedback can be received. Further, the model parameters of the abnormal object recognition model can be corrected according to the image to be processed, the object recognition result and the response result. Thus, the model performance can be continuously optimized in the model application process of the abnormal object recognition model.

[0095] S320: Acquire an image to be processed including an object to be identified.

[0096] S330, processing the image to be processed based on the abnormal object recognition model obtained by pre-training to obtain an object recognition result corresponding to the object to be recognized; wherein the object recognition result includes the recognition category and bounding box information corresponding to the object to be recognized.

[0097] S340: When the object to be identified is determined to be an abnormal object according to the identification category, determine a target processing device corresponding to the abnormal object and a target device position of the target processing device according to the identification category and the bounding box information.

[0098] S350: Setting a target processing device according to the location of the target device, and processing the abnormal object based on the set target processing device.

[0099] The technical solution of the embodiment of the present invention obtains an abnormal object recognition model through training, thereby achieving the effect of training an abnormal object recognition model that can quickly and accurately recognize abnormal objects based on constructed sample data, effectively improving the accuracy and robustness of the model, so that the trained abnormal object recognition model can support more complex abnormal object recognition tasks, effectively improving the model performance.

[0100] Embodiment 4

[0101] Figure 5 Schematic diagram of the structure of an abnormal object processing device provided by the fourth embodiment of the present invention. Figure 5 As shown, the device includes: an image acquisition module 410, an object recognition module 420, a processing device determination module 430 and an object processing module 440.

[0102] Among them, the image acquisition module 410 is used to acquire an image to be processed including an object to be identified; the object identification module 420 is used to process the image to be processed based on a pre-trained abnormal object identification model to obtain an object identification result corresponding to the object to be identified; wherein the abnormal object identification model is used to identify abnormal objects associated with the target ship; the abnormal object is an object that needs to be prevented from approaching or boarding the target ship; the object identification result includes the identification category and bounding box information corresponding to the object to be identified; the processing device determination module 430 is used to determine the target processing device and the target device position corresponding to the abnormal object according to the identification category and the bounding box information corresponding to the abnormal object when it is determined that the image to be processed includes an abnormal object based on the identification category; wherein the target processing device is used to process the abnormal object so that the abnormal object is away from the target ship; the object processing module 440 is used to set the target processing device according to the target device position, and process the abnormal object based on the set target processing device.

[0103] The technical solution of the embodiment of the present invention obtains an image to be processed including an object to be identified, and further processes the image to be processed based on a pre-trained abnormal object recognition model to obtain an object recognition result corresponding to the object to be identified; wherein the object recognition result includes a recognition category and bounding box information corresponding to the object to be identified, and further, when it is determined based on the recognition category that the image to be processed includes an abnormal object, the target processing device and the target device position corresponding to the abnormal object are determined according to the recognition category and bounding box information corresponding to the abnormal object, and further, the target processing device is set according to the target device position, and the corresponding abnormal object is processed based on the set target processing device, so as to solve the problem that the abnormal object processing method in the related art is time-consuming and labor-intensive, the object recognition accuracy is poor, and it may even affect the personal safety of the crew, and realizes the effect of quickly identifying abnormal objects associated with the ship based on the neural network model, improves the recognition efficiency of abnormal objects, and further improves the processing efficiency of abnormal objects, and realizes the effect of effectively processing abnormal objects without human intervention, and further improves the navigation safety of the ship.

[0104] Optionally, the processing device determination module 430 includes: a device category determination unit and a processing device determination unit.

[0105] a device category determination unit, configured to determine a target device category corresponding to the abnormal object according to the identification category corresponding to the abnormal object and preset object device correspondence information; wherein the object device correspondence information is used to indicate a correspondence between the identification category of the abnormal object and the device category corresponding to the object processing device for processing the abnormal object;

[0106] A processing device determination unit is used to determine a target processing device according to the bounding box information corresponding to the abnormal object and the target device category, and to determine a target device position according to the bounding box information corresponding to the abnormal object and a device association parameter corresponding to the target processing device.

[0107] Optionally, the processing device determination unit includes: a device position acquisition subunit, a distance determination subunit and a processing device determination subunit.

[0108] a device position acquisition subunit, configured to acquire a current device position of at least one candidate processing device belonging to the target device category and pre-set on the target ship;

[0109] a distance determination subunit, configured to determine, for at least one of the candidate processing devices, a distance between the candidate processing device and the abnormal object according to the current device position and the bounding box information;

[0110] The processing device determination subunit is used to determine a minimum distance from at least one of the distances, take the minimum distance as a target distance, and take the candidate processing device corresponding to the target distance as a target processing device.

[0111] Optionally, the processing device determination unit includes a device position determination subunit.

[0112] The device position determination subunit is used to determine the target device position according to the current device position of the target processing device, the device movement range and the bounding box information corresponding to the abnormal object.

[0113] Optionally, the object processing module 440 includes: a device moving unit.

[0114] The device moving unit is used to move the target processing device from the current device position to the target device position.

[0115] Optionally, after determining that the object to be identified is an abnormal object according to the identification category, the device further includes: an image annotation module, a training sample set updating module and a model updating module.

[0116] Optionally, the device also includes: a model training module.

[0117] A model training module is used to train an abnormal object recognition model;

[0118] The model training module includes: a training sample acquisition unit, a result output unit and a model parameter correction unit.

[0119] A training sample acquisition unit, configured to acquire a plurality of training samples, wherein the training samples include: a sample image, wherein the sample image includes an abnormal object, and a theoretical annotation result corresponding to the abnormal object, wherein the theoretical annotation result includes theoretical category and theoretical bounding box information;

[0120] A result output unit, for inputting the sample images in the training samples into a pre-built deep learning model for the plurality of training samples, to obtain actual output results corresponding to the sample images;

[0121] A model parameter correction unit is used to determine a loss value according to the actual output result and the theoretical labeling result, and to correct the model parameters in the deep learning model based on the loss value, and to take the convergence of the loss function in the deep learning model as a training target to obtain an abnormal object recognition model.

[0122] Optionally, the device also includes: a model parameter updating module.

[0123] A model parameter updating module is used to deploy the abnormal object recognition model in the target ship and update the model parameters of the abnormal object recognition model based on the image to be processed, the object recognition result and the response result.

[0124] Optionally, the image types of the to-be-processed image and the sample image include color images and / or infrared images.

[0125] The abnormal object processing device provided in the embodiment of the present invention can execute the abnormal object processing method provided in any embodiment of the present invention, and has the corresponding functional modules and beneficial effects of the execution method.

[0126] Embodiment 5

[0127] Figure 6 A schematic diagram of the structure of an electronic device 10 that can be used to implement an embodiment of the present invention is shown. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device can also represent various forms of mobile devices, such as personal digital processing, cellular phones, smart phones, wearable devices (such as helmets, glasses, watches, etc.) and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely examples and are not intended to limit the implementation of the present invention described and / or required herein.

[0128] like Figure 6 As shown, the electronic device 10 includes at least one processor 11, and a memory connected to the at least one processor 11, such as a read-only memory (ROM) 12, a random access memory (RAM) 13, etc., wherein the memory stores a computer program that can be executed by at least one processor, and the processor 11 can perform various appropriate actions and processes according to the computer program stored in the read-only memory (ROM) 12 or the computer program loaded from the storage unit 18 to the random access memory (RAM) 13. In the RAM 13, various programs and data required for the operation of the electronic device 10 can also be stored. The processor 11, the ROM 12, and the RAM 13 are connected to each other through a bus 14. An input / output (I / O) interface 15 is also connected to the bus 14.

[0129] A number of components in the electronic device 10 are connected to the I / O interface 15, including: an input unit 16, such as a keyboard, a mouse, etc.; an output unit 17, such as various types of displays, speakers, etc.; a storage unit 18, such as a disk, an optical disk, etc.; and a communication unit 19, such as a network card, a modem, a wireless communication transceiver, etc. The communication unit 19 allows the electronic device 10 to exchange information / data with other devices through a computer network such as the Internet and / or various telecommunication networks.

[0130] The processor 11 may be a variety of general and / or special processing components with processing and computing capabilities. Some examples of the processor 11 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various special artificial intelligence (AI) computing chips, various processors running machine learning model algorithms, a digital signal processor (DSP), and any appropriate processor, controller, microcontroller, etc. The processor 11 executes the various methods and processes described above, such as an abnormal object processing method.

[0131] In some embodiments, the abnormal object processing method can be implemented as a computer program, which is tangibly contained in a computer-readable storage medium, such as a storage unit 18. In some embodiments, part or all of the computer program can be loaded and / or installed on the electronic device 10 via the ROM 12 and / or the communication unit 19. When the computer program is loaded into the RAM 13 and executed by the processor 11, one or more steps of the abnormal object processing method described above can be performed. Alternatively, in other embodiments, the processor 11 can be configured to execute the abnormal object processing method in any other appropriate manner (for example, by means of firmware).

[0132] Various implementations of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), systems on chips (SOCs), load programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various implementations can include: being implemented in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which can be a special purpose or general purpose programmable processor that can receive data and instructions from a storage system, at least one input device, and at least one output device, and transmit data and instructions to the storage system, the at least one input device, and the at least one output device.

[0133] Computer programs for implementing the methods of the present invention may be written in any combination of one or more programming languages. These computer programs may be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing device, so that when the computer program is executed by the processor, the functions / operations specified in the flow chart and / or block diagram are implemented. The computer program may be executed entirely on the machine, partially on the machine, partially on the machine and partially on a remote machine as a stand-alone software package, or entirely on a remote machine or server.

[0134] In the context of the present invention, a computer-readable storage medium may be a tangible medium that may contain or store a computer program for use by or in combination with an instruction execution system, device or equipment. A computer-readable storage medium may include, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, devices or equipment, or any suitable combination of the foregoing. Alternatively, a computer-readable storage medium may be a machine-readable signal medium. A more specific example of a machine-readable storage medium may include an electrical connection based on one or more lines, 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), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0135] To provide interaction with a user, the systems and techniques described herein may be implemented on an electronic device having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and a pointing device (e.g., a mouse or trackball) through which the user can provide input to the electronic device. Other types of devices may also be used to provide interaction with the user; for example, the feedback provided to the user may be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user may be received in any form (including acoustic input, voice input, or tactile input).

[0136] The systems and techniques described herein may be implemented in a computing system that includes backend components (e.g., as a data server), or a computing system that includes middleware components (e.g., an application server), or a computing system that includes frontend components (e.g., a user computer with a graphical user interface or a web browser through which a user can interact with implementations of the systems and techniques described herein), or a computing system that includes any combination of such backend components, middleware components, or frontend components. The components of the system may be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include: a local area network (LAN), a wide area network (WAN), a blockchain network, and the Internet.

[0137] A computing system may include a client and a server. The client and the server are generally remote from each other and usually interact through a communication network. The client and server relationship is generated by computer programs running on the corresponding computers and having a client-server relationship with each other. The server may be a cloud server, also known as a cloud computing server or cloud host, which is a host product in the cloud computing service system to solve the defects of difficult management and weak business scalability in traditional physical hosts and VPS services.

[0138] It should be understood that the various forms of processes shown above can be used to reorder, add or delete steps. For example, the steps described in the present invention can be executed in parallel, sequentially or in different orders, as long as the desired results of the technical solution of the present invention can be achieved, and this document does not limit this.

[0139] The above specific implementations do not constitute a limitation on the protection scope of the present invention. It should be understood by those skilled in the art that various modifications, combinations, sub-combinations and substitutions can be made according to design requirements and other factors. Any modification, equivalent substitution and improvement made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for handling abnormal objects, characterized in that: Applied to target ships, including: Get the image to be processed that includes the object to be recognized: The image to be processed is processed based on a pre-trained abnormal object recognition model to obtain an object recognition result corresponding to the object to be recognized; wherein the abnormal object recognition model is used to recognize abnormal objects associated with the target ship; the abnormal objects are objects that need to be prevented from approaching or boarding the target ship; the object recognition result includes recognition categories and bounding box information corresponding to the object to be recognized; In the case where it is determined based on the recognition category that the image to be processed includes an abnormal object, a target processing device and a target device position corresponding to the abnormal object are determined according to the recognition category and the bounding box information corresponding to the abnormal object; wherein the target processing device is used to process the abnormal object so that the abnormal object is away from the target ship; The target processing device is set according to the position of the target device, and the corresponding abnormal object is processed based on the set target processing device.

2. The abnormal object processing method according to claim 1, characterized in that: The step of determining a target processing device and a target device position corresponding to the abnormal object according to the identification category and the bounding box information corresponding to the abnormal object comprises: Determine the target device category corresponding to the abnormal object according to the identification category corresponding to the abnormal object and preset object device corresponding information; wherein the object device corresponding information is used to indicate the corresponding relationship between the identification category of the abnormal object and the device category corresponding to the object processing device for processing the abnormal object; A target processing device is determined according to the bounding box information corresponding to the abnormal object and the target device category, and a target device position is determined according to the bounding box information corresponding to the abnormal object and a device association parameter corresponding to the target processing device.

3. The abnormal object processing method according to claim 2, characterized in that: The determining the target processing device according to the bounding box information corresponding to the abnormal object and the target device category includes: Acquire a current device position of at least one candidate processing device belonging to the target device category and pre-installed on the target ship; For at least one of the candidate processing devices, determining a distance between the candidate processing device and the abnormal object according to the current device position and the bounding box information; A minimum distance is determined from at least one of the distances, the minimum distance is used as a target distance, and the candidate processing device corresponding to the target distance is used as a target processing device.

4. The abnormal object processing method according to claim 3 is characterized in that: The device-associated parameter includes a movable range of the device; and determining the target device position according to the bounding box information corresponding to the abnormal object and the device-associated parameter corresponding to the target processing device, comprises: The target device position is determined according to the current device position of the target processing device, the movable range of the device, and the bounding box information corresponding to the abnormal object.

5. The abnormal object processing method according to claim 3, characterized in that: The step of setting the target processing device according to the position of the target device comprises: The target processing device is moved from the current device location to the target device location.

6. The abnormal object processing method according to claim 1, characterized in that: Also includes: Train and obtain an abnormal object recognition model; The training obtains an abnormal object recognition model, including: Acquire multiple training samples, wherein the training sample set includes multiple training samples, and the training samples include: a sample image, the sample image includes an abnormal object, and a theoretical annotation result corresponding to the sample image, the theoretical annotation result includes theoretical category and theoretical bounding box information; For the plurality of training samples, the sample images in the training samples are input into a pre-built deep learning model to obtain actual output results corresponding to the sample images; A loss value is determined according to the actual output result and the theoretical annotation result, and the model parameters in the deep learning model are corrected based on the loss value. The convergence of the loss function in the deep learning model is used as a training goal to obtain an abnormal object recognition model.

7. The abnormal object processing method according to claim 6, characterized in that: Also includes: The abnormal object recognition model is deployed in the target ship, and model parameters of the abnormal object recognition model are updated based on the image to be processed, the object recognition result and the response result.

8. The abnormal object processing method according to claim 1, characterized in that: The image types of the images to be processed include color images and / or infrared images.

9. An abnormal object processing device, characterized in that: Configured in the target ship, including: An image acquisition module is used to acquire an image to be processed including an object to be identified: An object recognition module is used to process the image to be processed based on a pre-trained abnormal object recognition model to obtain an object recognition result corresponding to the object to be recognized; wherein the abnormal object recognition model is used to recognize abnormal objects associated with the target ship; the abnormal objects are objects that need to be prevented from approaching or boarding the target ship; and the object recognition result includes recognition categories and bounding box information corresponding to the object to be recognized; a processing device determination module, configured to determine a target processing device and a target device position corresponding to the abnormal object according to the identification category and the bounding box information corresponding to the abnormal object, when it is determined that the image to be processed includes an abnormal object based on the identification category; wherein the target processing device is configured to process the abnormal object so that the abnormal object is away from the target ship; The object processing module is used to set the target processing device according to the position of the target device, and process the corresponding abnormal object based on the set target processing device.

10. An electronic device, characterized in that: The electronic device comprises: at least one processor; and a memory communicatively connected to the at least one processor; wherein, The memory stores a computer program executable by the at least one processor, and the computer program is executed by the at least one processor so that the at least one processor can execute the abnormal object processing method according to any one of claims 1 to 8.

11. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer instructions, and the computer instructions are used to enable a processor to implement the abnormal object processing method according to any one of claims 1 to 8 when executed.

12. A computer program product, characterized in that The computer program product comprises a computer program, which, when executed by a processor, implements the abnormal object processing method according to claims 1-8.

Images