A method and device for detecting the shielding of a monitoring camera of a flood control dam of an open-pit mine

Abstract

The application provides a kind of open-pit mine flood control dam monitoring camera shielding detection method and device, the method includes: obtaining water surface image under different environment and water level conditions, identifying the scene of water surface image, and under respective scene, water surface image is carried out binaryzation processing and augmentation processing, generate water surface image set;For the first image corresponding to the first scene in water surface image set, the first image is subjected to water surface semantic segmentation processing, to generate binary segmentation image;Construct two-class neural network, carry out preliminary detection to whether monitoring camera is shielded, obtain first shielding identification result;According to the first confidence of scene identification, the second confidence of water surface semantic segmentation processing and the third confidence of shielding detection, determine whether monitoring camera is shielded.The application can accurately identify whether the camera is shielded automatically, reduce cost and workload.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of water surface detection, and particularly relates to a method for detecting shielding of a monitoring camera of a flood control dam of an open-pit mine. BACKGROUND

[0002] The feature of China's energy structure is to take coal as the main energy source for a long period of time. Open-pit mining of coal is an important mining method, and the proportion of open-pit mining of coal gradually increases due to the advantages of fast construction speed, low production cost, high labor efficiency and the like. Since the open-pit mine is generally located in a low-lying gully area, rainwater is easy to gather, so a flood control dam needs to be arranged around the open-pit mine to achieve the purpose of flood control and ensure the safety of open-pit mining of coal. After the flood control dam is arranged, the water level of the flood control dam needs to be detected in real time to timely find an abnormal situation that the water level exceeds the warning line.

[0003] In the prior art, the water level of the flood control dam is often observed by a camera to determine whether the water level of the flood control dam exceeds the limit, or the water level of the flood control dam is determined whether to exceed the limit according to an image captured by the camera and a method of deep learning.

[0004] However, since the production dust and water vapor of the mining area are large, the camera may be shielded, and thus the image captured by the camera greatly influences the recognition result. The traditional method is to regularly arrange staff to inspect the camera, which undoubtedly increases the workload. In addition, when shielding may occur at the edge of the camera, it is difficult to recognize. SUMMARY

[0005] In view of the above problems, a method and device for detecting shielding of a monitoring camera of a flood control dam of an open-pit mine are provided to detect whether the camera is shielded in the process of water surface detection.

[0006] The first aspect of the present application provides a method for detecting shielding of a monitoring camera of a flood control dam of an open-pit mine, comprising:

[0007] According to the water surface images of the monitoring camera under different environments and water level conditions, the scenes of the water surface images are recognized, and the water surface images are subjected to binarization processing and augmentation processing under respective scenes to generate a water surface image set;

[0008] For a first image corresponding to a first scene in the water surface image set, the first image is subjected to water surface semantic segmentation processing to generate a binary segmentation image;

[0009] A binary classification neural network is constructed to preliminarily detect whether the monitoring camera is shielded to obtain a first shielding recognition result;

[0010] According to the first confidence of scene recognition, the second confidence of water surface semantic segmentation processing, the third confidence of occlusion detection and the first occlusion recognition result, it is determined whether the monitoring camera is occluded.

[0011] Optionally, the scene of the water surface image is identified, including:

[0012] A candidate scene classification neural network is constructed, based on a deep learning method, using a transfer learning method, and the candidate scene classification neural network is trained using a multi-weather dataset (MWD). After training, a scene classification neural network is obtained.

[0013] According to the scene classification neural network, the scene of the water surface image is identified.

[0014] Optionally, the water surface image is binarized, including:

[0015] In any scene, the water surface region and the dam body region of the water surface image are segmented and labeled to generate a binarized image.

[0016] Optionally, according to the first confidence of scene recognition, the second confidence of water surface semantic segmentation processing, the third confidence of occlusion detection and the first occlusion recognition result, it is determined whether the monitoring camera is occluded, including:

[0017] The first confidence from scene recognition and the second confidence from water surface semantic segmentation are weighted and added to obtain a fourth confidence, expressed as:

[0018] C4 = λ1C cla + λ2C seg ,

[0019] Wherein, C4 is the fourth confidence, C cla is the first confidence, C seg is the second confidence, λ1 and λ2 are determined by experience, and their sum is 1.

[0020] When the first occlusion recognition result is occlusion, the first occlusion recognition result is taken as 1, otherwise as 0. The product of the assignment of the first occlusion recognition result and the third confidence is taken as a fifth confidence, expressed as:

[0021] C5 = R·C wea ,

[0022] Wherein, C5 is the fifth confidence, C wea is the third confidence, and R is the value of the first occlusion recognition result.

[0023] According to the fourth confidence and the fifth confidence, a total confidence is obtained, which is formulated as:

[0024] C=C4+C5,

[0025] The total confidence is compared with a preset threshold to determine whether the monitoring camera is blocked.

[0026] Optionally, the comparison of the total confidence with the preset threshold to determine whether the monitoring camera is blocked comprises:

[0027] If the total confidence is greater than the preset threshold, it is determined that the monitoring camera is blocked.

[0028] If the total confidence is not greater than the preset threshold, it is determined that the monitoring camera is not blocked.

[0029] Optionally, the method further comprises:

[0030] According to the binary segmented image, it is determined whether the water surface is abnormal.

[0031] Optionally, the determination of whether the water surface is abnormal according to the binary segmented image comprises:

[0032] According to a clustering algorithm, points in the binary segmented image are divided into several classes, and classes not belonging to the water surface are removed.

[0033] The maximum range of the flood control dam water area is determined, and the number of pixel points belonging to the water class in the binary segmented image that falls outside the maximum range is calculated.

[0034] If the number of pixel points belonging to the water class in the binary segmented image that falls outside the maximum range exceeds the preset pixel point threshold, the water level corresponding to the binary segmented image is out of limit.

[0035] The second aspect of the application proposes a device for detecting the blocking of a monitoring camera of a flood control dam of an open-pit mine, characterized in that it comprises:

[0036] An acquisition module is configured to acquire water surface images under different environmental and water level conditions according to a monitoring camera, identify scenes of the water surface images, and perform binaryzation processing and augmentation processing on the water surface images under respective scenes to generate a set of water surface images.

[0037] A semantic segmentation module is configured to perform water surface semantic segmentation processing on a first image corresponding to a first scene in the set of water surface images to generate a binary segmented image.

[0038] A first blocking detection module is configured to construct a binary classification neural network, perform primary detection on whether the monitoring camera is blocked, and obtain a first blocking identification result.

[0039] The second occlusion detection module is configured to determine whether the monitoring camera is occluded according to the first confidence of scene recognition, the second confidence of water surface semantic segmentation processing, and the third confidence of occlusion detection.

[0040] The third aspect of the present application provides a computer device, comprising a memory, a processor and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the method of any one of the first aspect.

[0041] The fourth aspect of the present application provides a non-transitory computer readable storage medium, which stores a computer program, wherein the computer program is executed by a processor to implement the method of any one of the first aspect.

[0042] The technical scheme provided by the embodiments of the present application at least brings the following beneficial effects:

[0043] When detecting the occlusion of the camera, the confidence discrimination factor is introduced, and each processing process is weighted, so that the occlusion of the camera can be detected and recognized more accurately, and the cost is reduced.

[0044] Additional aspects and advantages of the present application will be in part apparent and in part pointed out hereinafter. BRIEF DESCRIPTION OF DRAWINGS

[0045] The above and / or additional aspects and advantages of the present application will become apparent and be readily appreciated from the following description, taken in conjunction with the accompanying drawings, in which:

[0046] Figure 1 is a flow chart of a method for detecting occlusion of a monitoring camera of a flood control dam of an open-pit mine according to an embodiment of the present application;

[0047] Figure 2 is a block diagram of an occlusion detection device for a monitoring camera of a flood control dam of an open-pit mine according to an embodiment of the present application:

[0048] Figure 3 is a block diagram of an electronic device. DETAILED DESCRIPTION

[0049] The embodiments of the present application are described in detail below, and examples of the embodiments are shown in the accompanying drawings, wherein the same or similar reference signs represent the same or similar elements or elements having the same or similar functions throughout. The embodiments described below by referring to the accompanying drawings are exemplary and are intended to explain the present application, and cannot be understood as limiting the present application.

[0050] Figure 1is a flowchart of a method for detecting occlusion of a monitoring camera of a flood control dam of an open-pit mine according to an embodiment of the present application, and the method comprises the following steps:

[0051] In step 101, water surface images under different environmental and water level conditions are acquired by the monitoring camera, the scenes of the water surface images are identified, and the water surface images are subjected to binarization processing and augmentation processing under respective scenes to generate a water surface image set.

[0052] In the embodiment of the present application, the monitoring camera is arranged with multiple IMU sensors at the water bank, the shooting angle and shooting position of the monitoring camera are adjusted until the monitoring area can completely cover the dam body of the entire flood control dam, and water surface images under different environmental and water level conditions are acquired by the monitoring camera.

[0053] Subsequently, the scene in which the water surface image is located is identified.

[0054] In a possible embodiment, a VGG-16 is used as a basic network architecture, a SE module incorporating a compressed excitation is used as a channel attention mechanism to construct a candidate scene classification neural network, and based on a deep learning method, a transfer learning manner is used to train the candidate scene classification neural network by using a multi-weather dataset MWD, and after the training is completed, a scene classification neural network is obtained.

[0055] It should be noted that the multi-weather dataset MWD contains 65,000 images of 6 common categories, including sunny, cloudy, rainy, snowy, misty, and thunderstorm weather, and the trained scene classification neural network can identify the scene in which the water surface image is located. The purpose of this is that when the scenes are different, the contents of the water surface images are greatly different, and when subsequent binarization and semantic segmentation are performed, if the same processing is performed on each scene, it may have a great impact on the subsequent water surface identification process.

[0056] In the embodiment of the present application, the water surface region and the dam body region of the water surface image are segmented and labeled in any scene to generate a binarized image.

[0057] It should be noted that since the problem of whether the water surface is out of limit is observed, the camera can meet the requirement that the water surface and the dam body are contained in the captured image when the camera is adjusted.

[0058] In a possible embodiment, the water surface region is labeled as red and the dam body region is labeled as black for region differentiation.

[0059] In addition, the water surface image and the binarized image are subjected to augmentation processing.

[0060] It should be noted that the conventional augmentation processing includes rotation, cutting, multi-cutting map splicing, adjusting contrast and adjusting brightness, and one or more of the above augmentation processing is used in the present application to increase the data set capacity, and the specific processing is determined according to the actual image.

[0061] In step 102, for a first image corresponding to a first scene in the water surface image set, the first image is subjected to water surface semantic segmentation processing to generate a binary segmentation image.

[0062] In a possible embodiment, the first image corresponding to the first scene is a binary image corresponding to lightning weather, and the water surface semantic segmentation network trained by using the lightning weather data set is used to segment the binary image.

[0063] The semantic segmentation method is a typical computer vision problem, which involves taking some raw data as input and converting them into masks with highlighted regions of interest. In the embodiments of the present application, the purpose of the water surface semantic segmentation processing of the first image is to more significantly distinguish the dam body from the water surface.

[0064] It should be noted that the present application can determine whether the water surface is out of limit according to the binary segmentation image, and specifically:

[0065] According to the clustering algorithm, the points in the binary segmentation image are divided into several classes, and the classes not belonging to the water surface are removed;

[0066] The maximum range of the flood control dam water area is determined, and the number of pixel points belonging to the water class in the binary segmentation image falling outside the maximum range is calculated;

[0067] If the number of pixel points belonging to the water class in the binary segmentation image falling outside the maximum range exceeds the preset pixel point threshold, the water level corresponding to the binary segmentation image is out of limit.

[0068] In a possible embodiment, the clustering algorithm is a clustering algorithm based on a GCN graph neural network.

[0069] In step 103, a binary classification neural network is constructed to perform primary detection on whether the monitoring camera is blocked to obtain a first blocking identification result.

[0070] In the embodiments of the present application, the binary classification neural network uses a neural network to preliminarily identify the water surface image captured by the monitoring camera. When it is detected that the water surface image is incomplete, it is preliminarily judged that the monitoring camera is blocked. The angle of judgment here is from the completeness of the image.

[0071] In step 104, according to the first confidence of scene recognition, the second confidence of water surface semantic segmentation processing, the third confidence of blocking detection and the first blocking identification result, it is determined whether the monitoring camera is blocked.

[0072] In the embodiments of the present application, since the occlusion can occur at the edge of the camera, the binary classification neural network may fail to make a correct judgment. Therefore, a confidence factor is introduced to re-judge the first occlusion recognition result.

[0073] Specifically, the first confidence from the scene recognition and the second confidence from the water surface semantic segmentation are weighted and added to obtain a fourth confidence, and the expression is as follows:

[0074] C4=λ1C cla +λ2C seg ,

[0075] wherein C4 is the fourth confidence, C cla is the first confidence, C seg is the second confidence, and λ1 and λ2 are determined by experience and their sum is 1.

[0076] When the first occlusion recognition result is occlusion, the first occlusion recognition result is taken as 1, otherwise as 0. The product of the value of the first occlusion recognition result and the third confidence is taken as a fifth confidence, and the expression is as follows:

[0077] C5=R·C wea ,

[0078] wherein C5 is the fifth confidence, C wea is the third confidence, and R is the value of the first occlusion recognition result.

[0079] The total confidence is obtained according to the fourth confidence and the fifth confidence, and the formula is as follows:

[0080] C=C4+C5,

[0081] The total confidence is compared with a preset threshold to determine whether the monitoring camera is occluded.

[0082] It should be noted that the first confidence, the second confidence and the third confidence are known according to the prediction accuracy of the model in the model training process.

[0083] In the embodiments of the present application, if the total confidence is greater than the preset threshold, it is determined that the monitoring camera is occluded.

[0084] If the total confidence is not greater than the preset threshold, it is determined that the monitoring camera is not occluded.

[0085] It should be noted that the preset threshold is not fixed, but is determined according to the confidence of the model in the actual situation.

[0086] In a possible embodiment, the preset threshold is 1.5, that is, if the total confidence is greater than 1.5, it is determined that the monitoring camera is finally occluded.

[0087] The confidence degree discrimination factor is introduced, and each processing process is weighted, so that the camera occlusion can be more accurately detected and identified, and the cost is reduced.

[0088] Figure 2 is a block diagram of an open-pit mine flood control dam monitoring camera occlusion detection device according to an embodiment of the present application, comprising:

[0089] The acquisition module 210 is configured to acquire water surface images under different environment and water level conditions according to the monitoring camera, identify the scenes of the water surface images, and perform binaryzation processing and augmentation processing on the water surface images under the respective scenes to generate a water surface image set.

[0090] The semantic segmentation module 220 is configured to perform water surface semantic segmentation processing on a first image corresponding to a first scene in the water surface image set to generate a binary segmentation image.

[0091] The first occlusion detection module 230 is configured to construct a binary classification neural network to preliminarily detect whether the monitoring camera is occluded to obtain a first occlusion identification result.

[0092] The second occlusion detection module 240 is configured to determine whether the monitoring camera is occluded according to a first confidence degree of scene recognition, a second confidence degree of water surface semantic segmentation processing, and a third confidence degree of occlusion detection.

[0093] As to the device in the above embodiment, the specific manner in which the various modules perform operations has been described in detail in the embodiment relating to the method, and will not be described in detail here.

[0094] Figure 3 A schematic block diagram of an example electronic device 700 that can be used to implement embodiments of the present disclosure is shown. The electronic device is intended to represent various forms of digital computers, such as laptops, desktops, tablets, personal digital assistants, servers, blade servers, mainframes, and other appropriate computers. The electronic device can also represent various forms of mobile devices, such as personal digital assistants, cellular telephones, smartphones, wearable devices, and other similar computing devices. The components shown here, their connections and relationships, and their functions, are meant to be examples only, and are not meant to limit implementations of the present disclosure described and / or claimed in this document.

[0095] As Figure 3As shown, the device 700 includes a computing unit 701 that can perform various appropriate actions and processes in accordance with a computer program stored in a read-only memory (ROM) 702 or a computer program loaded from a storage unit 708 into a random access memory (RAM) 703. Various programs and data required for the operation of the device 700 can also be stored in the RAM 703. The computing unit 701, the ROM 702, and the RAM 703 are connected to each other through a bus 704. An input / output (I / O) interface 705 is also connected to the bus 704.

[0096] A plurality of components in the device 700 are connected to the I / O interface 705, including an input unit 706, such as a keyboard, a mouse, etc., an output unit 707, such as various types of displays, speakers, etc., a storage unit 708, such as a magnetic disk, an optical disk, etc., and a communication unit 709, such as a network card, a modem, a wireless communication transceiver, etc. The communication unit 709 allows the device 700 to exchange information / data with other devices through a computer network, such as the Internet, and / or various telecommunication networks.

[0097] The computing unit 701 can be various general and / or special purpose processing components having processing and computing capabilities. Some examples of the computing unit 701 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various special-purpose artificial intelligence (AI) computing chips, various computing units running machine learning model algorithms, a digital signal processor (DSP), and any appropriate processor, controller, microcontroller, etc. The computing unit 701 performs various methods and processes described above, such as the voice instruction response method. For example, in some embodiments, the voice instruction response method can be implemented as a computer software program that is tangibly embodied in a machine-readable medium, such as the storage unit 708. In some embodiments, part or all of the computer program can be loaded and / or installed on the device 700 via the ROM 702 and / or the communication unit 709. When the computer program is loaded into the RAM 703 and executed by the computing unit 701, one or more steps of the voice instruction response method described above can be performed. Alternatively, in other embodiments, the computing unit 701 can be configured to perform the voice instruction response method by any other appropriate means, such as by means of firmware.

[0098] The various embodiments of the systems and techniques described above can be implemented in digital electronic circuitry, integrated circuitry, a field programmable gate array (FPGA), an application specific integrated circuit (ASIC), a system on a chip (SOC), a load programmable logic device (CPLD), computer hardware, firmware, software, and / or combinations thereof. These various embodiments can include implementation in one or more computer programs that are executable and / or interpretable on a programmable system including at least one programmable processor, which can be special or general purpose, coupled to receive data and instructions from, and to transmit data and instructions to, a storage system, at least one input device, and at least one output device.

[0099] Program code for carrying out methods of the present disclosure can be written in any combination of one or more programming languages. The program code can be provided to a processor or controller of a general purpose computer, special purpose computer, or other programmable data processing apparatus to produce a machine, such that the program code, when executed by the processor or controller, produces the functions / operations specified in the flowcharts and / or the block diagrams. The program code can be executed entirely on a machine, partially on a machine, partially on a machine as a stand-alone software package, partially on a machine and partially on a remote machine or entirely on a remote machine or server.

[0100] In the context of the present disclosure, a machine-readable medium can be a tangible medium that contains or stores a program for use by or in connection with an instruction execution system, apparatus, or device. The machine-readable medium can be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium can include but is not limited to an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any suitable combination of the foregoing. More specific examples of the machine-readable storage medium will include one or more lines of electrical connections, portable computer disks, hard disk drives, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or Flash memory), optical fibers, portable compact disc read-only memories (CD-ROMs), optical storage devices, magnetic storage devices, or any suitable combination of the foregoing.

[0101] To provide for interaction with a user, the systems and techniques described here can be implemented on a computer 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 a trackball) by which the user can provide input to the computer. Other kinds of devices can be used to provide for interaction with a user as well; for example, feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form, including acoustic, speech, or tactile input.

[0102] The systems and techniques described here can be implemented in a computing system that includes a back end component (e.g., as a data server), or that includes a middleware component (e.g., an application server), or that includes a front end component (e.g., a user computer having a graphical user interface or a Web browser through which a user can interact with an implementation of the systems and techniques described here), or any combination of such back end, middleware, or front end components. The components of the system can 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), the Internet, and a blockchain network.

[0103] The computer system can include clients and servers. A client and server are generally remote from each other and typically interact through a communication network. The relationship of client and server arises by virtue of computer programs running on the respective computers and having a client-server relationship to each other. A server can be a cloud server, also known as cloud computing server or cloud host, which is a host product in the cloud computing service system, to solve the defects of large management difficulty and weak business scalability in traditional physical host and VPS (Virtual Private Server, or VPS for short) services. The server can also be a server of a distributed system, or a server combined with a blockchain.

[0104] It should be understood that various forms of flow shown above can be used with reordering, adding or deleting steps. For example, the steps recited in the present disclosure can be performed in parallel, in series, or in different orders, as long as the desired results of the technical solutions of the present disclosure can be achieved, which are not limited herein.

[0105] The above detailed description does not limit the scope of the disclosure. Various modifications, combinations, sub-combinations and alternatives can be made to the detailed description. Any modification, equivalent replacement and improvement etc. made within the spirit and principle of the disclosure shall be included in the scope of the disclosure.

Claims

1. A method for detecting occlusion of a monitoring camera at an open-pit mine flood control dam, characterized in that, include: Based on water surface images acquired by surveillance cameras under different environmental and water level conditions, the scene of the water surface image is identified, and binarization and augmentation processing are performed on the water surface image under each scene to generate a water surface image set. For the first image corresponding to the first scene in the water surface image set, perform water surface semantic segmentation processing on the first image to generate a binary segmentation image; A binary classification neural network is constructed to perform a preliminary detection of whether the surveillance camera is obstructed, and a first occlusion recognition result is obtained. Based on the first confidence level of scene recognition, the second confidence level of water surface semantic segmentation, the third confidence level of occlusion detection, and the first occlusion recognition result, it is determined whether the surveillance camera is occluded, specifically including: The first confidence score from scene recognition and the second confidence score from water surface semantic segmentation are weighted and summed to obtain the fourth confidence score, expressed as follows: ， in, The fourth confidence level This represents the first confidence level. This is the second confidence level. , Based on experience, their sum is 1; When the first occlusion recognition result indicates occlusion, the first occlusion recognition result is set to 1; otherwise, it is set to 0. The product of the first occlusion recognition result and the third confidence level is used as the fifth confidence level, expressed as: ， in, For the fifth confidence level, The third confidence level, The value of the first occlusion recognition result; Based on the fourth and fifth confidence levels, the total confidence level is obtained, which is formulated as follows: = ， The total confidence level is compared with a preset threshold to determine whether the surveillance camera is obstructed.

2. The method according to claim 1, characterized in that, The scenario for recognizing the water surface image includes: A candidate scene classification neural network is constructed. Based on deep learning, the transfer learning method is used to train the candidate scene classification neural network using the MWD multi-weather dataset. After training, the scene classification neural network is obtained. The scene classification neural network identifies the scene of the water surface image.

3. The method according to claim 1, characterized in that, The binarization process of the water surface image includes: In any scene, the water surface area and the dam area of ​​the water surface image are separated and labeled to generate a binarized image.

4. The method according to claim 1, characterized in that, The step of comparing the total confidence level with a preset threshold to determine whether the surveillance camera is obstructed includes: If the total confidence level is greater than a preset threshold, it is determined that the surveillance camera is blocked. If the total confidence level is not greater than a preset threshold, then the surveillance camera is determined to be unobstructed.

5. The method according to claim 1, characterized in that, Also includes: Based on the binary segmentation image, determine whether the water surface is abnormal.

6. The method according to claim 5, characterized in that, The step of determining whether the water surface is abnormal based on the binary segmentation image includes: The points in the binary segmented image are divided into several classes according to the clustering algorithm, and the classes that do not belong to the water surface are removed; Determine the maximum extent of the flood control dam's water area, and calculate the number of pixels in the binary segmentation image that belong to the water category and fall outside the maximum extent. If the number of pixels belonging to the water category in the binary segmented image that fall outside the maximum range exceeds a preset pixel threshold, then the water level corresponding to the binary segmented image exceeds the limit.

7. An obstruction detection device for a monitoring camera at an open-pit mine flood control dam, characterized in that, The device is used to implement the occlusion detection method for monitoring cameras at open-pit mine flood control dams as described in claim 1, and the device includes: The acquisition module is used to acquire water surface images under different environmental and water level conditions from the monitoring camera, identify the scene of the water surface image, and perform binarization and augmentation processing on the water surface image in each scene to generate a water surface image set. The semantic segmentation module is used to perform water surface semantic segmentation processing on the first image corresponding to the first scene in the water surface image set to generate a binary segmented image. The first occlusion detection module is used to construct a binary classification neural network to perform a preliminary detection of whether the surveillance camera is occluded, and obtain a first occlusion recognition result. The second occlusion detection module is used to determine whether the surveillance camera is occluded based on the first confidence level of scene recognition, the second confidence level of water surface semantic segmentation processing, and the third confidence level of occlusion detection.

8. A computer device, characterized in that, The method includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the computer program, it implements the method as described in any one of claims 1-6.

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

Images