A liquid leakage detection method based on deep learning, an electronic device, and a storage medium

Abstract

This application relates to a deep learning-based leak detection method, electronic device, and storage medium. The deep learning-based leak detection method includes: acquiring a first image of a target to be detected; dividing the first image into a leak detection area and a rainwater detection area; and performing leak detection on the divided first image using a pre-trained leak detection model to determine a first leak judgment result. This application solves the problem of false alarms in leak detection methods during rainy days and washing / rinsing scenarios in practical applications, reducing false alarms in rainy or washing / rinsing conditions and improving the alarm accuracy of the entire system.

Description

Technical Field

[0001] This application relates to the field of leakage control technology, and in particular to a leakage detection method, electronic device, and storage medium based on deep learning. Background Technology

[0002] In oil and gas, chemical and other production processes, there is often a need to prevent liquid leakage in pipelines where there are equipment such as storage tanks, valves, and flanges. In the past, this was done by manual on-site inspections. Now, some intelligent algorithms are being deployed in video systems for automated inspections.

[0003] However, current remote automated video surveillance systems often generate numerous false alarms in rainy weather or during cleaning and rinsing scenarios, failing to distinguish between rainwater and leaks caused by liquid leakage. The system frequently receives alarms, only to discover upon inspection that they are false alarms. With so many false alarms, users may become desensitized to them, developing a distrust of the system and consequently ignoring even legitimate alarms.

[0004] Reducing false alarms in rainy or washing conditions and improving the overall system's alarm accuracy has become a crucial factor in the application of remote automatic video inspection in various industries.

[0005] Currently, there is no effective solution to the problem of missed detections in leakage detection methods during rainy days and washing / rinsing scenarios in practical applications. Summary of the Invention

[0006] This embodiment provides a deep learning-based leakage detection method, electronic device, and storage medium to address the problem of missed detection in leakage detection methods during rainy days and washing / rinsing scenarios in practical applications.

[0007] Firstly, this embodiment provides a deep learning-based leakage detection method, the method comprising:

[0008] Acquire the first image of the target to be detected;

[0009] The first image of the target to be detected is divided into a leakage detection area and a rainwater detection area;

[0010] The first target image to be detected after region division is used to detect leakage through a pre-trained leakage detection model to determine the first leakage judgment result.

[0011] In some embodiments, the process of performing leak detection on the first target image after region segmentation using a pre-trained leak detection model to determine the first leak judgment result includes:

[0012] The leakage detection model is used to detect leakage in the leakage detection area of ​​the first target image to be detected, and leakage detection results are obtained.

[0013] The leakage detection model is used to detect rainwater in the rainwater detection area of ​​the first target image to be detected, and a first rainwater detection result is obtained.

[0014] Based on the leakage detection results and the rainwater detection results, the first leakage judgment result is determined.

[0015] In some embodiments, determining the first leakage judgment result based on the leakage detection result and the rainwater detection result includes:

[0016] Set the confidence threshold for leak detection;

[0017] The first leakage judgment result is determined based on the leakage detection result, the first rainwater detection result, and the leakage detection confidence threshold.

[0018] In some embodiments, the method further includes:

[0019] When the leakage detection result indicates that leakage has occurred, the coordinates of the location point of the leakage area corresponding to the first target image to be detected are obtained.

[0020] In some embodiments, the method further includes:

[0021] Obtain a training sample set, which includes target image data with leakage detection area and rainwater detection area;

[0022] The sample set is augmented to obtain an expanded sample set.

[0023] The expanded sample set is labeled to obtain the expanded labeled sample set;

[0024] The leakage detection model is trained based on the expanded labeled sample set.

[0025] In some embodiments, the method further includes:

[0026] Acquire the image of the second target to be detected;

[0027] The leakage detection model is used to detect rainwater in the second target image to obtain the second rainwater detection result.

[0028] Secondly, this embodiment provides a method for determining leakage, the method comprising:

[0029] Determine the result of the first leak assessment;

[0030] Obtain the second rainwater detection result;

[0031] Based on the first leakage judgment result and the second rainwater detection result, a second leakage judgment result is obtained; the second leakage judgment result is used to identify whether leakage has occurred in the leakage detection area of ​​the first target image to be detected.

[0032] Wherein, the first leakage judgment result is the first leakage judgment result obtained by the deep learning-based leakage detection method described in the first aspect; the second rainwater detection result is the second rainwater detection result obtained by the deep learning-based leakage detection method described in the first aspect.

[0033] Thirdly, this embodiment provides an intelligent analysis server, the server comprising:

[0034] The acquisition unit is used to acquire the first image of the target to be detected;

[0035] A segmentation unit is used to divide the first target image to be detected into a leakage detection area and a rainwater detection area;

[0036] The judgment unit is used to perform leakage detection on the first target image to be detected after region division through a pre-trained leakage detection model, and determine the first leakage judgment result.

[0037] Fourthly, this embodiment provides an electronic device including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the deep learning-based leakage detection method described in the first aspect; or, the processor is configured to run the computer program to execute the leakage judgment method described in the second aspect.

[0038] Fifthly, this embodiment provides a storage medium storing a computer program that, when executed by a processor, implements the deep learning-based leakage detection method described in the first aspect; or, when executed by a processor, the computer program implements the deep learning-based leakage judgment method described in the second aspect.

[0039] Compared with related technologies, the leakage detection method, leakage judgment method, intelligent analysis server, electronic device and storage medium based on deep learning provided in this embodiment acquire a first target image to be detected; divide the first target image to be detected into a leakage detection area and a rainwater detection area; and perform leakage detection on the first target image after area division through a pre-trained leakage detection model to determine the first leakage judgment result. This solves the problem of false alarms in leakage detection methods in many rainy days and washing and rinsing scenarios in practical applications, reduces false alarms in rainy days or washing and rinsing situations, and improves the alarm accuracy of the entire system.

[0040] Details of one or more embodiments of this application are set forth in the following drawings and description to make other features, objects and advantages of this application more readily apparent. Attached Figure Description

[0041] The accompanying drawings, which are included to provide a further understanding of this application and form part of this application, illustrate exemplary embodiments and are used to explain this application, but do not constitute an undue limitation of this application. In the drawings:

[0042] Figure 1 This is a hardware structure block diagram of a terminal that executes a deep learning-based leakage detection method according to this embodiment;

[0043] Figure 2 This is a structural block diagram of a leakage detection system according to this embodiment;

[0044] Figure 3 This is a flowchart of a deep learning-based leakage detection method in this embodiment;

[0045] Figure 4 This is a flowchart of a leakage detection method according to this embodiment;

[0046] Figure 5 This is a schematic diagram of the structure of a leakage detection system according to a preferred embodiment of the present invention;

[0047] Figure 6 This is a flowchart of a leakage detection method according to a preferred embodiment of the present invention;

[0048] Figure 7 This is a structural block diagram of an intelligent analysis server in this embodiment. Detailed Implementation

[0049] To better understand the purpose, technical solution, and advantages of this application, the application is described and illustrated below in conjunction with the accompanying drawings and embodiments.

[0050] Unless otherwise defined, the technical or scientific terms used in this application shall have the general meaning as understood by one of ordinary skill in the art to which this application pertains. Words such as “a,” “an,” “an,” “the,” “the,” and “these,” used in this application, do not indicate quantitative limitation and may be singular or plural. The terms “comprising,” “including,” “having,” and any variations thereof used in this application are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or device that comprises a series of steps or modules (units) is not limited to the listed steps or modules (units) but may include steps or modules (units) not listed, or may include other steps or modules (units) inherent to such processes, methods, products, or devices. The terms “connected,” “linked,” and “coupled,” used in this application, are not limited to physical or mechanical connections but may include electrical connections, whether direct or indirect. The term “multiple” used in this application refers to two or more. The "and / or" operator describes the relationship between related objects, indicating that three relationships can exist. For example, "A and / or B" can represent three cases: A alone, A and B simultaneously, and B alone. Typically, the character " / " indicates that the objects before and after it are in an "or" relationship. The terms "first," "second," and "third," etc., used in this application are merely for distinguishing similar objects and do not represent a specific ordering of the objects.

[0051] The method embodiments provided in this example can be executed on a terminal, computer, or similar computing device. For example, it can run on a terminal. Figure 1 This is a hardware structure block diagram of a terminal executing a deep learning-based leakage detection method according to this embodiment. For example... Figure 1 As shown, a terminal may include one or more ( Figure 1 Only one is shown in the diagram. A processor 102 and a memory 104 for storing data are also included. The processor 102 may be, but is not limited to, a microprocessor (MCU) or a programmable logic device (FPGA). The terminal may also include a transmission device 106 for communication functions and an input / output device 108. Those skilled in the art will understand that… Figure 1 The structure shown is for illustrative purposes only and does not limit the structure of the terminal described above. For example, the terminal may also include components that are larger than... Figure 1 The more or fewer components shown, or having the same Figure 1 The different configurations shown are illustrated.

[0052] The memory 104 can be used to store computer programs, such as application software programs and modules, like the computer program corresponding to a deep learning-based leakage detection method in this embodiment. The processor 102 executes various functional applications and data processing by running the computer program stored in the memory 104, thus implementing the aforementioned method. The memory 104 may include high-speed random access memory and non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memory. In some instances, the memory 104 may further include memory remotely located relative to the processor 102, and these remote memories can be connected to the terminal via a network. Examples of such networks include, but are not limited to, the Internet, corporate intranets, local area networks, mobile communication networks, and combinations thereof.

[0053] The transmission device 106 is used to receive or send data via a network. This network includes a wireless network provided by the terminal's communication provider. In one example, the transmission device 106 includes a Network Interface Controller (NIC), which can connect to other network devices via a base station to communicate with the Internet. In another example, the transmission device 106 can be a Radio Frequency (RF) module used for wireless communication with the Internet.

[0054] This embodiment provides a leakage detection system. Figure 2 This is a structural block diagram of a leakage detection system according to this embodiment, as shown below. Figure 2 As shown, the system includes: multiple front-end devices 10, a server 20, and a management platform 30.

[0055] Specifically, multiple front-end devices 10 communicate with server 20 through wireless communication modules, and server 20 collects and analyzes video data at predetermined intervals.

[0056] Server 20 is trained on data from a database using a deep learning algorithm based on a convolutional neural network (CNN) with two or more layers to create an artificial intelligence model for automatically detecting video data about liquid leakage collected by front-end device 10.

[0057] Server 20 utilizes data from a database containing initial data regarding liquid leaks. Server 20 sends detected anomalies along with images of liquid leaks captured by front-end device 10 to management platform 30.

[0058] When the server 20 detects the video frame data transmitted by the current front-end device 10, it acquires the first target image to be detected transmitted by the current front-end device 10, divides the first target image to be detected into a leakage detection area and a rainwater detection area, and performs leakage detection on the first target image after the area division through a pre-trained leakage detection model to determine the first leakage judgment result.

[0059] Specifically, the front-end device 10 monitors areas at risk of liquid leakage, captures relevant videos, and transmits the videos to the server 20. The server 20 divides the currently received video frame into a leakage detection area and a rainwater detection area. Under normal circumstances, the leakage detection area is a rectangular area or other graphic area drawn in the middle of the video frame, and the rainwater detection area is a rectangular area or other graphic area drawn around the leakage detection area in the video frame. The first target image to be detected after the area division is used to detect leakage through a pre-trained leakage detection model. When the leakage detection area detects a certain amount of liquid, while the rainwater detection area does not contain liquid, it can be determined that leakage has occurred and false alarms caused by rainwater erosion are excluded.

[0060] After completing the leakage detection in the video frames transmitted by the current front-end device 10, the server 20 can also perform rain detection on the video frames transmitted by adjacent front-end devices 10 to further determine whether there is rain erosion. The leakage judgment result of the current front-end device 10 and the rain judgment result of the adjacent front-end device 10 are sent to the management platform for secondary judgment to further reduce false alarms of leakage detection caused by rain.

[0061] This embodiment provides a deep learning-based leakage detection method. Figure 3 This is a flowchart of a deep learning-based leakage detection method according to this embodiment, as shown below. Figure 3 As shown, the process includes the following steps:

[0062] Step S302: Obtain the first image of the target to be detected.

[0063] Specifically, the server acquires the first image of the target to be detected. This first image refers to the target image used to detect whether leakage exists. It can be obtained from a video recording captured by a front-end device. The front-end device captures video of the target area, which is then transmitted back to the server via a network cable. The server uses a leakage detection model to detect each frame of the video (i.e., the first image of the target to be detected). The front-end device includes, but is not limited to, cameras and network video recorders.

[0064] Step S304: Divide the first target image to be detected into a leakage detection area and a rainwater detection area.

[0065] Specifically, the server divides the first target image into a leakage detection area and a rainwater detection area based on the position and angle of the front-end device and the predicted position of the monitoring target in the front-end device's image. Generally, the leakage detection area is a rectangular area or other graphic area drawn in the middle of the video frame, and the rainwater detection area is a rectangular area or other graphic area drawn around the leakage detection area in the video frame.

[0066] Step S306: The first target image to be detected after region division is used to perform leakage detection through a pre-trained leakage detection model to determine the first leakage judgment result.

[0067] Specifically, the server performs leak detection on the first target image after region segmentation using a pre-trained leak detection model. The first leak determination result is that leaks occur in the first target image when liquid is detected in the leak detection area and not in the rainwater detection area. By performing liquid detection in both the leak detection area and the rainwater detection area separately, and generating corresponding judgment results based on the liquid detection results in both areas, leaks can be confirmed and false alarms caused by rainwater erosion can be eliminated.

[0068] Through the above steps S302 to S306, the first target image to be detected is acquired; the first target image to be detected is divided into a leakage detection area and a rainwater detection area; the first target image to be detected after area division is used to perform leakage detection through a pre-trained leakage detection model to determine the first leakage judgment result; this solves the problem of false alarms in leakage detection methods in many rainy days and washing and rinsing scenarios in practical applications, reduces false alarms in rainy days or washing and rinsing situations, and improves the alarm accuracy of the whole system.

[0069] In some embodiments, the process of detecting leakage in the first target image to be detected using a leakage detection model and determining the first leakage judgment result includes: performing leakage detection on the leakage detection area of ​​the first target image to be detected using the leakage detection model to obtain a leakage detection result; performing rainwater detection on the rainwater detection area of ​​the first target image to be detected using the leakage detection model to obtain a first rainwater detection result; and determining the first leakage judgment result based on the leakage detection result and the rainwater detection result.

[0070] Specifically, the leakage detection result is as follows: when the detected liquid level in the leakage detection area of ​​the first target image reaches a set threshold, leakage occurs; when the detected liquid level in the leakage detection area of ​​the first target image does not reach the set threshold, leakage does not occur. Similarly, the rainwater detection result is as follows: when the detected liquid level in the rainwater detection area of ​​the first target image reaches a set threshold, rainwater occurs; when the detected liquid level in the leakage detection area of ​​the first target image does not reach the set threshold, rainwater does not occur. The first leakage judgment result includes... If the liquid level in both the leakage detection area and the rainwater detection area of ​​the first target image does not reach the set threshold, the first leakage judgment result is that no leakage has occurred; if the liquid level in the leakage detection area reaches the set threshold, but the liquid level in the rainwater detection area does not reach the set threshold, the first leakage judgment result is that leakage has occurred; if the liquid level in both the leakage detection area and the rainwater detection area of ​​the first target image reaches the set threshold, the first leakage judgment result is that it is uncertain whether leakage has occurred; the set thresholds for the liquid level in the leakage detection area and the rainwater detection area are not necessarily the same.

[0071] In some embodiments, determining the first leakage judgment result based on the leakage detection result and the rainwater detection result includes: setting a leakage detection confidence threshold; and determining the first leakage judgment result based on the leakage detection result, the first rainwater detection result, and the leakage detection confidence threshold.

[0072] Specifically, a leakage detection confidence threshold is set; the first target image to be detected is subjected to leakage detection through the leakage detection model, and the detection box scores of the leakage detection area and the rainwater detection area are output; the first leakage judgment result is determined based on the detection box scores of the leakage detection area, the detection box scores of the rainwater detection area, and the leakage detection confidence threshold.

[0073] Furthermore, when the detection frame score of the leakage detection area reaches the set leakage detection confidence threshold, while the detection frame score of the rainwater detection area is less than the set leakage detection confidence threshold, it can be determined that leakage has occurred and false alarms caused by rainwater erosion can be ruled out.

[0074] In some embodiments, the method further includes:

[0075] When the leakage detection result indicates that leakage has occurred, the coordinates of the location point of the leakage area corresponding to the first target image to be detected are obtained.

[0076] Specifically, when the detection frame score of the leakage detection area reaches the set leakage detection confidence threshold, while the detection frame score of the rainwater detection area is less than the set leakage detection confidence threshold, a leakage is determined to have occurred. Then, the coordinates of the location point of the leakage area corresponding to the first target image to be detected are returned so that the user can confirm the location of the leakage and carry out repairs.

[0077] In some embodiments, the method further includes:

[0078] Obtain a training sample set, which includes target image data with leakage detection area and rainwater detection area;

[0079] The sample set is augmented to obtain an expanded sample set;

[0080] The expanded sample set is labeled to obtain the expanded labeled sample set;

[0081] The leakage detection model is trained based on the expanded labeled sample set.

[0082] Specifically, the server can obtain a training sample set from the video data captured historically by the front-end device. The training sample set includes target image data with leakage detection area and rainwater detection area. The image set of leakage detection area and rainwater detection area is subjected to local data augmentation processing to obtain an expanded sample set. The expanded sample set is labeled to obtain an expanded labeled sample set.

[0083] In some embodiments, the method further includes: acquiring a second target image to be detected; performing rainwater detection on the second target image to be detected using a leakage detection model to obtain a second rainwater detection result.

[0084] Specifically, the second target image to be detected is a video frame extracted from the video recording taken by a front-end device adjacent to the front-end device that acquired the first target image to be detected. Rain detection is performed on the video frames taken by the adjacent front-end device to help determine whether rain has occurred. When the liquid in both the leakage detection area and the rain detection area of ​​the first target image to be detected reaches the set threshold, the first leakage judgment result is uncertain whether leakage has occurred. Based on the acquired second rain detection result, if the second rain detection result is no rain, then leakage is determined to have occurred; if the second rain detection result is rain, then the management platform determines that leakage has not occurred.

[0085] By performing a secondary assessment of rainwater detection results from neighboring areas, false alarms in leak detection caused by rainwater can be further reduced.

[0086] This embodiment also provides a method for determining leakage. Figure 4 This is a flowchart of a leakage detection method according to this embodiment, such as... Figure 4 As shown, the process includes the following steps:

[0087] Step S402: Determine the first leakage judgment result.

[0088] Specifically, the management platform obtains the first leakage judgment result obtained by the server from the leakage detection of the first target image to be detected.

[0089] Step S404: Obtain the second rainwater detection result.

[0090] Specifically, the management platform obtains the second rainwater detection result obtained by the server from rainwater detection of the second target image.

[0091] Step S406: Based on the first leakage judgment result and the second rainwater detection result, obtain the second leakage judgment result; the second leakage judgment result is used to identify whether leakage has occurred in the leakage detection area of ​​the first target image to be detected.

[0092] Specifically, when the first leakage judgment result indicates that leakage has occurred, the management platform receives the second rainwater detection result obtained from the intelligent analysis server. If the second rainwater detection result indicates that no rain has occurred, the management platform determines that leakage has occurred; if the second rainwater detection result indicates that rain has occurred, the management platform determines that leakage has not occurred.

[0093] Through steps S402 to S406 above, the management platform performs a secondary judgment on the rainwater and leakage alarms reported by the intelligent analysis server based on the rainwater detection results of the adjacent area, further reducing false alarms of leakage detection caused by rainwater.

[0094] The present embodiment will now be described and illustrated through preferred embodiments.

[0095] Figure 5 This is a schematic diagram of a leakage detection system according to a preferred embodiment of this invention. Figure 5 As shown, the system includes: a front-end high-definition camera 40, an intelligent analysis server 50, and a management platform 60.

[0096] The front-end high-definition camera 40 is matched with the required number of high-definition cameras according to the requirements of the leakage detection area to achieve high-definition video data acquisition.

[0097] The intelligent analysis server 50 backend intelligent analysis server adopts a visual intelligent analysis algorithm to set up leakage detection areas and rainwater detection areas on high-definition video images, realize single-area leakage detection and rainwater detection, and store real-time video to achieve high reliability and high availability of the storage system.

[0098] The intelligent analysis server 50 makes a judgment by comparing the detected items with the actual given threshold, and the judgment conditions are as follows:

[0099] 1) If the leakage area in a certain detection zone reaches the alarm threshold, but no alarm is triggered in the rainwater detection zone, it is determined that a leakage has occurred, and a leakage alarm is issued to prompt the on-duty personnel to conduct video verification and confirmation.

[0100] 2) If the leakage area in a certain detection area reaches the alarm threshold, the rainwater detection area will also trigger an alarm. It is necessary to rule out whether a sudden large-scale leakage has occurred. If rainwater is not detected in the rainwater detection areas of other nearby cameras, it indicates that a large-scale leakage has occurred here, triggering a leakage alarm and reminding the on-duty personnel to conduct video verification and confirmation.

[0101] 3) In other cases, no alarm should be reported.

[0102] Management Platform 60: The platform features modular deployment. It includes a video surveillance system management module for unified management of front-end camera devices, supporting recording and playback functions; a customizable voice module for configuring alarm alert voices; and the ability to perform secondary assessments of nearby areas on rain and leakage alarms reported by the intelligent analysis server, further reducing false alarms.

[0103] It should be noted that the steps shown in the above process or in the flowchart of the accompanying figures can be executed in a computer system such as a set of computer-executable instructions, and although a logical order is shown in the flowchart, in some cases the steps shown or described may be executed in a different order than that shown here.

[0104] This preferred embodiment provides a method for detecting leakage. Figure 6 This is a flowchart of a leakage detection method according to a preferred embodiment of the present invention, as follows: Figure 6 As shown, the process includes the following steps:

[0105] Step S602: Draw the rain detection area and the leakage detection area on the first video frame.

[0106] Step S604: Perform leakage detection on the leakage detection area of ​​the first video frame.

[0107] Step S606: Perform rain detection on the rain detection area of ​​the first video frame.

[0108] Step S608: When the leakage detection area of ​​the first video frame detects leakage, while the rainwater detection area detects no rainwater, an alarm signal is triggered.

[0109] Specifically, if the leakage area in a certain detection area corresponding to the first video frame reaches the alarm threshold, and no alarm is triggered in the rainwater detection area, it is determined that a leakage has occurred, and a leakage alarm is issued to prompt the on-duty personnel to review and confirm via video.

[0110] By drawing rainwater detection area and leakage detection area on the first video frame, leakage detection is performed in the leakage detection area of ​​the first video frame, and rainwater detection is performed in the rainwater detection area of ​​the first video frame. When the leakage detection area of ​​the first video frame detects leakage, but the rainwater detection area detects no rain, an alarm signal is triggered. This solves the problem of missed alarms in leakage detection methods in many rainy days and washing and rinsing scenarios in practical applications, reduces false alarms in rainy days or washing and rinsing situations, and improves the alarm accuracy of the entire system.

[0111] In some embodiments, when the leakage detection area of ​​the first video frame detects leakage and the rain detection area detects rain, rain detection is performed on the second video frame captured by the adjacent camera and uploaded to the management platform for secondary judgment.

[0112] Specifically, if the leakage area in a certain detection zone reaches the alarm threshold, the rainwater detection zone will also trigger an alarm. It is necessary to rule out whether a sudden large-scale leakage has occurred. If rainwater is not detected in the rainwater detection zones of other nearby cameras, it indicates that a large-scale leakage has occurred here, triggering a leakage alarm and reminding the on-duty personnel to conduct a video review and confirmation.

[0113] This embodiment also provides an intelligent analysis server for implementing the above embodiments and preferred embodiments; details already described will not be repeated. The terms "module," "unit," "subunit," etc., used below refer to combinations of software and / or hardware that perform a predetermined function. Although the apparatus described in the following embodiments is preferably implemented in software, hardware implementation, or a combination of software and hardware, is also possible and contemplated.

[0114] Figure 7 This is a structural block diagram of an intelligent analysis server according to this embodiment, such as... Figure 7 As shown, the intelligent analysis server includes:

[0115] Acquisition unit 70 is used to acquire the first target image to be detected;

[0116] The segmentation unit 80 is used to divide the first target image to be detected into a leakage detection area and a rainwater detection area;

[0117] The judgment unit 90 is used to perform leakage detection on the first target image to be detected after the region division through a pre-trained leakage detection model, and determine the first leakage judgment result.

[0118] It should be noted that the above modules can be functional modules or program modules, and can be implemented through software or hardware. For modules implemented through hardware, the above modules can reside in the same processor; or the above modules can be located in different processors in any combination.

[0119] This embodiment also provides an electronic device including a memory and a processor, the memory storing a computer program and the processor being configured to run the computer program to perform the steps in any of the above method embodiments.

[0120] Optionally, the electronic device may further include a transmission device and an input / output device, wherein the transmission device is connected to the processor and the input / output device is connected to the processor.

[0121] Optionally, in this embodiment, the processor can be configured to perform the following steps via a computer program:

[0122] S1, acquire the first image of the target to be detected.

[0123] S2, divide the first target image to be detected into a leakage detection area and a rainwater detection area.

[0124] S3, the first target image to be detected after region division is used to perform leakage detection through a pre-trained leakage detection model to determine the first leakage judgment result.

[0125] Optionally, in this embodiment, the processor may also be configured to perform the following steps via a computer program:

[0126] S11, Determine the first leakage judgment result.

[0127] S12, obtain the second rainwater detection result.

[0128] S13, based on the first leakage judgment result and the second rainwater detection result, obtain a second leakage judgment result; the second leakage judgment result is used to identify whether leakage has occurred in the leakage detection area of ​​the first target image to be detected.

[0129] It should be noted that the specific examples in this embodiment can refer to the examples described in the above embodiments and optional implementations, and will not be repeated in this embodiment.

[0130] Furthermore, in conjunction with the deep learning-based leakage detection method provided in the above embodiments, this embodiment can also provide a storage medium for implementation. This storage medium stores a computer program; when executed by a processor, the computer program implements any of the deep learning-based leakage detection methods described in the above embodiments.

[0131] Optionally, in conjunction with the leakage detection method provided in the above embodiments, this embodiment can also provide a storage medium for implementation. The storage medium stores a computer program; when executed by a processor, the computer program implements any of the leakage detection methods in the above embodiments.

[0132] It should be understood that the specific embodiments described herein are merely illustrative of the application and not intended to limit it. All other embodiments derived by those skilled in the art based on the embodiments provided in this application without inventive effort are within the scope of protection of this application.

[0133] Obviously, the accompanying drawings are merely some examples or embodiments of this application. Those skilled in the art can apply this application to other similar situations based on these drawings without any creative effort. Furthermore, it is understood that although the work done in this development process may be complex and lengthy, for those skilled in the art, certain design, manufacturing, or production modifications made based on the technical content disclosed in this application are merely conventional technical means and should not be considered as insufficient disclosure of this application.

[0134] The term "embodiment" in this application refers to a specific feature, structure, or characteristic described in connection with an embodiment that may be included in at least one embodiment of this application. The appearance of this phrase in various places in the specification does not necessarily imply the same embodiment, nor does it imply that it is mutually exclusive with or independent of other embodiments. It will be clearly or implicitly understood by those skilled in the art that the embodiments described in this application may be combined with other embodiments without conflict.

[0135] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of patent protection. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the appended claims.

Claims

1. A liquid leakage detection method based on deep learning, characterized in that, The method includes: Acquire the first image of the target to be detected; The first image of the target to be detected is divided into a leakage detection area and a rainwater detection area; The first target image to be detected after being divided into regions is used to detect leakage through a pre-trained leakage detection model to determine the first leakage judgment result. The process of performing leakage detection on the first target image after region segmentation using a pre-trained leakage detection model, and determining the first leakage judgment result, includes: Set a leakage detection confidence threshold; perform leakage detection on the first target image through the leakage detection model, and output the detection box scores of the leakage detection area and the rainwater detection area; determine the first leakage judgment result based on the detection box scores of the leakage detection area, the rainwater detection area, and the leakage detection confidence threshold; when the detection box score of the leakage detection area reaches the set leakage detection confidence threshold, while the detection box score of the rainwater detection area is less than the set leakage detection confidence threshold, then leakage is determined to have occurred and false alarms caused by rainwater erosion are excluded. 2.The liquid leakage detection method based on deep learning according to claim 1, characterized in that, The method further includes: When the leakage detection result indicates that leakage has occurred, the coordinates of the location point of the leakage area corresponding to the first target image to be detected are obtained. 3.The liquid leakage detection method based on deep learning according to claim 1, characterized in that, The method further includes: Obtain a training sample set, which includes target image data with leakage detection area and rainwater detection area; The sample set is augmented to obtain an expanded sample set. The expanded sample set is labeled to obtain the expanded labeled sample set; The leakage detection model is trained based on the expanded labeled sample set. 4.The liquid leakage detection method based on deep learning according to claim 1, characterized in that, The method further includes: Acquire the image of the second target to be detected; The leakage detection model is used to detect rainwater in the second target image to obtain the second rainwater detection result.

5. A method of determining a liquid leakage, characterized by, The method includes: Determine the result of the first leak assessment; Obtain the second rainwater detection result; Based on the first leakage judgment result and the second rainwater detection result, a second leakage judgment result is obtained; the second leakage judgment result is used to identify whether leakage has occurred in the leakage detection area of ​​the first target image to be detected. Wherein, the first leakage judgment result is the first leakage judgment result obtained by the deep learning-based leakage detection method according to any one of claims 1 to 4; the second rainwater detection result is the second rainwater detection result obtained by the deep learning-based leakage detection method according to claim 4.

6. An intelligent analytics server, characterized by The server includes: The acquisition unit is used to acquire the first image of the target to be detected; A segmentation unit is used to divide the first target image to be detected into a leakage detection area and a rainwater detection area; The judgment unit is used to perform leakage detection on the first target image to be detected after region division through a pre-trained leakage detection model, and determine the first leakage judgment result. The process of performing leakage detection on the first target image after region segmentation using a pre-trained leakage detection model, and determining the first leakage judgment result, includes: Set a leakage detection confidence threshold; perform leakage detection on the first target image through the leakage detection model, and output the detection box scores of the leakage detection area and the rainwater detection area; determine the first leakage judgment result based on the detection box scores of the leakage detection area, the rainwater detection area, and the leakage detection confidence threshold; when the detection box score of the leakage detection area reaches the set leakage detection confidence threshold, while the detection box score of the rainwater detection area is less than the set leakage detection confidence threshold, then leakage is determined to have occurred and false alarms caused by rainwater erosion are excluded. 7.An electronic device comprising a memory and a processor, the electronic device characterized by, The memory stores a computer program, and the processor is configured to run the computer program to execute the deep learning-based leakage detection method according to any one of claims 1 to 4; or, the processor is configured to run the computer program to execute the leakage judgment method according to claim 5.

8. A computer-readable storage medium having stored thereon a computer program, characterized in that, When the computer program is executed by the processor, it implements the steps of the deep learning-based leakage detection method according to any one of claims 1 to 4; or, when the computer program is executed by the processor, it implements the steps of the leakage judgment method according to claim 5.

Images