Garbage identification method and system for trash rack of hydraulic power plant

Abstract

The invention relates to the technical field of hydraulic power plant garbage identification, in particular to a hydraulic power plant trash rack garbage identification method, which comprises the following steps: acquiring a shot image at a trash rack; classifying the shot images by adopting a classification model, and outputting a classification result; based on the classification result, judging whether garbage exists in the shot image or not; processing the shot image to obtain a foreground image and a background image under the condition of judging that the garbage exists; and calculating the area of the foreground image, comparing the area with a preset threshold value, and if the area is greater than the preset threshold value, generating alarm information and sending the alarm information to a terminal. The invention provides an efficient, intelligent and low-cost garbage identification scheme for the trash rack of the hydraulic power plant, overcomes the defects of dependence on manpower, response lag, low accuracy and the like in the prior art, and has remarkable application prospects and popularization values in automatic operation and maintenance of the hydraulic power plant and intelligent water conservancy construction.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of security monitoring, and relates to a method and system for identifying garbage on a trash rack of a hydropower plant. BACKGROUND

[0002] Hydroelectric power generation, as an important part of renewable energy, is widely used in areas rich in water resources. Its principle is mainly to convert the potential energy of high water flow into kinetic energy through a water turbine, and then drive a generator to complete the output of electric energy. However, in natural river channels, reservoirs or diversion channels, water flow often carries a large amount of floating debris, such as waterweeds, branches, dead leaves, bird and fish carcasses, plastic products and other garbage generated by human life or industry. If these floating objects enter the water inlet or water turbine blade area of the hydroelectric power generation system, they can easily cause equipment blockage, poor rotation, and even cause the unit to shut down, affecting the overall power generation efficiency and power grid power supply stability.

[0003] To prevent floating objects from entering the water turbine, a floating trash rack is commonly used as a protective device in the prior art. This device can float on the water surface according to the water level change and form a physical barrier in front of the water inlet, which can intercept floating objects and play a protective role. However, over time, floating objects will continue to accumulate in front of the trash rack, and once the critical amount is exceeded, not only will it affect the water flow rate and increase the water head loss, but it may also cause the trash rack to be "blocked", thereby threatening the safety of the power plant operation.

[0004] Currently, whether the trash rack is blocked by floating objects is still mainly determined by manual inspection and visual judgment. This method not only consumes a large amount of human resources and increases the cost of operation and maintenance, but also has many problems such as delayed response, large subjective error in judgment, and inability to effectively patrol at night or in bad weather. Especially during the flood or peak of the rainy season, the accumulation rate of garbage is much higher than usual, and manual intervention methods are difficult to meet the real-time processing needs, so an intelligent automatic judgment and early warning mechanism is needed to replace the traditional manual monitoring method. SUMMARY

[0005] The method and system for identifying garbage on a trash rack of a hydropower plant provided by the present application break through the drawbacks of traditional reliance on manual judgment.

[0006] The present application is achieved by the following technical solutions: A method for identifying garbage on a trash rack of a hydropower plant, comprising the following steps: S1, obtaining a shooting image at the trash rack; S2, using a classification model to classify the shooting image and output a classification result; S3, determining whether there is garbage in the shooting image based on the classification result; S4, in the case of determining that there is garbage, processing the photographed image to obtain a foreground image and a background image; S5, calculating an area of the foreground image, comparing the area with a preset threshold, and if the area is greater than the preset threshold, generating an alarm information and sending it to the terminal.

[0007] Further, step S1 uses a camera to photograph the trash rack at a preset time period, obtains the photographed image, and uploads the photographed image to a background system.

[0008] Further, the photographed image is a top view image of the trash rack.

[0009] Further, the classification model of step S2 is a pre-trained convolutional neural network (CNN) model, and the pre-training step includes: S201, collecting historical photographed images of the trash rack and labeling the historical photographed images to obtain a training data set; S202, dividing the data set into a training set and a validation set in a ratio of 8:2; S203, training the CNN model based on the training set and verifying the trained CNN model based on the validation set, iteratively adjusting the model parameters until the CNN model converges.

[0010] Further, step S4 of processing the photographed image specifically includes the following steps: S401, cropping the photographed image with garbage; S402, calculating the background threshold of the cropped image; S403, based on the background threshold, performing binaryzation processing on the cropped image to obtain the background image and the foreground image.

[0011] Further, the background threshold is calculated as follows: Define the pixel gray value in the cropped image as Then Where represents the upper limit value of the horizontal coordinate of the pixel point, and represents the upper limit value of the vertical coordinate .

[0012] Further, the binaryzation processing is calculated as follows: ; wherein represents the pixel value after binaryzation processing.

[0013] Further, the calculation of the area of the foreground image includes counting the number of foreground pixels in the foreground image, and calculating the area of the foreground image based on the area of a single pixel.

[0014] The application also includes a trash identification system for a trash screen of a hydropower plant, which is applicable to any of the above methods, comprising: an image acquisition module for acquiring a shooting image at the trash screen; a classification processing module for receiving the shooting image and performing image classification using a pre-constructed classification model, and outputting a classification result; a judgment module for judging whether there is trash in the shooting image based on the classification result; an image processing module for processing the shooting image to output a foreground image and a background image if the judgment module judges that there is trash; an area calculation module for calculating the area of the foreground image based on the foreground image; an alarm module for comparing the area with a preset threshold, and generating an alarm information and sending it to a terminal if the area is greater than the preset threshold.

[0015] The application has the following advantages: (1) The trash identification method for a trash screen of a hydropower plant proposed by the application can periodically shoot the trash screen by deploying an image acquisition device, and automatically classify and identify the image using a pre-trained convolutional neural network (CNN) model, which can judge whether there is trash in real time, replace manual inspection, and greatly reduce the operation and maintenance labor cost; (2) The trash identification method and system for a trash screen of a hydropower plant proposed by the application not only identify whether there is trash, but also further combine image processing algorithms to accurately calculate the area of the foreground region, use a set threshold to judge whether the trash is accumulated to the degree that needs to be cleaned, and only generate an alarm information when necessary, avoiding false positives and frequent disturbances; (3) The trash identification method for a trash screen of a hydropower plant proposed by the application uses fixed installation parameters to calculate the area of a single pixel in the image processing stage, does not need to rely on high-precision measuring equipment or complex geometric models, can be quickly deployed in different hydropower plant scenes, and has good engineering practicability; (4) The trash identification system for a trash screen of a hydropower plant proposed by the application supports dynamic adjustment of the shooting frequency, can automatically increase the sampling rate after identifying the trash, or realize self-adaptive adjustment by linking hydrological parameters in high-risk environments such as heavy rain and flood, and improve the real-time performance and response efficiency of the identification system; (5) The trash identification system for the trash screen of the hydropower plant can timely monitor the trash accumulation condition in front of the trash screen, assists the operation and maintenance personnel to reasonably dispatch and clean, avoids the water head loss or equipment failure caused by the trash blockage, and improves the operation efficiency and power generation safety level of the power station from the source; The trash identification scheme for the trash screen of the hydropower plant is efficient, intelligent and low in cost, overcomes the deficiencies of the prior art, such as dependence on manual operation, response lag and low accuracy, and has a significant application prospect and popularization value in the automatic operation and maintenance and intelligent water conservancy construction of the hydropower station. BRIEF DESCRIPTION OF DRAWINGS

[0016] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings needed in the embodiment description will be briefly introduced. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative labor.

[0017] Figure 1 The flow chart of the trash identification method for the trash screen of the hydropower plant is provided. Figure 2 The terminal device schematic diagram of the trash identification method and system for the trash screen of the hydropower plant is provided. Figure 3 The readable storage medium schematic diagram of the trash identification method and system for the trash screen of the hydropower plant is provided. In the figure, 200 is a terminal device, 210 is a memory, 211 is a RAM, 212 is a cache, 213 is a ROM, 214 is a program / utility, 215 is a program module, 220 is a processor, 230 is a bus, 240 is an external device, 250 is an I / O interface, 260 is a network adapter, 300 is a program product. DETAILED DESCRIPTION

[0018] In order to make the purpose, technical solutions and advantages of the present application more clear, the present application will be further described in detail below with examples and drawings, the schematic embodiments of the present application and the description thereof are only used to explain the present application, and do not limit the present application.

[0019] Example 1 The present application provides a trash identification method and system based on image recognition and area judgment, which combines a convolutional neural network (CNN) model to automatically classify, locate and calculate the area of trash in the image, and determines whether to trigger an alarm according to the area. The method has the advantages of high intelligence, strong adaptability and low false alarm rate. The present application will be described in detail below with specific examples.

[0020] In this embodiment, the image acquisition device deployed above the trash rack first periodically takes overhead images of the trash rack, which covers the entire grid area, ensuring that it can fully reflect whether there are floating objects, attachments or accumulations on the trash rack. The camera communicates with the background processing system through the network and uploads the collected images to the background server in real time or at a fixed time. This embodiment adopts a serial processing flow of image acquisition-classification recognition-image processing-area calculation-threshold judgment-warning, effectively realizing non-manual and automatic garbage monitoring during the operation of the hydropower plant.

[0021] In the image acquisition stage, the system is provided with a preset shooting period, which can be flexibly configured according to the operating environment, such as once every 3 hours, once every 6 hours, or set to 3 times a day according to the daily average garbage generation. The system supports dynamic adjustment of the shooting frequency on demand. When it is detected that suspicious targets exist in a plurality of consecutive images, the shooting period is automatically shortened to increase the sampling rate, so as to grasp the garbage accumulation trend faster. During special weather such as heavy rain and flood, the system can also be linked with water level or flow monitoring data to automatically start high-frequency acquisition mode to improve the real-time and accuracy of identification.

[0022] When the shooting image is uploaded to the background, it enters the classification and recognition stage. The present application uses a trained convolutional neural network (CNN) classification model to classify and process images. The CNN model can effectively distinguish between clean grid images and images with garbage by deeply abstracting and extracting features such as image texture, color distribution and edge structure. In the training stage, a large number of labeled historical trash rack images are used to construct a training set and a validation set with a ratio of 8:2. The classification labels are divided into two categories: no garbage and garbage, and the classification accuracy is required to be above 90%. The classification output is a probability vector or a binary decision label. The model training uses a cross-entropy loss function and is iteratively trained through the Adam optimization algorithm. The training convergence standard is that the validation set loss function no longer decreases or there is no significant improvement in continuous multiple iterations. After the model training is completed, it is deployed to the backend server to receive and classify images in real time.

[0023] If the classification result is garbage, it enters the image processing stage. The system automatically crops the image to remove irrelevant background areas, including shore structures and water surface reflection areas, to enhance the accuracy of subsequent image segmentation. The image cropping area is based on a preset size template, which generally retains the main body of the grid area in a center-symmetric manner, and its size is set to the maximum value that contains all possible garbage accumulation areas. Then, the background threshold T of the cropped image is calculated to separate the foreground (garbage part) from the background (clean water surface and grid body).

[0024] The calculation method of the background threshold T is: , Where is the average value of the image Gray value of coordinate point, And The horizontal and vertical dimensions of the calculation window. Based on The binarization processing is carried out, and the pixel points The original value is retained, and otherwise it is set to 0, to obtain the foreground image By this method, the main area of garbage can be effectively stripped from the complex background.

[0025] In order to reduce the probability of misidentification, the system adopts a moderate conservative cutting area size, and optimizes the uniform light period under the collection condition, so as to improve the stability of the background modeling. Since the overall process of the system is area judgment, no additional noise filtering or median processing is introduced after binarization processing, and only in specific scenarios, convolution kernel smoothing or inflation operation is used for supplementary enhancement. In this embodiment, no noise post-processing is performed by default, so as to simplify the calculation process.

[0026] After completing the foreground image, the system counts the pixel points in the foreground area. The corresponding area of each foreground pixel point in the image can be determined by factors such as shooting height, camera field of view angle and image resolution. Specifically, if the installation height of the camera is H, the width of the shooting area is W, and the corresponding image width is w pixels, then the width of a single pixel is W / w meters, and the area is (W / w) x (W / h). Since the image shooting and installation parameters are basically fixed, the area of a single pixel can be pre-calibrated or automatically converted without the need for dynamic measurement. In engineering deployment, this value is written into the system configuration file as an image processing parameter, and is automatically read when called. The system accumulates the number of all foreground pixel points, and multiplies the single-point area to obtain the total garbage area.

[0027] In this embodiment, the basis for whether to alarm is whether the foreground area is greater than a set threshold. The preset threshold is set according to user demand, operation experience or historical data, such as 10% of the surface area of the general case of the grid, which can be used as a reference threshold, or set to 0.3 m², 0.5 m², etc. The garbage retention capacity of different trash racks is different, and the set threshold supports user configuration. If the system determines that the area exceeds the threshold, it is determined that the garbage is seriously blocked, and the alarm module is triggered. This module can send garbage alarm information to the operation terminal through SMS, network notification or industrial SCADA platform push, etc., to remind the operator to clean up in time, and to ensure the smooth flow of water and the safety of the equipment.

[0028] To realize the above method, the present application constructs a trash rack garbage identification system for a hydropower plant, which comprises an image acquisition module, a classification processing module, a judgment module, an image processing module, an area calculation module and an alarm module. The functions of each module are as follows: The image acquisition module is used to be deployed above the trash rack, and the industrial camera is used to acquire images at a set frequency and upload them to the background; The classification processing module loads a pre-trained CNN model to automatically identify the presence or absence of garbage, and supports multi-thread high-concurrency processing. The judgment module makes existence judgment according to the classification result to decide whether to enter the image segmentation process. The image processing module performs standardization cropping, gray scale calculation, background separation and other operations on the image, and outputs foreground and background images. The area calculation module calculates the actual area of the target pixel based on the foreground image. The alarm module generates a garbage jam alarm according to the comparison result of the area and the set threshold, and sends it to the dispatching system or the operation and maintenance terminal.

[0029] Through the system and method of the embodiment, the automation monitoring capability of the trash screen operation of the hydropower plant can be significantly improved, and the artificial inspection intensity and the missed detection risk can be reduced. The system has good expansibility and generalization, and can be adapted to various types of hydropower station projects, and is suitable for different scenes such as main stream, branch stream and small hydropower.

[0030] Embodiment 2 Reference Figure 2 On the basis of embodiment 1, the embodiment provides a terminal device of a hydropower plant trash screen garbage identification method and system. The terminal device 200 includes at least one memory 210, at least one processor 220, and a bus 230 connecting different platform systems.

[0031] The memory 210 can include a readable medium in the form of a volatile memory, such as a RAM 211 and / or a cache 212 memory, and can further include a ROM 213.

[0032] The memory 210 also stores a computer program, which can be executed by the processor 220 to cause the processor 220 to execute any one of the above-mentioned hydropower plant trash screen garbage identification methods and systems in the embodiments of the application. The specific implementation manner and the achieved technical effects are consistent with those described in the embodiments of the above-mentioned applications, and some contents will not be repeated. The memory 210 can also include programs / utilities 214 having a set of (at least one) program modules 215, such as an operating system, one or more application programs, other program modules, and program data, each of which or some combination thereof can include the implementation of a network environment.

[0033] Correspondingly, the processor 220 can execute the above-mentioned computer program, and can execute the program / utilities 214.

[0034] The bus 230 may represent one or more of several types of bus structures, including a memory bus or memory controller, a peripheral bus, an accelerated graphics port, a processor, or a local bus using any of a variety of bus architectures.

[0035] The terminal device 200 can also communicate with one or more external devices 240, such as keyboards, pointing devices, Bluetooth devices, etc., and can also communicate with one or more devices that can interact with the terminal device 200, and / or communicate with any device that enables the terminal device 200 to communicate with one or more other computing devices (such as routers, modems, etc.). Such communication can be carried out through the I / O interface 250. In addition, the terminal device 200 can also communicate with one or more networks (such as local area networks (LANs), wide area networks (WANs) and / or public networks, such as the Internet) through the network adapter 260. The network adapter 260 can communicate with other modules of the terminal device 200 through the bus 230. It should be understood that although not shown in the figure, other hardware and / or software modules can be used in conjunction with the terminal device 200, including but not limited to: microcode, device drivers, redundant processors, external disk drive arrays, RAID systems, tape drives, and data backup storage platforms.

[0036] Example 3 refer to Figure 3 This embodiment proposes a computer-readable storage medium for a method and system for identifying garbage on a hydropower plant trash rack. Instructions are stored on the computer-readable storage medium. When the instructions are executed by the processor, any of the above-mentioned methods and systems for identifying garbage on a hydropower plant trash rack is implemented. The specific implementation method is consistent with the implementation method and the technical effect achieved in the above-mentioned application embodiments, and some contents will not be repeated here.

[0037] Figure 3A program product 300 for implementing the above application provided by the embodiment is shown, which can adopt a portable compact disc read-only memory (CD-ROM) and include program codes, and can run on a terminal device, such as a personal computer. However, the program product 300 of the present application is not limited to this, and in the embodiment, the readable storage medium can be any tangible medium containing or storing a program, which can be used by or in combination with an instruction execution system, device or apparatus. The program product 300 can adopt any combination of one or more readable media. The readable medium can be a readable signal medium or a readable storage medium. The readable storage medium can be, for example, but is not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device or apparatus, or any combination of the above. More specific examples (non-exhaustive list) of the readable storage medium include an electrical connection having one or more wires, a portable disc, 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 disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above.

[0038] The computer readable storage medium can include a data signal transported in a baseband or as part of a carrier wave, and can be readable by a computer. Such a propagated signal can take a wide variety of forms, including but not limited to electro-magnetic, optical, or any suitable combination thereof. A computer readable medium can also be any medium that can be read by a computer, including but not limited to any medium that stores, transmits, or receives the program code. The program code contained on the computer readable storage medium can be transmitted using any appropriate medium, including but not limited to wireless, wired, optical fiber cable, RF, etc., or any suitable combination thereof. The program code can be written in any combination of one or more programming languages, including an object oriented programming language such as Java, C++, etc., and conventional procedural programming languages, such as the "C" programming language or similar programming languages. The program code can execute entirely on the user's computing device, partly on the user's computing device, as a stand-alone software package, partly on the user's computing device and partly on a remote computing device or entirely on the remote computing device or server. In the latter scenario, the remote computing device can be connected to the user's computing device through any type of network, including a local area network (LAN) or a wide area network (WAN), or the connection can be made to an external computing device, such as through the Internet using an Internet Service Provider.

[0039] The above shows and describes the basic principles and main features of the present application and the advantages of the present application. Those skilled in the art should understand that the present application is not limited to the above embodiments, and the above embodiments and descriptions in the specification are only to illustrate the principles of the present application. Without departing from the spirit and scope of the present application, various changes and improvements can be made to the present application, and these changes and improvements all fall within the scope of the claimed present application. The scope of protection of the present application is defined by the appended claims and their equivalents.

Claims

1. A method for identifying trash racks in hydropower plants, characterized in that: The following steps are involved: S1. Acquire an image of the trash rack; S2. Classify the captured image using a classification model and output a classification result; S3. Based on the classification result, determining whether there is garbage in the captured image; S4. If it is determined that garbage exists, processing the captured image to obtain a foreground image and a background image; S5. Calculate the area of ​​the foreground image and compare the area with a preset threshold. If the area is greater than the preset threshold, generate an alarm message and send it to the terminal.

2. A method for identifying trash racks in a hydropower plant according to claim 1, characterized in that: In step S1, a camera is used to photograph the trash rack at a preset time period, to obtain the photographed images, and to upload the photographed images to a background system.

3. A method for identifying trash racks in a hydropower plant according to claim 1, characterized in that: The captured image is a top view image of the trash rack.

4. A method for identifying trash racks in a hydropower plant according to claim 1, characterized in that: The classification model in step S2 is a pre-trained convolutional neural network (CNN) model, and the pre-training step includes: S201, collecting historical images of trash racks and annotating the historical images to obtain a training data set; S202, dividing the data set into a training set and a validation set in a ratio of 8:2; S203: Train the CNN model based on the training set, verify the trained CNN model based on the verification set, and iteratively adjust model parameters until the CNN model converges.

5. The method for identifying trash racks in a hydropower plant according to claim 1, characterized in that: Step S4 processes the captured image and specifically includes the following steps: S401, cropping the captured image containing garbage; S402, calculating the background threshold of the cropped image; S403 : performing binarization processing on the cropped image based on the background threshold to obtain the background image and the foreground image.

6. A method for identifying trash racks in a hydropower plant according to claim 5, characterized in that: The background threshold The calculation of is as follows: Define the gray value of the pixel in the cropped image ,but: ;in Indicates the horizontal coordinate of the pixel The upper limit value of Indicates the vertical coordinate The upper limit value of .

7. A method for identifying trash racks in a hydropower plant according to claim 5, characterized in that: The calculation of the binarization process is as follows: ;in Indicates the pixel value after binarization.

8. A method for identifying trash racks in a hydropower plant according to claim 1, characterized in that ,The calculation area of ​​the foreground image includes counting the number of foreground pixels in the foreground image and calculating the area of ​​the foreground image based on the area of ​​a single pixel.

9. A trash rack identification system for a hydropower plant, characterized in that: The method according to any one of claims 1 to 8, comprising: An image acquisition module, used to obtain images taken at the trash rack; A classification processing module is used to receive the captured image, classify the image using a pre-built classification model, and output a classification result; a judgment module, configured to judge whether there is garbage in the captured image based on the classification result; An image processing module, configured to process the captured image and output a foreground image and a background image if the judging module determines that garbage exists; an area calculation module, configured to calculate the area of ​​a foreground image based on the foreground image; The alarm module is used to compare the area with a preset threshold, and if the area is greater than the preset threshold, generate an alarm message and send it to the terminal.

Images