Development And Realization Of Distribution Network Inspection Instrument Based On Vision System

The distribution network is an extremely important part of the transmission network.The daily inspection and maintenance of the distribution network is one of the important measures to ensure its long-term operation and stability.Increasing the automation ratio of the distribution network is a problem that the current power grid companys need to solve.However,in the current inspection process of the distribution network,my country is still in a relatively artificial environment,and the recording work of inspection often requires manual transcription and registration,and the inspection efficiency is low.Based on the above problems,in order to meet the actual needs of personnel,reduce the pressure of manual inspection,ensure efficiency and scientificity,and improve the degree of automation of distribution network,a distribution network inspection instrument based on visual system was designed.The specific research contents and results are as follows:1.According to the environment and conditions of the inspection instrument,the overall design of the system is carried out.The inspection instrument has the functions of infrared temperature measurement for the distribution network equipment and defect detection for the insulator,real-time data storage,Wi-Fi wireless transmission,and Control via Android APP.The patrol instrument designed in this topic is installed in the helmet,which is compact and convenient for the staff to carry.2.Designed the hardware circuit of the inspection instrument.The inspection instrument terminal uses a high-performance STM32F429 microcontroller as the main control chip,equipped with the RT-Thread real-time operating system developed by the Chinese open source community,which can simultaneously process tasks in multiple threads in real time,reducing the accumulation of tasks and queues in the work,Thus effectively avoiding the phenomenon of operating system stalls during multi-task high-load processing.In addition,in terms of images,it supports the collection of visible light images and is also equipped with an infrared shooting function,which can effectively detect abnormal heating and other failures that cannot be detected by the naked eye,and improve the accuracy and efficiency of inspections.With the help of high-speed Wi-Fi module and TCP protocol,wireless transmission not only meets the requirements of video transmission rate during operation,but also ensures the integrity and stability of data transmission.3.The software system used by the inspectors is designed,and the APP used on the Android device platform is developed to display the data transmitted by the data collection system in real time.When using an ordinary shooting system,the photographer needs to observe the image through the screen on the entire hardware,which brings great inconvenience to the staff.However,the APP mode allows the staff to observe the shooting effect more conveniently.The defect detection algorithm based on the convolutional neural network is transplanted into the Android APP,and the defect detection can be performed on the insulator during the inspection process,so that the function is also more perfect.4.Conducted defect detection research based on convolutional neural network,using defect detection algorithm based on convolutional neural network to conduct defect detection on insulator devices common in distribution network.By improving the SSD framework,it elaborated how the design of the MoblieNet network can reduce the amount of calculations and parameters,realize the lightweight of the network,and facilitate the transplantation of the algorithm to the Android APP,so as to realize the defect detection of the insulator image collected by the terminal of the inspection instrument.5.Complete the overall test of the system.The relevant test results show that the application of the insulator crack detection algorithm is normal,the infrared temperature measurement test results are normal,the inspection database is operating normally,and the Wi-Fi network is running stable.The operation test of each part shows that all functions can work normally,complete the design requirements of the system,and can better complete the collection and transmission tasks.