Research On Automatic Recognition System Of Pointer Instrument Reading In Hydro-Station

The stability of power system is closely related to the safe operation of hydro-station,and the operation status of hydroelectric power system equipment is mainly reflected by real-time data of monitoring instrument,generally apply the way of manual daily patrol record to ensure the safe operation of hydro-station.Due to the poor anti-electromagnetic interference ability of digital display instrument,the hydro-station itself is a powerful source of electromagnetic interference,so in the complex electromagnetic environment,many pointer monitoring instruments are arranged in the hydro-station,and the accuracy of these pointer instruments reading recognition directly affects the safe operation of hydro-station,at present,the reading of pointer instruments is mostly recorded manually,low accuracy of manual reading,high labor intensity,and it is susceptible to factors such as illumination,climate and personnel fatigue work,these influence factors may bring hidden dangers to the safe operation of hydro-station.In view of the above problems,in order to avoid the error caused by manual record,reduce the work intensity of personnel,improve the accuracy and objectivity of pointer instrument reading and promote the automation and intellectualization level of hydro-station monitoring management,the paper will study the automatic identification system of pointer meter readings in hydro-station.The main research work is as follows:(1)Aiming at the noise of Gauss or salt and pepper exist in instrument image,the filtering methods of spatial and frequency domain denoising techniques are compared and analyzed by simulation experiments,and according to the experimental results,propos a denoising method based on median filtering with wavelet soft threshold denoising.When the instrument image local too dark or too bright,adopt the homomorphic filtering algorithm to enhance the image contrast,which lays a foundation for subsequent image processing,recognition and analysis.(2)In order to determine the core technologies of instrumentation imagefeature extraction and recognition,after research with analysis and the simulation experiments compared,adopt the hough transform method to extract the straight line of the instrument pointer and fit the circle of the instrument dial calibration distribution.According to calculating method for deflection angle of instrument pointer,adopt the angle calculating method to automatic calculate the instrument reading,through test inspection the accuracy and precision of the instrument reading calculation results,the results show that the recognition accuracy is about 92%,and the relative error is within 3%,these performances can basically satisfy the working requirements of daily inspections of hydro-station.(3)Comprehensive analyz the actual functional requirement of the automatic identification system of pointer instrument reading in hydro-station,the whole process ofautomatic recognition algorithm processing based on instrument reading,utilize the MatLab software platform to design and develop a graphical user interface processing system with simple interface,simple operation and strong orientation,then conduct the system function operation test.And the test results show that the system satisfy the design requirements,and can achieve accurate recognition of instrument reading,and there are no karton,delay and other phenomena in the operation of the system,therefore,it lays the foundation for the automatic identification of hydro-station pointer instrument reading and the informationization and intellectualization of hydro-station.