Research On Remote Reading Recognition And Location Of Single Pointer And Digital Instrument

In substation,in order to ensure the safety and stability of power supply,it is necessary to maintain the equipment daily.There are many shortcomings in using traditional manual methods to detect the instruments of substation.In order to realize the automation management of substation and reduce the burden of personnel,the patrol robot is used to collect remote reading data of instrument,and realize the daily maintenance and timely warning of equipment.In substations,there are many types of instrumentation equipment,most of which are single pointer and digital instruments.It is of great significance to carry out the research on the reading recognition of pointer and digital instrument.However,in the substation,the environment is complex for robots.The collected instrumentation images are affected by the robot’s attitude,illumination,object occlusion,dial stain,lens blurring and so on,which is a huge challenge for remote instrument recognition.The main contributions of this paper are as follows:1.Research on location technology of remote instrument recognition method.Mobile inspection robot needs to automatically identify the instrument to be tested when it carries out autonomous routine inspection of the instrument.In the complex environment,the problem of object detection under the influence of distance,attitude,illumination,object occlusion,focusing blur,noise and other factors is considered.The method of coarse positioning and fine positioning is adopted to improve the success rate of instrument recognition.2.To improve the accuracy of image registration,a feature point matching algorithm combining SURF feature points with LSD line segment features is adopted.3.A binarization method based on edge gray surface fitting is proposed.The experimental results show that the algorithm has good robustness to illumination and can effectively reduce the influence of illumination change on the binarization of instrument surface.4.A character segmentation algorithm is proposed,which can effectively solve the problem of character segmentation such as character cohesion and character oversegmentation,and has a certain anti-interference ability to noise.