Research On The Number Identification Of Pointers Based On Digital Features

As a universal measurement tool,pointer meter has many advantages such as convenient use and simple operation,especially in industry.In recent years,the accuracy of digital readouts by pointers has become a focus of increasing concern.The most effective technique for solving this problem is to replace human eye readings with automatic readings.Therefore,how to realize the automatic recognition of the number of pointer instruments has become an important topic for scholars at home and abroad.Therefore,this dissertation focuses on the automatic identification technology of pointer instruments,and introduces Snake algorithm and gravity method to achieve better recognition performance.First,using the improved Snake algorithm based on correction force,the dial and pointer of the pointer instrument are double-positioned.In this method,the Hough transform is used to coarsely position the disk,and then the Snake algorithm is used to precisely locate the instrument disk.Then the improved center of gravity method is used to realize the digital character recognition of the pointer instrument.The digital recognition method of pointer meter based on the improved gravity center method is implemented by three steps: improved center of gravity calculation,horizontal projection processing,and instrument expression reading,so as to obtain a more accurate pointer instrument display number and visually display it.Because of the use of digital features,the number of pointer instruments represents higher recognition accuracy and faster recognition speed.Secondly,in order to verify the effectiveness of the research work in this paper,experimental researches are carried out in two aspects: first,comparing the traditional Snake algorithm with the improved Snake algorithm proposed in this paper,the performance difference between the pointer instrument disc and the scale positioning;Second,compare the template matching method,neural network method and the improved center of gravity method proposed in this paper.The experimental results show that the improved Snake algorithm proposed in this paper can more accurately locate the pointer-type instrument disc and scale features than the traditional Sanke algorithm,and the processing speed is faster;the improved gravity center method proposed in this paper compares to template matching.Law,neural network method,with faster recognition speed and higher recognition accuracy.