Recognition Of The Analog Display Instrument Based On Deep Learning

With the rapid development of artificial intelligence,nowadays many traditional industries gradually from artificial transform into artificial intelligence.So it's intend to making the simple and repetitive working procedure to finish by artificial intelligence in production of industry.In the field of science and technology,the deep learning algorithm has been improved and applied in many fields,especially in image processing and pattern recognition.In this paper,a kind of intelligent system that is based on deep learning and applied to recognition of the analog display instrument designed.Main work of this paper is as follows.Firstly,locating instrument from the image gotten by camera that use the Faster-RCNN algorithm based on deep learning,and removing the interference information and make the subsequent image processing more easily.Next,according to the feature and geometric characteristics,using the connected domain analysis,image segmentation,and least squares method respectively,to locate meter center,pointer segmentation and refinement;Then,segmenting digital dial and scale,and recognizing them automatically by convolutional neural network LeNet-5 based on deep learning,which can combine with statistical calibration line to calculate the index value.At last,spreading the fan-shaped scale line into a rectangular range,according to the pointer position,it can get the reading eventually.This paper creatively proposes the method of using Faster-RCNN algorithm to locate the instrument,which can locate various types and sizes of the instrument in different images rapidly and accurately and has strong anti-interference.As for recognition,using algorithm of LeNet-5 to recognize digital and then calculate index value,it can make the system applicable to instrument of different range and uneven scale,which greatly improves the application range,and really realize recognition of the analog display instrument intelligently.