Research On Image Recognition System Under Low Light Based On Embedded Linux

Because of large amount of varieties and in quantity, Anolog measurement instruments have become one of the most popular measurement instruments in the industry field. For example, piezometer, centesimal meter, voltage meter, amperemeter and so on.This thesis has realized meter image capture system based on embedded Linux and uses many grayimage processing technology, such as gray processing, image filtering, enhancement, edge detect and Hough processing, to complete the processing of image under low light and recognize pointer angle automatically.Firstly,this paper briefly introduces the image capture system's realization scheme, then presents characteristic of the embedded system development.Based on this, the whole system design is divided into hardware design and software design two parts. In hardware design part, explains especially the interface of S3C2410 CPU chip with memory and network controller chip, In software design part, discusses USB CMOS image sensor driver realization and how to use VideoLinux API functions to realize image capturing.Secondly,presents especially image processing and recognition means of the voltage meter pointer position. Be aimed at the flaw of the conventional image recognition method under better illumination condition, the article puts forwards a new method, which is applicable for identifying pointer area under bad illumination condition.Through a great deal of different conditional experiments,this method has many advantages,such as simple computing,good applicability and good processing result. This method solves the problem of identifying pointer area under bad illumination condition and establishes the basic for identifying the pointer position and computing the pointer angle.