An image-based intelligent system for pointer instrument reading
Zhijuan Yang, Wennan Niu, Xiaojiang Peng, Yongqiang Gao, Yu Qiao, Yi Dai · 2014
This paper proposes an image-based system for pointer instrument reading (PIR). The main contributions are twofold. First, we introduce binary descriptor such as ORB(Oriented FAST and Rotated BRIEF) for registering the images of pointer instruments. Second, we propose a Circle-based Regional cumulative Histogram (CRH) method to finely locate the pointer. The CRH method is robust to illumination, and is relative fast compared to previous methods. Experimental results demonstrate the effectiveness and practicability of our image-based PIR system.