An Ellipse Fitting with PSO for Automatic Reading Recognition of Pointer Instruments

Automatic reading recognition technology of pointer instruments based on computer vision is important for the development of industrial information technology. An ellipse fitting with Particle Swarm Optimization (PSO) for automatic reading recognition is designed to realize the pointer line detection and angle calculation for circular pointer instruments in this paper. Firstly, in order to eliminate the non-instrument arc segment in the instrument edge image, the concavities and convexities of the panel contour are classified by the gradient calculation of the edge image for the detected instrument; and then the residual elliptic arc is fit by PSO to overcome some adverse effects such as light, environment and noise on the contour detection image and to realize rapid smooth fitting for the instrument panel contour. Finally, the interment reading is determined by the measurement range and the angle between pointer straight line and scale line, according to the processed results of scale zone and pointer zone. Experimental results indicated that the proposed method is to realize fast and accurate reading detection of the circular pointer instrument without dependence on visual shooting angle.