Improving the measurement accuracy of an absolute imaging position encoder via a new edge detection method

Abstract

To improve the measurement accuracy of an absolute imaging position encoder, a novel edge detection method is developed to accurately detect the edges of grating lines based on empirical mode decomposition (EMD). First, according to the characteristics of the absolute position coding image (APCI), the vertical projection of the grey-level APCI is achieved by conventional grey-level transformation. Then, the vertical projection contour is decomposed by an EMD method and a finite number of intrinsic mode functions (IMFs) are achieved. Next, the last two IMFs are used to describe the trend of the projection contour. The trend is used as adaptive thresholds to divide the vertical projection contour into different intervals. Finally, the second derivate of the vertical projection contour is applied to obtain the accurate edges in each interval. Experimental results indicate that the proposed method is more robust to heavy noise, bad illumination and a slope compared with the existing methods with respect to edge detection. Also, the proposed method achieves better measurement accuracy than the existing methods.