Precise positioning of mark point based on the improved cramer-rao lower bound algorithm

The rapid positioning algorithm of the Surface Mount Technology (SMT) circuit board in high-speed SMT production is not so precise and its adaptability is poor. Thus, a new circular mark point positioning algorithm was proposed to solve rapid positioning in SMT circuit board in this study. Based on the image features, this algorithm was used to implement the initial segmentation. A group of dominant point set off the edge was extracted based on the canny operator with the introduction to Gaussian kernels of multi-scale space. Circle fitting was implemented by using the improved Cramer-Rao lower bound (CRLB) operators. Thus the mark center is located accurately and the precise positioning of the SMT circuit board is ensured. The high-precision surface mounting of circuit board components was implemented, which improves the quality of SMT products. In addition, the practicability, accuracy and the robustness of the algorithm were verified through the application of the algorithm in actual SMT production. The algorithm has a higher detection accuracy compared with the Linear Spectral Frequency (LSF) algorithm and Adaptive Filter (AF) algorithm. This algorithm speed is 1.3 times faster than hough algorithm and LSF algorithm. Copyright © 2016 American Scientific Publishers All rights reserved.