Research On The Visual Matching Algorighm Of Automatic Wirebonder

Speed and accuracy of automatic wire bonder visual matching algorithm are key technical parameters for automatic wire bonder, directly determine the automatic degree of the automatic wire bonder and direct impact on the wire bonding speed, quality With the automatic wire bonder to a higher speed and better precision, the automatic wire bonder visual matching algorithm must meet higher requirements. In this paper, a new visual algorithm based on geometric pattern is introduced. This algorithm can achieve sub-pixel accuracy in high speed under different light contrast and great changed angle.With the image affine principle and the method of least squares, this paper builds a mathematical model for matching. According to this new model, the software system architecture is designed. This paper presents edge detection algorithm to meet sub-pixel precision and high speed requirements. Based on gray gradient algorithm, this edge detection algorithm includes low-pass filtering, Image enhancement and peak detection and sub-pixel edge detection to detect the edge sub-pixel location. This paper designs a pattern domain based on the characteristics of template edge points and the minimum distance vector. In this paper, the data structure of the pattern domain is build, and through series of operations such as seed module, link module, the pattern domain is constructed for the real-time matching. According to this mathematical model, this paper designs a real-time matching module and its sub modules, through fuzzy logic operations and various matrix operations to achieve matching speed and accuracy requirements.Finally, through experiments this paper verifies that this matching algorithm can achieve the proposed performance and can be applied to the automatic wire bonder.