Objective: In order to realize edge detection for micro accessory

a new edge detecting algorithm of micro accessory is proposed based on the process characteristic of micro accessory. Method: First

this algorithm extracts the transition region of different parts. The polynomial regression mathematics models of transition region in the different parts are built and then the accurate position is got by process character from the mathematics model. Result: Experimental results indicate that the measurement error of 1mm standard gauge block is 0.395μm. Conclusion: It draws that the process characteristic has greater influence for the edge of micro accessory and so the edge detection of micro accessory should consider the influence of process characteristic by the common process methods—lapping

stamping and wire-cutting. The actual processing condition and statistics are added in this algorithm. The edge detection result can reach the inferior pixel level by building the mathematics model of transition region.