Aiming at problems of low efficiency and low precision of two-dimensional（ 2 D） morphology measurement of large size workpiece

a 2 D mophology measurement system of large size workpiece is designed based on machine vision. This system mainly includes hardware system design and software system design. The camera position is precisely located by optical-electricity encoder. The large size workpiece is divided into several parts to be captured

and coordinate transformation math model is constructed

size measurement of large size workpiece is realized. A subpixel straight line/arc detection algorithm based on improved random sample consensus（ RANSAC） is designed

aiming at problem of low precision of normal linear/arc detection algorithm

which can get exact equation of the straight line/arc. Experimental results show that the system has high precision algorithm and good stability.