Depth of field extension method for visual measurement system based on coded aperture

Measurement error occurs when the machine vision system measures objects beyond the depth of field, resulting in a reduction of measurement accuracy. Aiming at the problem that the depth of field of vision system is difficult to meet the demand, a depth of field extension method for vision measurement system based on coded aperture is proposed. The method designs the number of equivalent edge transition pixels to evaluate the quality of images, establishes a multi-objective evaluation function to optimize the design of the coded aperture, proposes a regularization method based on L 0 prior to deblur defocus images, and builds the coded aperture visual measurement system. Based on the experimental results, compared with the circular aperture, the number of equivalent edge transition pixels of the image obtained by the proposed method decreases by $67.8 \%$, and the variance of the dimension solution in the vertical and horizontal directions decreases by $87.6 \%$ and $82.5 \%$, so the image quality and the stability of the dimension solution are improved.