Mirror-like workpiece surface defect detection method and device based on multiple illumination conditions

The invention discloses a mirror-like workpiece surface defect detection method and device based on multiple illumination conditions. The method comprises the following steps: detecting a first position based on white light to obtain a first image, and detecting the first position based on cosine fringe light to obtain a second image. Processing the first image based on a first detection model to identify a non-deformation defect, and obtaining the type, coordinate and size of the non-deformation defect; and meanwhile, processing the second image based on a phase deviation conversion method to determine whether a deformation defect exists or not. And if the deformation defect exists, determining a defect area, and triggering binary stripe light to carry out area expansion detection on the defect area so as to obtain a third image. And processing the third image based on the second detection model to identify the deformation defect, and obtaining the type, coordinate and size of the deformation defect. And rep