The real-time performance of a universal thresholding algorithm formachine vision with nonstructured light

摘要

The extraction of useful image data is achieved by a nonlinearthresholding technique which has been shown to yield useful practicalresults in a number of different applications through an ad-hocexperimental approach to the establishment of the appropriatethresholding algorithm. The paper illustrates the systematicestablishment of the nonlinear algorithm and the resulting real-timeperformance of the optimal threshold function for an input image with asimple sample object. The use of the method for inspection applicationsis discussed briefly