Review of edge detection algorithms for application in miniature dimension measurement modules

Abstract

Continuous improvement of semiconductor chip technology allows more and more complex functions to be performed by an increasing number of modules with limited space and power. The deep miniaturisation and cost reduction of such modules has a positive impact on their application areas. They can be used, for example, inside machine tools for dimensional inspection of workpieces or deformation of the tool tip. Modules of this type must contain as many standard components as possible and, in particular, a processing unit responsible, among other things, for processing the optical sensor signal. For this reason, this study reviews and compares algorithms for object edge detection useful in embedded systems based on standard single chips, cores based on Cortex-M4F. The complexity of processing and the quality of algorithm output data was analysed. The results of experimental studies are presented. It was found that the required processing times significantly reduce the use of single chip embedded systems only for edge detection on small images.