INTEGRATED RECOGNITION OF TEXTURE'S IMAGES OF WORN SURFACES AND DAMAGED CUTTING TOOLS
Abstract
Algorithms to the formation of descriptions of texture images, based on the construction of co-occurrence matrices, Fourier analysis of the energy spectrum and Markov random fields were considered. A comparative analysis of algorithms for the implementation of statistical and spectral approaches with the recognition of surfaces worn or damaged cutting tools was performed. Recommendations for parameter setting and the joint application of the considered algorithms were given. Accuracy of recognition was improved.