The Decision-Making Technique Based on Computer Vision and Selection of Pareto-Optimal Alternative of Technological Production Parameters

The article proposes a decision-making technique in the field of rolling production process control based on the analysis of rolling tape surface image and the construction of technological parameter Pareto-optimal set. The substantiation of the automated process control system modernization scheme based on the introduction of a decision support system with computer vision for the rolling mill operator is presented. As a result of the technique development, the wavelet model of the rolling tape surface image is modified; the model presented in the pixel packet form; the classification information field has been expanded to improve recognition accuracy of rolling tape surface defects; the algorithm for forming the defect region and refining spatial features was developed; the algorithm for the construction of technological production parameter Pareto-optimal set has been developed. The application of the proposed technique made it possible to increase the accuracy of the rolling process control.