Algorithm for the automated determination of heating areas during metal processing with numerically controlled machines

摘要

The article developed an approach to automated analysis images obtained in the visible and thermal imaging range by a system consisting of two separate cameras. For the generated images, a technique is proposed for the parallel application of algorithms for primary data processing (filtering and blur restoration method), searching for communication elements between data and performing geometric transformations over them, necessary for their combination. To identify elements with elevated temperatures, a primary (normal) thermal gradient mask is created. For simplicity, the resulting mask image is simplified. The simplification method is based on a nonlinear change in the image's bitness by analyzing histograms of color gradients. Next, a mask of threshold values for objects recorded by the optical system is formed. The paper proposes an algorithm for the averaged comparison of the received data and identifying areas where the values are exceeded. The advantage of the developed approach is its versatility. The use of one tuned system makes it possible to signal the cutting tool's wear, the presence of inclusions (inhomogeneities) in the workpiece part, wear of the bearings or elements of the robotic complex (machine tool). The set of test data provides an example of applying the proposed approach for analyzing the production process of cutting metal using a waterjet machine and milling a part on a numerically controlled machine.