Road surface diversion analysis while driving an autonomous vehicle

The main purpose of the study was to solve the problem of analyzing the road surface diversion in real time mode while driving an autonomous vehicle using machine vision combined with light line projection. The objects of the study were the machine vision video images and the subjects were the parameters and its processing algorithms. The results of assessing the deviation of the road surface from the horizontal level are shown based on processing of video images of the specific segments of a scale open-pit road model in laboratory conditions. For each specific segment, such as the beginnings and endings of longitudinal and transverse slopes, specific conditions by which the machine vision system processes changes in the geometry of the light line projections. The values of the perimeter and geometric shapes of the light line projection, as well as the values of their curvature and the degree of their pattern matching were used as the main parameters. The Shi-Tomasi angle detection algorithm was used to calculate the perimeter. A block diagram and a description of the software module functions are presented for analyzing the road surface diversion. The obtained results in the form of software and the underlying algorithms can be used to solve industrial problems concerned with controlling autonomous vehicles.