Industrial Automatic Control Systems and Controllers Annotation << Back Extracting Vehicle Candidates from a Sequence of Digital Scene Shots V.S. Tormozov The article considers the current task of extracting and counting candidates of vehicles for the sequence of digital pictures of the scene of the transport section of the road network (RN). The relevance is based on the applicability of the developed solution to reduce the workload of the RN of cities, the implementation of the traffic jam information system, and the algorithmic and software support of transport systems, which is of concern to the administration of many modern cities. The invention proposes a modified algorithm for extraction of vehicle candidates diff ering from existing extraction algorithms by a clustering procedure of deterministic descriptors of candidate objects. The main stages of vehicle detection are considered. The research results given in the article depict the principal applicability of the proposed algorithms for the practical task of collecting information on the characteristics of RN traffic flows. Keywords: candidate extraction; candidate localization; contour analysis; digital image processing; machine vision; intelligent transport system. DOI: 10.25791/asu.1.2021.1249 Pp. 21-27.