Abstract
The aim of the work is to study the effectiveness of algorithms for detecting objects of interest in images obtained by remote surveillance systems. The main properties of objects in relation to the background are their compactness and isolation. The algorithms in question use multithresholding, which yields a set of binary slices. It allows to carry out morphological processing of objects on each slice for the purpose of the analysis of their geometrical characteristics and the subsequent selection. As a result, an adaptive object detection threshold can be set. The use of area selection can significantly reduce the number of false alarms during detection and use lower‑level thresholds, which increases the probability of correct detection of objects of interest. Results of processing of test models of images, and also results of detection of objects on the real image showing efficiency of the considered algorithm are presented.
Highlights
Введение Задачи обнаружения, выделения и локализации объектов интереса на изображениях различной при‐ роды весьма актуальны и интенсивно исследуются
The aim of the work is to study the effectiveness of algorithms for detecting objects of interest in images obtained by remote surveillance systems
The algorithms in question use multithresholding, which yields a set of binary slices
Summary
Ю. Обнаружение объектов на изображениях с использованием селекции по площади // Вопросы радиоэлектроники. АО «Научно-производственное предприятие «Радар ммс», 2 Санкт-Петербургский государственный электротехнический университет «ЛЭТИ»
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
