Abstract
Abstract. Detection of moving objects in video sequence received from moving video sensor is a one of the most important problem in computer vision. The main purpose of this work is developing set of algorithms, which can detect and track moving objects in real time computer vision system. This set includes three main parts: the algorithm for estimation and compensation of geometric transformations of images, an algorithm for detection of moving objects, an algorithm to tracking of the detected objects and prediction their position. The results can be claimed to create onboard vision systems of aircraft, including those relating to small and unmanned aircraft.
Highlights
In this paper we present our solution to moving object detection and multiple object tracking tasks
Because of our set of algorithm is developed for onboard vision system it can works in some conditions: low contrast of objects of interest in relation to the background, inhomogeneity of the optical medium and various atmospheric phenomena, crossing trajectories of tracking objects, occlusion of objects with elements of the background, quickly changes of the observed scene
The estimation end compensation of geometrical transform are performed in two steps: evaluation and compensation of rotation, evaluation and compensation of shift. The advantages of this approach are subpixel precision, acceptable computational complexity. This algorithm based on phase correlation allows us to estimate the vector of parameters of geometrical transformations with a sufficiently high accuracy, which allow us to obtain background estimation to solve the task of detection of moving objects
Summary
In this paper we present our solution to moving object detection and multiple object tracking tasks. This set of algorithm is developed to work on onboard video system. Exist some object detection algorithms based on colour information (Lefevre, 2003), (Gorry, 2007), but usually we can’t use colour image in onboard technical vision systems because of limited computing resources. There are approaches to detection objects using machine learning based algorithms (Papageorgiou, 1998) To use such approaches, it is necessary to create a training samples, which is not always easy to do because of the wide variety of objects of interest and observation conditions. The set of algorithm composed from three main part: algorithm for estimation and compensation of geometric transformations of images, algorithm for detecting of moving objects and algorithm for object tracking and predicting their position in case of occlusion
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
