Detección Y Seguimiento de Personas Basado en Vectores de Movimiento

Abstract

Nowadays the security requirements demands motion detection systems more reliable, precise, easy to handle and to able to adapt in any scene or monitoring area. Image processing area has been great advances and it makes possible the creation of algorithms and systems which can satisfy these requirements. However, complexity, high cost of acquisition or the limitation in their operation make these systems to have great number of false alarms. However the advances of the digital technology during the last several years have allowed important improvements in the development of security and access control systems by means of images and video processing. This fact has done it possible the development of low cost and efficient motion detection and tracking systems for applications that require access control and monitoring of access restricted places. Because the necessity of low cost and high performance motion detection and tracking systems has increased during the last years, several systems have been in the literature, based on the analysis on the information contained in image sequences to detect the movement of persons and vehicles, etc. This paper proposes a human and vehicle motion detection and tracking system, based on motion vectors, which is able to discriminate between allowed and not allowed movements without the intervention of any human operator. Evaluation results show the desirable features of proposed method.