59958

Detecção de Mudança para Identificação de Objetos Móveis em Vídeo UHD

Favoritar este trabalho

Several sensors provide a very large range of data with different characteristics. In this line, this work aims to demonstrate an application for Data Fusion. An algorithm in OpenCV designed to detect moving objects in Ultra HD video purchased by Iris camera, installed on the Zvezda module of the International Space Station (ISS). The study area is located in Vancouver, Canada. The images are from Deimos-2 satellite in 1C level. The first stage of the research was to detect moving objects between successive frames of high resolution video. The detection of moving objects resulted in new images containing motion information, thus it was possible to separate the moving objects in the scene. In order to highlight the changes were applied one thresholding in the change detection image creating a binary image. So as to reduce the possibility of false positives, especially those generated by the movement of the sensor, was applied a morphological operation erosion. Finally, a new video was created showing the moving objects over time, allowing monitoring of these objects. In the second stage, there was the match between the georeferenced panchromatic image and the video frames of moving object, which allowed the identification of the object position associated with the each frame of video to monitor the displacement of objects.