Abstract: Object Detection using Haar feature-based cascade classifiers is an effective object detection method, "Rapid Object Detection using a Boosted Cascade of Simple Features". It is a machine learning based approach where a cascade function is trained from a lot of positive and negative images. It is then used to detect objects in other images. Moving object identification and tracking motion is the base source to extract vital information regarding moving objects from sequences in continuous image-based surveillance systems. An advanced approach to motion detection for automatic image analysis has been presented in the paper. This achieves complete detection of moving object which is robust against of changes in brightness, dynamic variations in the surrounding environment and noise from the background. The proposed method is a pixel dependent and non-parametrized approach that is based on first frame to build the model. The detection of the foreground which represents the object and background which is the surrounding of the environment starts once the subsequent frame is captured. It utilizes unique tracking methodology that identifies and eliminates the ghost object from dissolving into the background of the frame. The proposed algorithm has been test implemented on several open source image by imposing single set of variables to overcome shortcomings of relevant and recently developed techniques.

Keywords: Object Tracking, Motion Detection, Background Subtraction, Normalized Cut Segmentation, Video Surveillance, etc.

PDF | DOI: 10.17148/IJARCCE.2020.9525

Open chat
Chat with IJARCCE