Object Detection In Video Streaming Using Machine Learning And Cnn Techniques

Abstract

Object detection is affecting a lot of areas where images play a crucial role. It therefore becomes necessary to address this issue by building effective image detection in video streaming systems, that can detect such traces of objects that may cause harm for flight landing and take-off, either by identifying the object type, detecting image size or evaluating under different environmental conditions . There are a number of techniques like Background subtraction, similarity matching, convolutional neural networks, end-to-end feed forward neural network to detect objects in video streaming but with limitations of correctly identifying object in live video steam under illumination changes, non-stationary backgrounds and similar looking background pixels and foreground pixels form a complex background. Hence it is now required to develop system that requires less training and no human intervention for object detection in live video streaming. For this study we have taken the application of runways where providing minute object detection is necessary for flight safety.