Intelligent Object Detection and Extraction Method for Surveillance Applications in Smart City

Abstract

Detecting and segmenting moving objects are advanced capacity required by greater part of machine learning, artificial intelligence and computer vision. Presently, there are numerous object extraction methods have been looked into many years and equipment devices improved for extracting moving objects without shadow and ghost artifacts. However, an assortment of issues exists, for example, shadows pixels also segmented as foreground object regions. In this paper, we suggest a robust and simple automatic object detection framework for indoor environment based on image enhancement and background subtraction. Extracting dynamic objects in video sequences are a challenging task in smart vision applications and machine learning research area. Our proposed method could help to detect human shape rapidly and accurately in urban surveillance systems without any noise. The method is implemented and tested widely used datasets and our system proved to achieve good performance solving object segmentation problems in indoor and outdoor scenes for delivering object identification and tracking tasks.