ABSTRACT:- The detection and recognition of objects is very important and challenging task in computer vision, as there is an increasing interest in building autonomous mobile system. To make mobile service robot truly ubiquitous to complex and dynamic indoor environments, they should be able to understand the surrounding environment beyond the ability to avoid obstacles, autonomously navigate, or build maps. Several researchers have proposed different techniques for recognizing and localizing indoor objects for mobile robot navigation. Different object detection algorithms from the early computer vision approaches until recent deep learning approaches are reviewed and compared. Based on the type of feature used the algorithms are classified in to two. The first class is based on a local feature like SIFT, SURF, etc. this class also includes the techniques that fuse these local features with 3D or Depth images. The second class is based on deep features, using deep neural networks for detecting objects in the images which id further classified into two based on whether these algorithms use one or two stages for detecting the object. Object detection for mobile robot navigation can be used for Assisting an elderly person or person with disability (Blind) to navigate in indoor environment, Home security and surveillance.

Keywords: Object Detection, Deep Learning, Object Localization, Mobile Robot Navigation, Semantic Navigation.