Object Detection Methods for Improving Smart City Safety

Abstract

The safety measures in a city are of major concern to its people. A smart surveillance system is necessary to make the safety measures robust and reliable. Currently, there are many object detection methods available for image analysis. Image analysis is only possible when we can detect the number of objects in the image, type of objects, as well as their location. Image analysis methods include Fast RCNN, YOLO, region-based convolution neural network (RCNN), and so on. These object detection methods are based on deep learning (DL) techniques. Video analysis requires the detection of moving objects. These methods involve background subtraction and extracting the foreground objects for motion analysis. There are several deep learning (DL)-based methods for video analysis such as the YOLO algorithm for detecting objects in a video frame and identifying their location. This chapter describes significant object detection methods in images and video that can be used in surveillance systems to improve the safety measures in a city.