Development and implementation of vision-based techniques for target identification

• Main Author: Thenmugilan Gandhy
• Other Authors: Erdal Kayacan
• Format: Final Year Project
• Language: English
• Published: 2015
• Subjects: DRNTU::Engineering::Computer science and engineering::Computing methodologies::Image processing and computer vision; DRNTU::Engineering::Computer science and engineering::Computing methodologies::Pattern recognition
• Online Access: http://hdl.handle.net/10356/64607
• Institution: Nanyang Technological University

Object detection is one of the complex problems in the area of computer vision. As new developments are coming through, application of object detection and localisation extends to multiple fields. One of its applications would be aerial surveillance by processing real time images to detect and locate a specific object. In this project, an aerial boat detection system is imitated and an algorithm for a simple boat detection system is presented. In this system, SURF (Speeded Up Robust Features) detector, bag of keypoints and support vector machine (SVM) algorithms are utilized. Development of boat detection system is made up of training phase and detection phase. Experimental evaluations show that system can detect and locate a floating paper boat in an image with an average accuracy of 90.1%. The system has been optimized by varying SVM’s parameters and the results are presented in this report.