A Capstone Project submitted in partial fulfillment of the requirements for the degree of Bachelor of Science in Engineering (Mechatronics), School of Engineering and Technology

For 3D reconstruction the mostly used methods are the stereoscopic vision and the Kinect sensor. Even though another method was introduced in early 90s which is 3D Reconstruction by zooming, it has not shown much progress due to the lack of 3D information in the Zooming method. Therefore, the number of researches that has been conducted in this particular area is scarce. For the senior Capstone project I have selected the topic of 3D Reconstruction by motorized Zooming lenses, in order to research further about the interested area and there by find a better way of obtaining the 3D information by Zooming lenses using monocular vision. To obtain this purpose, a single fixed position camera has been used. The camera is capable of taking a set of images by varying the focal length of the lenses. And the other process consists of face detection using OpenCV Haar cascade classifier, Feature detection of the captured images by using Surf descriptor. The detected features has been mapped by Flann base matcher technique. RGB intensity base algorithm has been used as a reconstruction technique. Finally OpenGL techniques has been used to render the final 3D model.

Capstone Project (B. Sc.) - Asian Institute of Technology, 2014