Final Year Project

Project Title

Crack Identification of Concrete Surfaces Using Convolution Neural Networks

Description

Automation is one of the rapidly emerging fields of research in the contemporary context. Our capstone design project is about automating the process of crack identification on roads. The algorithm works on convolution neural networks which is a part of deep (machine) learning. Crack detection on long patches of road proves to be a challenge if done manually. To counter this problem, we have developed a specialized rover that can navigate through large distances while acquiring images of the road. These images are then fed to our central processing system can identify cracks of different shapes, depths and lengths. Through GPS tagging, the exact location of the cracks can be provided to the maintenance crew. This is a joint project with Syed Muhammad Kazim Raza Rizvi and Ambreen Aslam.