Neural network vision of an intelligent non-destructive evaluation autonomous vehicle

Automated inspection of aircraft and spacecraft fasteners is of great importance in the evaluation for possible fracture or failure on the craft's fuselage before they are launched and between missions. The purpose of this thesis is to study an appropriate algorithm that can potentially be implemented to guide any of the NDE vehicles previously created in private and government institutions. The algorithm is based on the current understanding of AI. Hence, Neural Networks in conjunction with sequential computation has been used for image acquisition and recognition of the fasteners. The purpose of the NN algorithm is to serve as vision system to guide autonomous vehicles to the rivets position on an aircraft surface. Although the actual test is beyond the scope of this work, the ultimate goal is to contribute in the creation a fully automated vehicle that can reliably test for damage.