THESIS TYPE: INDUSTRY, EXPERIMENTAL HARDWARE, EXPERIMENTAL SOFTWARE, EXPERIMENTAL – DESIGN, APPLIED EXPERIMENTAL, COMPANT EXPERIMENTAL

DESCRIPTION: The objective of the thesis is the creation of an intelligent system for the automatic analysis of car body demages. The system will receive structured light images of the different parts of the vehicle (e.g. bonnet, roof, mudguard) and will have to identify, through deep learning, the defects of the bodywork. An integral part of the thesis will be the validation of the system performance when the acquisition conditions vary (e.g. different levels of brightness, intensity or color of the structured light) as well as the validation of the optical acquisition system.

KEYWORDS: ARTIFICIAL INTELLIGENCE, AUTOMOTIVE, DEEP LEARNING, VIDEO ANALYSIS, DEEP NEURAL NETWORKS, HAIL, IDENTIFICATION, IDENTIFICATION OF DAMAGE, IMAGE PROCESSING, NEURAL NETWORKS, ARTIFICIAL NEURAL NETWORKS, CONVOLUTIONAL NEURAL NETWORKS, CONVOLUTIONAL NEURAL NETWORKS