This project aims, thanks to machine vision and artificial intelligence, to provide practical solutions in the domain of quality control in industrial manufacturing, pipeline inspection and weld radiographic testing. This research project is supporting digital processing for both fixed images and video sequences. The images can be obtained by CCD cameras or from the X-ray imaging modality, while the video sequences, obtained by endoscopes, concern fluid transportation pipelines of which the internal walls have to be inspected. Therefore, this project is articulated on the various steps found in machine vision chain. In our case, the used computer vision and artificial intelligence techniques aim to extract the visual information contained in an image or a video sequence in order to detect possible abnormality and then, proceed to their identification as defect or not. Among the techniques to implement, we can cite (1) defect image description by texture, color and shape, (2) defect detection by segmentation using statistical and/or deformable models (3) defect searching using Content Based Image Retrieval, (4) shape detection using the Radon transform and its generalized version, (5) defect detection and classification using machine learning techniques such as Deep Neural Networks (DNNs).