The identification of defects holds vital importance within the areas of image processing and computer vision, given their potential to compromise product quality and subsequently impact overall product performance. The proficient detection of defects facilitates informed manipulation of images, thereby augmenting visual perception and enhancing the precision of analytical processes. Concurrently, the cutting parameters pertinent to both turning and milling operations were allocated, encompassing cutting speed (m/min), cutter radius (mm), cutting depth (mm), and feed rate (mm/rev). Following this, texture analysis and defect detection were executed using deep learning algorithms. The model demonstrated substantial improvements throughout the training process. The model reached a training loss of 0.048, a training accuracy of 99.80%, a validation loss of 0.0091 and a validation accuracy of 99.65%. This research contributes to advancing quality assurance practices, especially in the automotive industry, through the integration of model explainability into defect detection systems.