Inspection, and its processes, is a fundamental step to ensure product quality in our modern industry. Traditional computer vision (CV) methods provide promising solutions in this regard, but the accuracy and speed of these inspection processes depend largely on rigorous implementation that is highly sensitive its traced context and well-defined situations. Additionally, these technologies demonstrate limitations when used in dynamic industrial environments, as manufacturing conditions and variability in the appearance of defects affect the accuracy of their detections/classifications. With the development of machine learning (ML), models have seen various improvements such as a continuous increase in performance along with a reduction in the need for manual adjustments which in turn yields a much greater capacity for generalization. This study aims to constitute a brief review and analysis of existing CV and ML methods applied to defect detection/classification applied to industry, more specifically detection/classification of defects in wood-derived boards such as melamine-faced chipboards (MFC) and medium-density fiberboards (MDF), with a larger focus on the former. In addition, it addresses advanced Automated Optical Inspection (AIO) systems applied to the detection/classification of this type of defects.