Fault detection is considered an important and challenging task to be incorporated in many industrial applications. It has gained interest in recent years, and many techniques have been proposed for developing an effective fault detection approach due to its significant importance in everyday life. This study presents an automated intelligent fault detection technique incorporating image processing and fuzzy logic. Image processing is the first step where features such as entropy estimation, color-based segmentation and depth estimation from gradients are obtained. The extracted features (number of {blobs, minima, maxima}, and estimated entropy) act as input to the fuzzy logic. The subsequent step incorporates fuzzy logic; the four inputs are fed to fuzzy which extract the fault and acts as knowledge rule-based tool and final step, i.e. the output generation, classifies it accordingly into four categories of faults (rust, bumps, hole, wrinkles/roller marks). The proposed method is compared with Linear Vector Quantization, and Multivariate Discriminant Function approaches. The method is tested on a database of 150 images. The proposed method demonstrated its significance and effectiveness with performance accuracy of 99%, 98%, 96.8% and 97.6% for rust, bumps, holes and wrinkles/roller marks respectively.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.