Deep learning based vision inspection system for remanufacturing application

Nwankpa, Chigozie and Eze, Solomon and Ijomah, Winifred and Gachagan, Anthony and Marshall, Stephen (2019) Deep learning based vision inspection system for remanufacturing application. In: 4th International Conference on Remanufacturing (ICoR), 2019-06-23 - 2019-06-25.

[thumbnail of Nwankpa-etal-ICOR2019-Deep-learning-based-vision-inspection-system]
Text. Filename: Nwankpa_etal_ICOR2019_Deep_learning_based_vision_inspection_system.pdf
Final Published Version

Download (660kB)| Preview


Deep Learning has emerged as a state-of-the-art learning technique across a wide range of applications, including image recognition, localisation, natural language processing, prediction and forecasting systems. With significant applicability, Deep Learning is continually seeking other new fronts of applications for these techniques. This research is the first to apply Deep Learning algorithm to inspection in remanufacturing. Inspection is a key process in remanufacturing, which is currently an expensive manual operation in the remanufacturing process that depends on human operator expertise, in most cases. This research further proposes an automation framework based on Deep Learning algorithm for automating this inspection process. The proposed technique offers the potential to eliminate human factors in inspection, save cost, increase throughput and improve precision. This paper presents a novel vision-based inspection system on Deep Convolution Neural Network (DCNN) for three types of defects, namely pitting, surface abrasion and cracks by distinguishing between these surface defected parts. The materials used for this feasibility study were 100cm x 150cm mild steel plate material, purchased locally, and captured using a web webcam USB camera of 0.3 megapixels. The performance of this preliminary study indicates that the DCNN can classify with up to 100% accuracy on validation data and above 96% accuracy on a live video feed, by using 80% of the sample dataset for training and the remaining 20% for testing. Therefore, in the remanufacturing parts inspection, the DCNN approach has high potential as a method that could surpass the current technologies, especially for accuracy and speed. This preliminary study demonstrates that Deep Learning techniques have the potential to revolutionise inspection in remanufacturing. This research offers valuable insight into these opportunities, serving as a starting point for future applications of Deep Learning algorithms to remanufacturing.