Defect detection of automotive leather based on Nanodet-Plus

During the processing of automotive leather, various defects such as folding and oil stains can occur for several reasons. Therefore, it is essential to detect such defects before leather manufacturing to prevent them and enhance the quality of leather products. However, the current defect detection methods are not only expensive but also inefficient, indicating a need for a more efficient alternative. To address this problem, we constructed an Automotive-leather-defect dataset for vehicles based on collected leather defect samples. Subsequently, we proposed a defect detection method using the NanoDet-Plus neural network model, which can enhance the accuracy and speed of the detection process.