객체 분할 전처리 및 CNN을 이용한 너트 도금 불량 판별
Classification of Defective Plating of Nuts Using Object Segmentation Pre-processing and Convolutional Neural Network
2
조회
0
다운로드
0
인용
0
참고문헌
키워드
초록
It is required to apply computer vision instead of human visual decision in the final stage of production automation. In this paper, we propose a pre-processing of deep learning to improve the classification performance of plating defective nuts, which are representative mechanical parts. The proposed pre-processing algorithm improves classification performance by separating the nut, which is an object, from background in the camera image of the inspection system so that the background image does not affect to train deep learning. The pre-processing algorithm generates a mask image by processing sequentially binarization, edge detection, mathematical morphology filtering and 2D connectivity, and performs masking on the input image to segment the object from the background. It is selected VGG-16 among the widely used deep learning models to train with the proposed segmentation pre-processing in consideration of computational burden and accuracy. As simulation results, it is shown that the proposed pre-processing can be effectively applied to classify plating defective nuts.
Abstract
It is required to apply computer vision instead of human visual decision in the final stage of production automation. In this paper, we propose a pre-processing of deep learning to improve the classification performance of plating defective nuts, which are representative mechanical parts. The proposed pre-processing algorithm improves classification performance by separating the nut, which is an object, from background in the camera image of the inspection system so that the background image does not affect to train deep learning. The pre-processing algorithm generates a mask image by processing sequentially binarization, edge detection, mathematical morphology filtering and 2D connectivity, and performs masking on the input image to segment the object from the background. It is selected VGG-16 among the widely used deep learning models to train with the proposed segmentation pre-processing in consideration of computational burden and accuracy. As simulation results, it is shown that the proposed pre-processing can be effectively applied to classify plating defective nuts.
참고문헌 (8)
- Liu, L., "Deep learning for generic object detection : A survey", International Journal of Computer Vision, 128, pp.261~318 (2019)
- Al-Garadi, "A survey of machine and deep learning methods for internet of things(IoT)security", IEEE Communications Surveys & Tutorials, 22(3), pp.1646~1685 (2020)
- Feng, Di, "Deep multi-modal object detection and semantic segmentation for autonomous driving:Datasets, methods, and challenges", IEEE Transactions on Intelligent Transportation Systems, pp.1~20 (2020)
- LeCun, Yann, "Gradient based learning applied to document recognition", Proceedings of the IEEE, 86(11), pp.2278~2324 (1998)
- A. Krizhevsky, "ImageNet classification with deep convolutional neural net-works", NIPS (2012)
- karen, S., "Very deep convolutional networks for large-scale image recognition", Proceedings of the International Conference on Learning Representations, pp.1~14 (2015)
- C. Szegedy, "Going Deeper with Convolutions", CVPR (2015)
- K. He, "Deep Residual Learning for Image Recognition", CVPR (2016)