A modified SSD method for Electronic Components Fast Recognition

Machine vision determines the machine's intelligence, and high reliability and robustness of recognition technology based on deep learning is the key to machine vision. Recently, accurate and rapid recognition model is the focus of researchers. In this paper, an advanced model based on deep learning was proposed, trying to explore good recognition performance for industrial environment. Firstly, an electronic component dataset (E.C.) is built. Secondly, an improved model based on SSD (single shot multi-box detector) is designed, which is conducted by adopting feature fusion strategy and adding visual reasoning techniques. Finally, the experiment on the dataset E.C. is completed, and the proposed method is compared with the current popular methods. The experiment shows that the detection accuracy and speed achieve better balance.