Research on target recognition method of steel bar end face based on deep learning
The recognition accuracy of the traditional visual image analysis method for the end face of the steel bar depends on the intensity of illumination and the level of image pixels.However,the changeable and uneven illumination of the site environment will greatly reduce the recognition accuracy of the steel bar contour.The method of deep learning is adopted to establish a reliable im-age training model,so that the detection system has the ability to learn the steel bar contour image independently,and constantly opti-mize the training parameters,so as to achieve the effect of improving the recognition accuracy of the steel bar end face contour.
deep learningneural networkactivation functiontraining modelindependent learningKmeans algorithm
万方数据
维普
