The Image Recognition Based on Restricted Boltzmann Machine and Deep Learning Framework
Renshu Wang, Bin Chen, Jingdong Guo, Jing Zhao · 2019
For the Unmanned Aerial Vehicle (UAV) has high mobility, it is adopted to reducing the difficulty in the patrolling of electric power line is a hard work by manual way. However, the judgment of damaged poles is still carried out by the patroller which is inefficient and fallible. So the method with artificial intelligence is considered to be introduced that a novel method is proposed in this paper to improve the recognition effect in complex background. Restricted Boltzmann Machine (RBM) is used to instead the full connected layers of faster regions with convolutional neural network (faster RCNN). For RBM has the ability of unsupervised learning, with the RBM and faster RCNN combined, it can reduce the training samples and influence of different background in the images to be identified. The experimental results show that the proposed model takes effects on the recognition of the wire poles in the distribution network which has practical value.