Research on Temporary Building Monitoring of Railway Construction Based on Remote Sensing Image

The demolition of temporary buildings is an important part of the railway construction environmental protection work. At present, the artificial field survey method has the problems of inefficiency and heavy tasks, and the difficult environment makes monitoring difficult. Aiming at the above problems, a method for detecting temporary building demolition based on remote sensing image is proposed. We use remote sensing technology to automatically monitor whether the adjacent buildings in each area are demolished on time to facilitate effective management of temporary land use ecological restoration. In this paper, we used the U-net network model, which is widely used in medical image segmentation, and its structure has been improved. First, the data set is created by cutting, denoising and marking the semantic information of the remote sensing image. The data set is divided into a training set and a test set, and the training set is enhanced. Then, the training set is input into the improved U-net model for training, which integrates multi-scale enhanced training and multiple loss calculation strategies. Finally, the prediction results are post-morphologically processed to obtain the final result, and detect the change. The experimental results show that the method achieves IoU, pixel accuracy and Kappa coefficient of 82.9%, 90.5% and 81.2%, which is better than FCN and U-net methods.