UAV remote sensing image Rural homestead detection based on deep learning

With the implementation of the rural revitalization strategy, the revitalization of rural homestead utilization is extremely important, master the regional location, quantity, size and distribution of rural homestead is a necessary prerequisite for the revitalization of rural homestead utilization. To detect rural homestead in UAV remote sensing image, it is necessary to seek a method that can automatically learn image features from a large number of samples and obtain the most effective feature information in the image, Deep learning is an algorithm that can automatically learn effective features from a large amount of data. In this paper, the SSD_MobileNet deep learning model is trained using massive UAV remote sensing data to realize intelligent detection of rural homestead. The experimental results show that the accuracy of the model is more than 90%, which can meet the needs of rural homestead detection in practical work.