智能技术学报

文章详情

稿件标题: DNN-Based Joint Classication for Multi-source Image Change Detection
稿件作者: Wenping Ma, Zhizhou Li, Puzhao Zhang, Tianyu Hu, Yue Wu
关键字词: change detection; multi-source image;deep neural network-s; feature learning
文章摘要: Multi-source change detection is an increasingly presented issue and it is of great significance in environmental and land exploration. Multi-source remote sensing images are obtained by different sensors, which usually are not completely consistent in terms of spatial resolution, spectral bands number in the same region. In this paper, we propose a novel joint classification frame work for multi-source image change detection, the multi-source image-pair are generated by different sensors, such as optical sensor and synthetic aperture radar, respectively. This frame-work is established for feature learning, which is based on deep neural networks. Firstly, in order to segment the optical image, deep neural networks are essential to extract deep features for clustering segmentation. Then the stacked denoising autoencoders are used to learn the capability of classification by training the reliable training examples, which are selected from optical image segmentation results that are uncha
收录刊物: 2017年2卷2期
稿件基金:
浏览次数: 101
下载次数: 45
点击下载