Preview
摘要：A Remote Sensing Image Semantic Segmentation Method by Combining De formable Convolution with Conditional Random Fields Zongcheng ZUO1,Wen ZHANG2,Dongying ZHANG31.Schood of Aeronautics and Astronoutics,Shanghai Jiao Tong Uniberily。Shanghai 200240,China;2.School of Remole Sensing and Informarion Engineering,Wuhan Unieriy,Wushan 430079,China;3.Schood of Hydrppower and Information Eninering.Huashong Uninersity of Science and Technologr,Wuhan 430074,China.Abstract:Curently,deep convolutional neunal networks have made great progres in the field of semantie segmentation.Because of the fixed convolution kemel geometry,standard convolution neural networks have been limited in the ability to simulate geomnetric transformations.Therefore。a deformable convolution is introduced to enhance the adaptability of coovolutional neheorks to spatial transformation.
关键词：CONVOLUTIONring.progr
分类号： O17[理学—基础数学]
- Near-zone Direct and Indirect Topographic Effects Based on the Rectangular Prism and Surface Element
- Satellite Positioning and Orbit Determination System(SPODS):Introduction and Evaluation
- Solar Radiation Pressure Modeling and Application of BDS Satellites
- Journal of Geodesy and Geoinformation Science Aims and Scopes
- Object Detection Research of SAR Image Using Improved Faster Region-Based Convolutional Neural Network
- Parameter Group Optimization by Combining CUBE with Surface Filtering and Its Application
- MHSS ARAIM Algorithm Combined with Gross Error Detection
- Automatic Matching of Multi-scale Road Networks under the Constraints of Smaller Scale Road Meshes
- Applicability of Bevis Formula at Different Height Levels and Global Weighted Mean Temperature Model Based on Near-earth Atmospheric Temperature
- Satellite Image Matching Method Based on Deep Convolutional Neural Network