SPEX: A Vision-Language Model for Land Cover Extraction on Spectral Remote Sensing Images
SPEX:一种用于光谱遥感图像土地覆盖提取的视觉-语言模型
机构 * College of Computer Science and Technology, Xinjiang University(新疆大学计算机科学与技术学院) ; iFlytek Co., Ltd.(iFlytek公司) ; National Engineering Research Center of Speech and Language Information Processing(语音与语言信息处理国家工程研究中心) ; School of Computer Science, Wuhan University(武汉大学计算机学院) ; Zhongguancun Academy(中关村学院) ; National Engineering Research Center for Multimedia Software(多媒体软件国家工程研究中心) ; Hubei Key Laboratory of Multimedia and Network Communication Engineering, Wuhan University(湖北省多媒体与网络通信工程重点实验室,武汉大学) ; State Key Laboratory of Information Engineering in Surveying, Mapping and Remote Sensing, Wuhan University(测绘遥感信息工程国家重点实验室,武汉大学)
AI总结 SPEX是一种专为光谱遥感图像土地覆盖提取设计的多模态视觉-语言模型,通过多尺度特征聚合和多光谱视觉预训练等技术,实现了精确的像素级解释和可解释的预测生成。
Comments Accepted to IEEE TGRS