【摘要】 洪涝、地震等自然灾害发生突然,特别是对城郊设防水平较低地区,短时间内能造成巨大的损失。及时、准确、快速地获得足够的灾情信息是减灾救灾的前提;依靠无人机采集的灾区遥感影像越来越成为减灾部门提取第一手灾情信息的首选数据源。结合无人机影像特点,优化利用影像局部不变特征进行特征匹配,通过RANSAC算子剔除匹配粗差,并用变换优化法求取最佳变换模型参数,然后采用基于动态规划的最佳拼接缝搜索策略和加权平均相结合的融合策略,在保证灾害应急精度要求的前提下,很好地消除了拼接缝和"融合鬼影"现象,为城郊区应急情况下的灾情信息获取提供了新的技术手段和支持。更多还原

【Abstract】 It is known that some natural disasters(like earthquake,flood) can cause a great damage in short time.Getting enough disaster information timely,accurately and fast is the premise of disaster mitigation.Rely on UAV fast stitching technique of remote sensing data,for disaster prevention and mitigation departments,has become a more and more important first choice information source of first-hand disaster information extraction.The paper considers UAV to SIFT image characteristics feature matching process optimization,and by using transform optimization method for calculating the optimum transformation model parameters,based on combination of the best dynamic programming seam-line search strategy and weighted average’s fusion strategy.It is good to eliminate the joining together fusion ghost phenomenon.The paper provides reliable technical support for the benefit of the disaster information source.更多还原