基于概率统计模型的遥感影像变化检测

基金项目: 国家863计划资助项目(2006AA12Z136)；国家教育部博士点专项基金资助项目(20060486041)

详细信息

Change Detection of Multi-time Remote Sensing Images Based on Statistics Models

¹School of Remote Sensing and Information Engineering,Wuhan University,129 Luoyu Road,Wuhan 430079,China

Funds: 国家863计划资助项目(2006AA12Z136)；国家教育部博士点专项基金资助项目(20060486041)

摘要: 基于概率统计理论,提出了一种基于t检验的遥感影像自动变化检测方法,并实现了其与相关系数法的有效结合,运用于复杂城区环境下地物类型的自动变化检测。对于高空间分辨率影像,有效引入纹理特征,减少了房屋阴影的影响,对最终变化检测结果起到了明显的增强效果。实验采用不同类型的数据,详细叙述了变化目标的提取以及本文方法的特点,结果真实反映了实际地物的变化,表明该方法具有很好的实用价值。
- 变化检测 /
- 遥感 /
- t检验 /
- 相关系数 /
- 纹理
Abstract: An approach for change detection of multi-time remote sensing images based on the theory of probability and statistics is proposed.A new change detection algorithm based on t-test is put forward and it can be combined with correlation coefficient algorithm efficiently.As to high-resolution remote sensing imagery,texture is a crucial feature and an effective measure of texture is introduced in order to enhance change detection results.The robustness of the proposed approach is tested and validated with QuickBird and SPOT satellite images.Experimental results,obtained on two test data sets,prove the validity and application value of the proposed approach.
- change detection /
- remote sensing /
- t-test /
- texture /
- correlation coefficient