遥感图像人机交互判读方法研究及其应用

Study on Image Interactive Interpretation Technology and Its Applications

对于大型复杂的遥感判读任务,往往需要灵活地运用不同的判读方法,才能取得令人满意的效果。然而传统目视判读和常规计算机分类都难以满足这一要求。笔者设计了屏幕目视判读、改进型监督自动分类、分区自动分类、辅助波段分类、动态变化判读、人机混合判读和多组分工判读等遥感图像人机交互判读方法,把计算机自动分类与人工判读的优势相结合,对于提高图像判读精度、速度和高效率的遥感专题信息生成,帮助判读人员更好地完成各种遥感图像的判读任务都具有重要的意义。该文探讨了以上方法的特点、适用范围与实现方法,并给出了一些实例加以说明。

For a large interpretation task of remotely-sensed imagery, it is necessary to utilize different approaches flexibly to get a good result. However, traditional visual interpretation and automatic classification aided by computer are difficult to do so. For this reason,the author developed 'Image Interactive Interpretation System'(3IS) in Windows NT and VC6.0 environment by OOP method, and designed some approaches including 'interactive visual interpretation','improved supervised classification','image-segmented classification','ancillary band-added classification','dynamic change interpretation', 'human-machine mixed interpretation' and 'grouping interpretation'. They can combine the advantages of automatic classification and visual interpretation, and make easier to use the knowledge and experience of interpreter, and referenced data. They can improve the accuracy of result for getting thematic information. This paper mainly presents the approaches, characteristics and applications of image interactive interpretation technology.