顾及资三卫星全色与多光谱数据的植被信息提取研究

Study on the Vegetation Information Extraction of the ZY-3 Satellite considering the Full Color and Multispectral Data

目前遥感影像数据量正以TB级的存储量不断增加,但是遥感信息提取的方法远远不能满足各行业的实际需求.因此,研究遥感影像的信息提取方法就显得尤为重要.以河北省雄安新区植被信息提取为例,顾及资源三号卫星全色影像的高分辨率,多光谱影像的空间信息、光谱信息和纹理信息,基于样本信息的面向对象分类技术,选择具有代表性的样本信息,在分类过程中引入NDVI信息作为补充规则进行植被信息半自动化提取,使分类结果精度达到了92.7%.实验结果表明,该技术方案操作简易,实用性较好,不仅有利于实时精确获取植被信息,方便未来城市生态环境的动态监测,且有助于推动未来城市土地规划和土地管理的发展.

At present, the data amount of image is increasing with the terabyte storage capacity, but the extraction method of remote sensing information is far from meeting the actual needs of various industries. Therefore, the information extraction study of remote sensing image becomes particularly important. In this paper, taking Xiong＇an New District of Hebei Province as an example, considering the high resolution of full color image, and the spatial information, spectral information, texture information of multi-spectral images of the ZY- 3 satellite, the object-oriented classification technology based on the sample information was adopted, and the representative of the sample information was selected, the NDVI information as a supplement of the rule was introduced in the classification process, finally the semi- automatic extraction of vegetation information was realized with the above method, and the accuracy of classification results reached 92.7%. The experimental results show that the technical scheme is simple and practical, which not only can help us to accurately obtain the vegetation information timely, but also promote the development of urban land planning and land management in the future.