基于分区和纹理特征的伊犁河谷遥感影像土地利用分类

摘要

The SPOT5 remote sensing image of Yining County in the Ili Valley of Xinjiang Uygur Autonomous Region was used in this paper.The texture information was extracted by means of Variance and Skewness algorithms supported by ERDAS IMAGINE 9.2 software.Based on the spectral and texture information,supervised classification was used to get the classification images of Yining County.Then the partition classification based on the spatial distribution of ground features was processed.The results showed that,the overall classification accuracy increased by 19.34％ and Kappa coefficient increased by 0.247.To test this classification method,the remote sensing image of Huocheng County was used,and a good result was also achieved.All these showed that the classification method could be prompted to the Ili Valley Land Use Classification.%利用新疆伊犁河谷的伊宁县SPOT5遥感影像,在ERDAS IMAGINE 9.2软件平台下,通过二次变异分析算法和三次非对称分析算法提取纹理信息.基于图像的光谱特征和纹理特征进行监督分类,然后根据地物的空间分布进行“分区分类处理”.结果表明,伊宁县的总体分类精度提高了19.34％,Kappa系数提高了0.247.利用霍城县遥感图像进行分类方法验证,同样取得了较好的效果,表明该方法可在伊犁河谷土地利用分类中进行推广应用.