中文摘要：

针对当前大多数裸地提取指数未考虑高反射率建筑与裸地在ETM＋遥感影像上的光谱相似性的问题,该文通过分析高反射率建筑与裸地的波谱特性,构建了裸地区域指数（BAI）和简化裸地区域指数（SBAI）,并使用土壤调节植被指数（SAVI）和改进的归一化差值水体指数（MNDWI）消除了植被和水体的影响。在此基础上,引用建筑指数（IBI）,通过指数函数图像拉伸的方法,增大裸地在BAI/SBAI与IBI上的差异,提出了增强型裸地区域指数（EBAI）和增强型简化裸地区域指数（ESBAI）,将裸地正值化、背景地物负值化,克服了大多数裸地指数阈值难以确定且易受噪声影响、精度差的缺点。实验结果表明,同参与对比的指数中具有最高精度的NDBaI相比,ESBAI生产者精度和使用者精度分别提升了10.26%、9.18%,EBAI分别提升了8.08%、7.78%。

英文摘要：

For most bare land index without considering the spectrum similarity between high reflectivity buildings and bare land on the ETM＋ remote sensing image,Bareness Area Index （BAD and Simplified Bareness Area Index （SBAI） are proposed by analyzing the spectrum characteristics of high reflectivity buildings and bare land. And also the effects caused by vegetation and water body to BAI and SBAI are decreased by using SAVI （Soil Adjusted Vegetation Index） and MNDWI （Modified Normal- ized Difference Water Index）. Based on the verified BAI and SBAI, IBI （Index-based Built-up Index） is introduced to build en- hanced bare land indexes. To get strong differences for bare land in each image, the images analyzed by IBI, BAI and SBAI are stretched with an exponential function. The ratios between stretched BAI and IBI, SBAI and IBI are normalized to construct EBAI （Enhanced Bareness Area Index） and ESBAI （Enhanced Simplified Bareness Area Index）. And finally a statistical analy- sis indicates that EBAI and ESBAI can make bare land positive and background object negative, and comparing with most other bare land indexes, EBAI and ESBAI have higher accuracy and are more easier to find a threshold to separate bare land and other land use type while suppressing noises effectively. The results of the experiment report that the producerrs accuracy and userts accuracy are improved by 10. 26% and 9.18% respectively for ESBAI, 8.08% and 7. 78% for EBAI, comparing with NDBaI （Normalized Difference Bareness Index）, which has the highest accuracy in the existing bare land index models.