CAJ | 학술논문

选择广西典型岩溶山区灌草植被类型设置91块样方，运用GreenSeeker手持光谱仪实测不同覆盖度植被的光谱响应并构建归一化植被指数( NDVI)、差值植被指数(DVI)、比值植被指数(RVI)、修正植被指数(MVI)和修改型土壤调节植被指数(MSAVI)等5种植被指数，同时运用数码相机垂直拍摄样方照片并进行监督分类，提取植被覆盖度，最后对样方的植被覆盖度和植被指数进行相关分析、建立回归模型和精度验证。结果表明：NDVI、RVI和MVI与植被覆盖度高度相关，相关系数超过0.9；而且用NDVI反演的植被覆盖度与实测值在α=0.05显著水平下进行t检验，结果无显著差异，平均精度达到95%以上，比其他植被指数模型更优秀。研究结论对广西岩溶石漠化地区植被覆盖度的快速、连续监测与评估具有重要参考价值。
선택엄서전형암용산구관초식피류형설치91괴양방，운용GreenSeeker수지광보의실측불동복개도식피적광보향응병구건귀일화식피지수( NDVI)、차치식피지수(DVI)、비치식피지수(RVI)、수정식피지수(MVI)화수개형토양조절식피지수(MSAVI)등5충식피지수，동시운용수마상궤수직박섭양방조편병진행감독분류，제취식피복개도，최후대양방적식피복개도화식피지수진행상관분석、건립회귀모형화정도험증。결과표명：NDVI、RVI화MVI여식피복개도고도상관，상관계수초과0.9；이차용NDVI반연적식피복개도여실측치재α=0.05현저수평하진행t검험，결과무현저차이，평균정도체도95%이상，비기타식피지수모형경우수。연구결론대엄서암용석막화지구식피복개도적쾌속、련속감측여평고구유중요삼고개치。
Ninety-one quadrats were set of shrub and herb vegetation types in Guangxi typical karst mountains area. By using GreenSeeker hand-held spectrometer, the spectral responses of shrub and grass in different vegetation coverage quadrats were measured, and the normalized data of five vegetation indexes were calculated such as NDVI (Normalized Difference Vegetation Index), DVI (Difference Vegetation Index), RVI (Ratio Vegetation Index), MVI (Modiifed Vegetation Index) and MSAVI (Modiifed Soil Adjusted Vegetation Index). At the same time, the photos were vertically taken from from above the quadrats by digital camera, Then the vegetation fraction (VF) could be extracted through a series of image pre-processing and image supervised classiifcation by ArcGIS10.1 and ENVI4.8 software. Finally, the correlation between Quadrats’ vegetation coverage and vegetation index were analyzed and the regression models of two variables were established, and the accuracy of the models were veriifed by means of small sample inspection. The results show that there were high correlations between NDVI, RVI, MVI and vegetation coverage, and the correlation coefifcient was greater than 0.9. The accuracy of vegetation coverage model inversed with NDVI was better than those of other models which inversed with rest vegetation indexes. The VF-NDVI model passed the check through t-test under the alpha=0.05 signiifcant levels and the overall average accuracy were 95%or more. The research results can be used as the important references for fast, continuous monitoring and evaluating the vegetation fraction in Guangxi karst mountainous area in China.