基于航空影像的城市绿地提取方法研究

Research on urban green space extraction method based on aerial image

基于航空摄影获取的研究区影像,文章利用易康(eCognition)软件分别计算了研究区内的4种植被指数(VDVI,NGRDI,EXG与NGBDI),通过面向对象的分类方法得到了研究区城市绿地的不同提取结果,并根据分类样本点数据进行了精度评价。其中,VDVI的绿地提取准确率最高,总体精度和Kappa系数分别为93%和0.780,NGBDI的绿地提取准确率最低,总体精度和Kappa系数分别为63.4%和0.234。研究表明,面向对象的城市绿地提取方法在实际应用中具备可行性,可为城市绿地的提取开辟新的途径。

Based on aerial photography images of the study area,the article used eCognition software to calculate the four vegetation indices(VDVI,NGRDI,EXG,and NGBDI)within the study area.Different extraction results of urban green spaces in the study area were obtained through object-oriented classification methods,and accuracy was evaluated based on the classification sample point data.Among them,VDVI has the highest accuracy in green space extraction,with an overall accuracy and Kappa coefficient of 93%and 0.780,respectively.NGBDI has the lowest accuracy in green space extraction,with an overall accuracy and Kappa coefficient of 63.4%and 0.234,respectively.Research has shown that object-oriented methods for extracting urban green spaces are feasible in practical applications and can open up new avenues for extracting urban green spaces.