查詢結果分析
來源資料
頁籤選單縮合
題名 | 植生指標對影像分類準確度影響之研究=Effect of the Vegetation Index on the Accuracy of Image Classification |
---|---|
作者 | 林世峻; 莊智瑋; 何世華; 林昭遠; | 書刊名 | 水土保持學報 |
卷期 | 40:3 2008.09[民97.09] |
頁次 | 頁315-327 |
分類號 | 436.7 |
關鍵詞 | 植生指標; 倒傳遞類神經網路; 最大概似法; 影像判釋; Vegetation index; Back propagation neural network; Maximum-likelihood decision rule; Image classification; |
語文 | 中文(Chinese) |
中文摘要 | 影像分類時易受空間解析度、波譜解析度等影響,進行分類時,加入輔助特徵值如( NDVI、CMFI )可稍有效提升影像精度。本研究以南投縣九份二山為研究試區,以最大概似法(Maximum-Likelihood Decision Rule)和倒傳遞類神經網路(Back Propagation Neural Network)進行影像分類之探討,結果顯示,除倒傳遞類神經網路搭配原始波段加入NDVI外(Kappa為0.5954),加入植生指標均可稍微提升分類精確度;在分類模式整體比較上,倒傳遞類神經網路稍優於最大概似法,但差距甚微。 |
英文摘要 | The accuracy of image classification is easily influenced by the image’s spatial and spectrum resolution. Feature values extracted from the image can be used effectively to promote the accuracy of the classification. Chiufenershan landslide area in Nantou County was chosen as the study site. The improvement of image classification can be achieved by using the maximum-likelihood decision rule and the back-propagation neural network after adding the calculation of the vegetation index.The results show that vegetation index can increase the accuracy of image classification , and the back-propagation neural network is better than maximum-likelihood decision rule in the performance of image classification by using vegetation index except NDVI (When Kappa is 0.5954). |
本系統之摘要資訊系依該期刊論文摘要之資訊為主。