本文以榆林市城區(qū)及其周邊范圍為實(shí)驗(yàn)區(qū)，以TM遙感圖像的第一主成分紋理信息、歸
一化植被指數(shù)和MNF變換得到的四個(gè)波段為數(shù)據(jù)源，采用支持向量機(jī)方法進(jìn)行分類(lèi)，并與最大似然法分類(lèi)和單純利用光譜信息的基于SVM分類(lèi)結(jié)果進(jìn)行比較。試驗(yàn)結(jié)果表明，將紋理分析方法應(yīng)用于圖像分類(lèi)中可區(qū)分光譜混淆的地類(lèi)；和傳統(tǒng)的分類(lèi)方法相比，采用支持向量機(jī)技術(shù)，使用光譜與紋理特征結(jié)合的分類(lèi)方法可以獲得更高的分類(lèi)精度。
關(guān)鍵詞：最小噪聲分離變換 灰度共生矩陣 支持向量機(jī)
Abstract: This paper discusses about the extraction of classification information of Yulin city and nearby area from an TM image and deals with the image classification based on the SVM method integrating the information of texture of PCA1, NDVI and MNF. And comparing to the results based on Maximum Likelihood and SVM method with single-source data. The results showed that the objects with same spectrum are distinguished by using texture analysis in image classification。 Compare with the traditional classification method, the classification based on the information of texture and spectrum using SVM could acquire higher classifying precision.
Key words: minimum noise fraction, gray level co-occurrence matrix, support vector machine