论文部分内容阅读
为了验证噪声对支持向量机分类器性能的影响,对“SVM可以有效用于含噪声和不确定性数据”这一观点进行定量分析评价,采用国产OMISII传感器获得的高光谱遥感数据进行了试验,为了更好地比较SVM分类器的抗噪性,先对原始数据进行支持向量机分类,然后在高光谱遥感影像中人为添加不同比例的椒盐噪声和条带噪声,然后进行支持向量机分类,并与传统的光谱角制图和最小距离分类算法进行比较。结果表明支持向量机具有明显的抗噪性,其分类性能受噪声影响较小,是一种有效的高光谱遥感影像分类器。
In order to verify the influence of noise on the performance of SVM classifier, the quantitative analysis and evaluation of “SVM can be effectively used for noise-containing and uncertainty data ” was carried out. Hyperspectral remote sensing data obtained from domestic OMISII sensors were used In order to compare the noise immunity of SVM classifier better, the original data were classified by SVM first, and then different proportion of salt, pepper and salt noise were added into the hyperspectral remote sensing image, and then SVM classifier , And compared with the traditional spectral angle mapping and minimum distance classification algorithm. The results show that SVM has obvious anti-noise performance and its classification performance is less affected by noise. It is an effective hyperspectral remote sensing image classifier.