论文部分内容阅读
为提高国产自主卫星在海洋溢油监测中的精度,提出一种基于非负矩阵分解和支持向量机的环境减灾星(HJ-1)海洋溢油遥感图像分类算法。针对HJ-1星遥感图像,首先利用非负矩阵分解算法进行图像特征提取,相对图像光谱和纹理等图像基本特征,构造具有针对性的溢油图像局部化非负特征,更符合遥感图像特征所对应的物理意义。进而在新特征的基础上,采用支持向量机实现遥感图像分类,解决小样本训练问题。通过墨西哥湾溢油遥感图像仿真实验比较,证明该方法在HJ-1星溢油图像分类中的有效性。
In order to improve the accuracy of domestic autonomous satellites in marine oil spill monitoring, a classification algorithm based on non-negative matrix factorization and support vector machine (SVRM) for environmental oil spill reduction (HJ-1) remote sensing images was proposed. For the HJ-1 satellite remote sensing image, firstly, the non-negative matrix factorization algorithm is used to extract the image features, relative image spectrum and texture, and to construct the targeted non-negative features of the oil spill images, which are more in line with the remote sensing image features Corresponding physical meaning. Based on the new features, SVM is used to classify remote sensing images and solve the problem of training small samples. Through the comparison of oil spill remote sensing images in the Gulf of Mexico, the effectiveness of the proposed method in HJ-1 star oil spill image classification is proved.