基于PCA变换的超谱图像降维算法仿真  被引量:7

Simulation of Hyperspectral Image Dimensionality Reduction Algorithm Based on PCA Transform

在线阅读下载全文

作  者:傅荣会[1] FU Rong-hui(Institute Computer Science,Neijiang Normal College,Neijiang Sichuan 641100,China)

机构地区:[1]内江师范学院计算机科学学院

出  处:《计算机仿真》2019年第11期234-238,共5页Computer Simulation

摘  要:针对当前图像降维算法复杂度和图像信息丢包率高的问题,提出基于PCA变换的超谱图像降维算法。采用矩匹配法与多波段匹配法对超谱图像条带噪声进行消除,基于消除结果根据分水岭算法对图像进行分割,利用分割过程中的图像目标对象内部标记使降维之后能够得到图像连续像素,与设定阈值之间进行比较,将变异比设定阈值小的临近像素集合判断为标记,利用所得标记控制分水岭分割,降低图像信息丢包率。结合图像分割,定义一幅超谱图像的原始矩阵,经PCA变换之后得到一个新矩阵,矩阵中各列就是变换之后的主成分,从大至小的顺序排列主成分,选择前二、三列组成数据图像包含的信息和原始图像相似程度最高,实现超谱图像降维。实验结果显示,上述算法降维复杂度低、图像信息完整性良好。Due to high complexity and high packet loss rate of image information, this article puts forward an algorithm of reducing dimensionality for hyperspectral image based on PCA transform. At first, the stripe noise of hyperspectral image was eliminated by the moment matching method and multi-band matching method. Based on the elimination result, the watershed algorithm was used to segment images. Then the internal tags in target object during the segmentation were used to obtain continuous pixels of image after dimension reduction. Compared with the set threshold value, the set of adjacent pixels whose variation was smaller than the setting threshold value was determined as the mark. Moreover, the watershed segmentation was controlled by the obtained mark to reduce the packet loss rate of image information. Based on the image segmentation, the original matrix of a hyperspectral image was defined. After PCA transformation, a new matrix was obtained. The columns in this matrix were the principal components after the transformation. Finally, the principal components were arranged from big to small. The first two or three columns of data were selected to compose the image. The information contained in the image is the most similar to the original image. Thus, the dimensionality reduction of hyperspectral image was achieved. Simulation results show that the proposed algorithm has lower dimensionality reduction and better integrity of image information.

关 键 词:变换 超谱图像 降维 

分 类 号:TP391[自动化与计算机技术—计算机应用技术]

 

参考文献:

正在载入数据...

 

二级参考文献:

正在载入数据...

 

耦合文献:

正在载入数据...

 

引证文献:

正在载入数据...

 

二级引证文献:

正在载入数据...

 

同被引文献:

正在载入数据...

 

相关期刊文献:

正在载入数据...

相关的主题
相关的作者对象
相关的机构对象