Skip to content

Application of Joint Skewness Algorithm to Select Optimal Wavelengths of Hyperspectral Image for Maize Seed Classification(基于联合偏度的高光谱图像波段选择对玉米种子分类研究)

License

Notifications You must be signed in to change notification settings

bleakie/JointSkewness

Repository files navigation

JointSkewness

基于联合偏度的高光谱图像波段选择对玉米种子分类研究

Application of Joint Skewness Algorithm to Select Optimal Wavelengths of Hyperspectral Image for Maize Seed Classification

论文地址:https://www.researchgate.net/publication/317750407_Application_of_Joint_Skewness_Algorithm_to_Select_Optimal_Wavelengths_of_Hyperspectral_Image_for_Maize_Seed_Classification

1.安装

(0)开发环境win+matlab

(1)LibSVM

(3)TensorToolbox

2.数据处理&算法

(1)由于涉及隐私数据,因此不公开原始数据,按照论文步骤直接得到结果(.mat文件),参见1.Materials and Methods

(2)张量的展开

MASK0

3.结果

在 10 个最优波段条件下,联合特征分类模型的识别精度达到了 96.28%,比光谱均值和图像熵的识别精度分别提高了 4.30%和20.38%,也高于全波段联合特征识别模型的 93.47%

(1)全波段单特征&多特征

MASK0

(2)单特征&多特征下的波段选择

MASK0

(3)单特征&多特征下的波段选择后详细分类结果

MASK0

(4)多特征&不同波段选择方法比较

MASK

About

Application of Joint Skewness Algorithm to Select Optimal Wavelengths of Hyperspectral Image for Maize Seed Classification(基于联合偏度的高光谱图像波段选择对玉米种子分类研究)

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published