摘要
提出了一种采用高光谱图像技术结合人工神经网络对油菜籽品种进行鉴别的方法．采集多个品种油菜籽400～1000cm范围的高光谱图像数据，通过主成分分析法（PCA）获得主成分图像，确定特征波长；采用基于灰度直方图和灰度共生矩阵联合的统计方法从特征图像中提取纹理特征参数，应用人工神经网络建立油菜籽品种鉴别模型．结果表明，模型训练时品种判别率为93．75％，预测的判别率为91．67％．说明高光谱图像技术对油菜籽品种具有较好的分类和鉴别作用．
Identification of rapeseed varieties by using hyperspectral imaging technique combined with artificial neural network （ANN） was proposed. Hyperspectral images of several rapeseed varieties in range 400-1000 nm were acquired, and then the principal component analysis （PCA） was performed to select three optimal band images. The texture parameters were extracted from the optimal band images based on gray level histogram and gray level co occurrence matrix （GLCM） statistical methods. The ANN model was used for the identification of rapeseed varieties. Detection results of ANN model showed that the discriminating rates of rapeseed varieties in the training and prediction sets were 93.75% and 91.67%, respectively. It is indicated that the hyperspectral imaging technology has a good classification and identification effects on rapeseed varieties.
出处
《浙江大学学报（农业与生命科学版）》
CAS
CSCD
北大核心
2011年第2期175-180,共6页
Journal of Zhejiang University:Agriculture and Life Sciences
基金
国家自然科学基金资助项目(60802038)
国家高技术研究发展计划863"资助项目(2006AA10Z234)
浙江省重大科技专项重点农业资助项目(2009C12002)
关键词
图像处理
高光谱图像
品种鉴别
主成分分析
油菜籽
image processing
hyperspectral imagery
variety identification
principal component rapeseed
