RGB图像分析法测定台风0613诱发的竹叶黄化和坏死(英文)

遭遇台风0613号袭击之后,日本山口市的许多乔木和灌木出现叶尖和叶缘甚至整叶黄化和坏死的症状。同一竹林林分中既有黄化叶片也有坏斑叶片。本文用非破坏、非接触的RGB图像分析法研究了这一现象。研究表明,室内拍摄的黄化竹单叶图像的G/R值与叶绿素计(SPAD)值之间呈紧密的正相关,平方回归相关系数0.961;与坏死斑面积率之间呈反逻辑斯蒂函数关系,平方回归相关系数0.958。用RGB图像分析法不仅可以估测叶片的黄化程度而且可也估测叶片的坏死斑面积。研究还发现,光线差异较大时拍摄的图片,相同叶片的G/L值的方差小于G/R值的方差值,尤其是绿色叶片。与G/R相比,黄化叶片的G/L值得到了与SPAD更紧密的相关性。室外拍摄的竹林冠分析表明,光线差别较大的图像G/L值与林分距海岸的距离的相关性也高于G/R值。

Measuring leaf necrosis and chlorosis of bamboo induced by typhoon 0613 with RGB image analysis

Symptoms of leaf necrosis or chlorosis of bamboo induced by Typhoon 0613 (T0613) were analyzed using RGB image analysis in Yamaguchi city,Japan. Results showed a closely positive relationship between Green/Red (G/R) value for indoor taking images of bam-boo individual leaves and chlorophyll meter value (SPAD) with regression coefficient of 0.961. The relation between G/R value of room taking images and Necrotic Area Percentage (NAP) for bamboo individual leaves showed an inverse logistic function relationship,with the correlated coefficient equaling to 0.958. Both leaf chlorosis and necrosis can be quantitatively estimated by RGB image analysis. Moreover,the variance of Green/Luminance (G/L) value for the same leaf was less than that of G/R for images taken in the conditions with large light difference,especially for green leaves. G/L value also exhibited a closer relationship with SPAD value of leaves with chlorosis than that of G/R values at the same condition. The relationship between G/L value for bamboo canopies and the Distance from Coastline (DC) was also closer than that of the G/R value for the images taken at field sites with big light difference.