Abstract:

A color feature extraction method of plant leaf has been designed according to the requirements of intelligent monitoring of plant growth. The overall color feature of plant leaf is extracted at first. The feature data can be used to identify a leaf is healthy, having overall color variation or having local color variation. If local color variation be detected, it will be segmented from the leaf and then its color and spatial distribution features will be computed. Tests have been conducted using leafs with different variations, in which the sunny side and the shade side of each leaf are set as two different series for comparison. The data show the feature parameters having been designed can provide data to judge a leaf is healthy or not, including the type and extent of color variation. The method is applicable to intelligent monitoring of plant growth and it can provide data for the following decision-making process.