Abstract:

Image is the largest information resource of intelligent forestry. The development and accuracy of image comprehension are directly related to the development and application process of intelligent information technology in forestry. In this article, we segmented the Populus alba×P. berolinensis seedling from an image by image comprehension method at first, and then analyzed the light need of it, and at last compared the image with the images of yellowing seedlings for drought. The results indicate that the seedling can be segmented well by minimum distance discrimination method based on L^*a^*b^* model, the gradual process of plant changing from green to yellow due to inadequate light can be statemented by the changes of the peak value in R component histogram. In addition, it may achieve better result using the the value of b^* component in L^*a^*b^* model to judge it is drought or lack of light leading to seedling yellowing.

Key words: Populus alba×, P. berolinensis, light need, image recogniton, image segmentation, L^*a^*b^* model