Classification of Density and Transparency of Needle Leaves Types Using AlexNet and VGG16 Architecture

This article discusses the application of digital image technology and deep learning using Convolutional Neural Network (CNN) in Forest Health Monitoring (FHM).Forest health monitoring is a method for measuring forest health, one of the parameters used is crown density and transparency.Measurement of these parameters is still done manually using magic cards so it is less effective and efficient, so it is necessary to apply digital images, one of which is the CNN algorithm to help measure the density scale and crown transparency.CNN architectures namely AlexNet and VGG16 are used to train the tree image recognition model.This research uses a tree image dataset with four types of needles grouped into classes based on crown density and transparency.The results showed that both CNN architectures achieved a good level of accuracy in classifying coniferous tree species based on crown density and transparency.VGG16 notably achieves higher accuracy than AlexNet.The results of model evaluation via the confusion matrix also provide insight into the model's performance in recognizing crown density and transparency classes.