Abstract: This project aims to utilise technologies such as computer vision, machine learning, as well as deep learning to solve a forestry and agricultural problem. The objective is to reduce the labour-intensive task of manually detecting and measuring fungal lesions (diseases) on specific tree species – the Eucalyptus, by using both real and augmented datasets. The methods used included augmenting a real image dataset with synthetic images using a two-step process: generating a metadata file that contains all the parameters extracted from the real image dataset and generating a synthetic image dataset based on the parameters specified in the metadata file. We then compare the results using the Structural Similarity Index (SSIM) to verify the validity and usefulness of the synthetic image dataset. Lastly, we train a Convolutional Neural Network (CNN) model using ResNet, first on the real image dataset, and then the synthetic image dataset, and ultimately a combination of the two datasets. The findings indicate that the integration of synthetic data dramatically improves the model’s accuracy as demonstrated by higher SSIM values and better lesion detection rates. Future work will involve using stable diffusion to enhance synthetic images for improved model training. Additionally, the model will be improved to not only detect lesions but also measure their size to determine the pathogen strain and resistance.