Fungal pathogens represent a serious threat to forest plantations, causing losses of up to 40% of global production. In response to this problem, deep learning and computer vision technologies have emerged as fast and accurate tools for disease detection. This paper presents a systematic review of the literature on fungal detection in forests, with a focus on the application of computer vision and deep learning. The study analyzed methodologies, model architectures, tools, and data sources. It was identified that convolutional neural network (CNN) architectures, along with Vision Transformers, are the most effective, while TensorFlow and PyTorch are the most widely used open-source frameworks. In addition, the review highlights that data collection is mainly carried out through drones and satellites equipped with multispectral cameras, enabling large-scale monitoring. Despite advances, the scarcity of labeled datasets remains a challenge, which is being addressed through the use of Generative Adversarial Networks (GANs) to create synthetic data. Early detection of the fungus Lasiodiplodia theobromae in Gmelina arborea wood with these methods has proven to be more accurate and efficient than traditional approaches, contributing to more effective and sustainable forest management.