In recent years, the agricultural sector has undergone a revolutionary shift toward ‘smart farming’, integrating advanced technologies to strengthen crop health and productivity significantly. This paradigm shift holds profound implications for food safety and the broader economy. At the forefront of this transformation is deep learning, a subset of artificial intelligence based on artificial neural networks, has emerged as a powerful tool in detection and classification tasks. Specifically, Convolutional Neural Networks (CNNs), a specialized type of deep learning and computer vision models, demonstrated remarkable proficiency in analyzing crop imagery, whether sourced from satellites, aircraft, or terrestrial cameras. These networks often leverage vegetation indices and multispectral imagery to enhance their analytical capabilities. Such model contributes to the development of systems that could enhance agricultural outcomes (Tian et al., 2020). This review encapsulates the current state of the art in using CNNs in agriculture. It details the image types utilized within this context, including, but not limited to, multispectral images and vegetation indices. Furthermore, it catalogs accessible online datasets pertinent to this field. Collectively, this paper underscores the pivotal role of CNNs in agriculture and highlights the transformative impact of multispectral images in this rapidly evolving domain.