I. M Muhammad
Computer Vision is an area of computer science that aims at making machines "see”. This technique uses cameras and computing systems to detect, identify and measure objects for further image processing without relying on human eyes. With the development of computer vision, it has been widely used in agriculture automation, which is a key factor in its development. This review will provide a systematic survey and analysis of the technologies and barriers encountered during the last three years along with the future prospects and opportunities to develop a valuable resource for researchers. Recent advances in unmanned aerial vehicle (UAV) systems have further enhanced data acquisition & analysis in an agricultural environment. Advanced sensing technologies, such as RGB, multispectral, hyperspectral and thermal images, have enhanced the accuracy of field monitoring and allowed for more accurate crop health, soil health and water stress variability assessments. Additionally, machine learning and deep learning have improved automation of tasks like disease detection, yield prediction, and spatial analysis of farmland conditions. However, implementing these developments in practice remains limited by their environmental sensitivity, computational needs, and inability to be generalized for wide-spread use in varying agricultural contexts. The analysis shows that the existing technology can support the development of agricultural automation for small scale field operations, providing advantages of low cost, high efficiency and precision. However, numerous challenges remain. At first, the technology will gradually be expanded to new application fields, meaning that more and more technical challenges will have to be overcome. It's essential to build large datasets. In addition, the rapid advancement of agricultural automation will lead to increasing needs for qualified workers. In the end, ensuring the reliability and performance of related technologies in a wide variety of difficult environments is a constant challenge. Based on the evaluation and discourse, we believe that computer vision will be more widely incorporated into the development of intelligent methods like deep learning, be applied across wider agricultural production management systems with rich datasets, be adopted more broadly to solve current agricultural problems, and be increasingly used to make agricultural production management moreeconomical, flexible, and reliable, contributing to the intelligentization of agricultural automation equipment and systems.