1.9 KiB
Ai in crop management
Team 8 survey a broad range of papers that discuss in depth various aspect and applications of AI in the horticulture space. To summarize I will briefly mention each of the application areas. The first application is that of pest management. Using AI can help with identifying pests in a cheaper and more precise way than manual methods. Having better monitoring allows for more precision application of the pesticides. Identifying diseases that are on your crops is vitally important to manage them and either cure them or mitigate damages. Using drone to capture images you can then use the YOLO model to identify and classify troubled plants. Using XAI methods on ML models can help give farmers real insight into the options and choices that they have. These methods involve using IoT senors to collect large amount of data and modelling this data with respect to crop type. From here you can apply XAI methods on the black box models to understand the effect of the fields environment on crops. The water use of agriculture and horticulture is shockingly high. Using AI to more precisely and efficiently use the Water could have very high benefits. The idea is to use various easily accessible data sources and applying ML models to predict irrigation needs. These have been shown to be quite effective however still far from off the shelf systems. Another use of remote sensing is to predict the yield that you will get. This is particularly useful when you rely on natural resources like rainfall instead of human powered irrigation. Relatively simple models like random forests perform with very high accuracy.
Overall there are many different ways that AI can help improve the efficiency and effectiveness of farming. It seems to be an application area that the risks involved with applying AI are the more mundane risks of unintended consequences, sufficiently wrong AI models and slow/inequable adoption.