AI x Agriculture
Agriculture is one of the most historical industries. The most traditional agriculture utilizes simple tools like sledge and livestock to increase efficiency, modern agriculture uses more advanced mechanical utilities such as electric tiller and tractor. Although the advanced tools significantly increase productivity, they still lack specification and management and thus causes pollution (ie. fertilizer, pesticide, contamination from livestock, etc.) and waste of energy. However, with AI tools, specifically prediction and agriculture robots, agriculture can achieve detailed management and thus reduce pollution.
Robotics is a broad application field including AI, mechanical engineering, electric engineering, etc. Unlike the fancy robots shown in movies, agriculture robots are more like machines that are informed AI and can finish some tasks. With the help of AI, robots can outperform human in managing and harvesting and significantly reduce human labor work. Managing plants need tremendous inspections and works such as checking pests and soil, but robots not only free human from the redundant work but also are more accurate. For example, R2Weed2 can detect weeds and remove them effectively, and it can also perform soil analysis, which is almost impossible for farmers to achieve for each small region of soil. The analysis results can be further used for other AI program to propose fertilizer or pesticide plan, which optimize environmentally harmful products efficiency and growth of plants.
Other AI tools can also help with agriculture such as machine learning (ML), computer vision (CV), etc. Roughly speaking, ML is a program that is fed with data to train and make prediction about new data, CV is to teach program to see things and do inference as human such as recognizing human faces. Scientists has been trying for decades about training ML to predict crop output in some given conditions like season, soil, etc. This helps farmers to decide what to grow on their land such that the output can be maximized. CV can be used in detection such as detecting pests, crop health, weeds (R2Weed2), etc., which helps farmers to detailly manage each plant. As the output is maximized in each land, fewer lands and resources will be used and wasted.
Despite the benefits of using AI in agriculture, there are some disadvantages such as AI may be less accurate and experienced than human when detecting weeds and pests. However, as AI is growing rapidly nowadays, more accurate algorithms will be proposed and hopefully AI can fully take charge of agriculture and reduce pollution.
CHENHAO ZHANG
