Weed control is fundamental to farming since weeds decline yields, increment creation costs, impede reap and lower item quality. Weeds additionally hinder water system water-stream, obstruct pesticide application, and harbor infection creatures. Early techniques for weed control included hand cultivation with hoes powered cultivation with cultivators, lethal wilting with high heat. However the results of these techniques are not significant, different means are maybe more commonplace today, especially the utilization of herbicide synthetic compounds. Automating the weed removal process is very important as we have both the need and technology to do the same, how, ever we need a machine-learning model, which can differentiate between the main crops and weed. Weed identification is the first step in automating the weed removal process. The proposed method facilitates the identification of weeds, and onions, in the plantation field using Machine learning in real-time.