Crop and Weed Image Classification Using Machine Learning Algorithms

While weeds are the bane of farmers and growth in crop yields, pesticides are the bane of the ecosystem due to their improper use especially as they can kill beneficial insects like pollinators. This project intends to design a machine learning-based automatic weed detection and precision pesticide application system to aid farmers. This way, instead of applying pesticides generally which is ineffective and hampering the ecosystem, targeted application of pesticides will be facilitated, which indeed would optimize the growth of the crops.These crops especially in Northern England study the effectiveness of machine learning and compare models - Predictive modelling and Custom body part segmentation, Random forests, Decision Trees, and Support Vector Machines - to evaluate which model accurately optimizes weed growth. This in turn can enhance and support the health of people as well as the environment.