AgroChromaBot: Smart Leaf Color-Sensing Robot for Selective Intelligent Pest Management

In modern agriculture, the uncontrolled and widespread application of pesticides has not only done bad things for the environment, but also made the farming activity higher-priced and influenced the wellbeing of crops. In order to overcome this issue, we present a new intelligent robotic system of pesticide spraying in the crop fields, which is for real-time leaf color diagnosis. The system uses a color identification algorithm, which reads color changes or discolorations in the leaves of plants that are usually an early sign of insect infestation or disease. The robotic system is designed to combine a high-resolution RGB camera, an on-board computer, and a spraying system on a four-wheeled rover. The experiments were conducted in different crop circumstances and the performance was measured in such parameters as the determination of the accuracy of spraying ($94.3 \%$), the reduction of pesticide consumption (up to $38 \%$), spraying profitability, and the rate of leaf condition improvement. Leaf color classification is $\mathbf{9 2. 3}$ percent. The system showed stable color identification in different lighting conditions and sturdy paths going on uneven ground. Depending on the grade of the color deformation as compared to the values of healthy leaves, the robot independently changes the strength of the spray and hits only the damaged areas. It is a selective method which significantly lowers chemical consumption, harmful effects on the environment and efficiency of operations. The cheap and scalable product has great prospects in the field of precision agriculture and sustainable farming techniques.