CROP AND WEED CLASSIFICATION USING MODIFIED CONVOLUTIONAL NETWORK

Abstract

Internet of things (IoT) in the horticulture field gives crops-arranged data sharing and programmed cultivating arrangements under single organization inclusion. The segments of IoT gather the recognizable data from various plants at various points. The data accumulated through IoT parts, for example, sensors and cameras can be utilized to be controlled for a superior cultivating focused dynamic cycle. The irksome technique for distinguishing sicknesses like rust, spots, and creepy crawly pervasion caused numerous misfortunes for farmers because of inappropriate determination and therapy. This examination plans to apply deep learning strategies to ease the issue. Legitimate tuning of hyper parameters can address overfitting along with a proper decision of analyzers bringing about a productive classifier. Additionally, this examination recognized potential future exploration attempts to apply the model for genuine situations. Weed location and classification are significant and pivotal strides for region explicit weed control. This diminishes the general expense and the adverse consequence of utilizing pointless herbicides on human health and crops.