Weedspedia: Deep Learning-Based Approach for Weed Detection using R-CNN, YoloV3 and Centernet

Weed identification is a crucial undertaking in agriculture, where management and identification of weed growth in fields need to be addressed through deep learning algorithms. This study uses the Region-Based Convolutional Neural Network (R-CNN), Yolo, and Centernet object recognition algorithms to identify objects in images. This study also proposes a novel method for weed identification based on weed size. Numerous image-processing applications, including those for healthcare, traffic analysis, and agriculture, frequently use R-CNN. The suggested technique provides a practical and affordable answer for weed identification in precision agriculture. Herbicide use by farmers can be reduced as a result of being able to more precisely identify and control weed growth. Identification of weeds aids in weed growth control, which can harm natural ecosystems, compete with crops, diminish yields, and increase maintenance costs. The methodology of measuring the length and breadth of a weed detected by R-CNN requires extra processing stages, which aid in understanding the growth and subsequent steps required to eliminate weeds dependent on the growth of the weeds. Researchers often begin by gathering a sizable collection of photos of various weed species to construct a deep learning model for weed identification. The paper focus on identification of weed using 3 pre trained models followed by growth prediction to take necessary steps in agricultural land.