Crop and Weed Classication Using Pixel-wise Segmentation on Ground and Aerial Images

Artificial Intelligence (AI) is a key tool in agriculture for implementing sustainable strategies for weed control. In traditional weed control, the agrochemical inputs are uniformly applied to the field, while innovative approaches using AI aim at minimizing the usage of chemical inputs thanks to local applications. In this paper, we focus on agricultural robotics systems that address the weeding problem by means of selective spraying or mechanical removal of the detected weeds. We present a set of deep learning-based methods designed to enable a robot to efficiently perform an accurate weed/crop classification from RGB or RGB+NIR (Near Infrared) images. In particular, we use two Convolutional Neural Networks (CNNs) to simplify and speed up the training process. A first encoder-decoder segmentation network is designed to perform a "plant-type agnostic" segmentation between vegetation and soil. Each plant is hence classified between crop and weeds by using a second network, depending on the type of pipeline, for patch-level or pixel-level classification. We introduce also a third CNN, specifically designed for setups with limited resources, like in small UAVs (Unmanned Aerial Vehicles), that exploits the proposed encoder-decoder segmentation network to efficiently estimate crop/weeds local statistics. Quantitative experimental results, obtained using multiple publicly available datasets demonstrate the effectiveness of the proposed approaches.