Self-supervised weed detection in vegetable crops using ground based hyperspectral imaging

Abstract

A critical step in treating or eradicating weed infestations amongst vegetable crops is the ability to accurately and reliably discriminate weeds from crops. In recent times, high spatial resolution hyperspectral imaging data from ground based platforms have shown particular promise in this application. Using spectral vegetation signatures to discriminate between crop and weed species has been demonstrated on several occasions in the literature over the past 15 years. A number of authors demonstrated successful per-pixel classification with accuracies of over 80%. However, the vast majority of the related literature uses supervised methods, where training datasets have been manually compiled. In practice, static training data can be particularly susceptible to temporal variability due to physiological or environmental change. A self-supervised training method that leverages prior knowledge about seeding patterns in vegetable fields has recently been introduced in the context of RGB imaging, allowing the classifier to continually update weed appearance models as conditions change. This paper combines and extends these methods to provide a selfsupervised framework for hyperspectral crop/weed discrimination with prior knowledge of seeding patterns using an autonomous mobile ground vehicle. Experimental results in corn crop rows demonstrate the system's performance and limitations.

Real-Time Semantic Segmentation of Crop and Weed for Precision Agriculture Robots Leveraging Background Knowledge in CNNs

“…Some approaches have been investigated to address to this problem such as [3], [20], [22]. Wendel and Underwood [22] address this issue by proposing a method for training data generation. They use a multi-spectral line scanner mounted on a field robot and perform a vegetation segmentation followed by a crop-row detection.…”

Weed Detection for Selective Spraying: a Review

“…This work reported about 97% accuracy using CNN in the detection of broadleaf and grass weeds without soil and soybean in the background. Wendel and Underwood [27] created a selfsupervised method to discriminate weeds in crop fields, without manual labeling. This work gathered training data to create a self-supervised classification framework that was adaptive to crop variation without generating new datasets manually.…”

Machine vision detection of pests, diseases, and weeds: A review

“…Wendel et al [42]developed a self-supervising hyperspectral autonomous mobile ground vehicle (Ladybird robot) for barnyard grass (Echinochloa crusgalli), curly dock (Rumex Crispus), and caltrop (Tribulus terrestris) weeds detection in the cornfield. The ladybird robot scans two rows of the field at a time mounted with hyperspectral line scanning camera (Resonon Pika II VNIR), and an RGB camera mounted underneath the robot to collect color images of the field.…”