Towards Autonomous Visual Navigation in Arable Fields

Abstract

Autonomous navigation of a robot in agricultural fields is essential for every task from crop monitoring through to weed management and fertilizer application. Many current approaches rely on accurate GPS, however, such technology is expensive and also prone to failure (e.g. through lack of coverage). As such, navigation through sensors that can interpret their environment (such as cameras) is important to achieve the goal of autonomy in agriculture.In this paper we introduce a purely vision based navigation scheme which is able to reliably guide the robot through rowcrop fields. Independent of any global localization or mapping, this approach is able to accurately follow the crop-rows and switch between the rows, only using on-board cameras. With the help of a novel crop-row detection and a novel crop-row switching technique, our navigation scheme can be deployed in a wide range of fields with different canopy types in various growth stages. We have extensively tested our approach in five different fields under various illumination conditions using our agricultural robotic platform (BonnBot-I). And our evaluations show that we have achieved a navigation accuracy of 3.82cm over five different crop fields.

Vision based Crop Row Navigation under Varying Field Conditions in Arable Fields

“…Ahmadi et al [21] developed a vision based navigation system for arable fields that uses colour based segmentation for crop row mask prediction. Their crop row line parameter estimation was based on the least square fitting of detected crop centres.…”

“…Ahmedi et al [21] has developed a vision based navigation scheme that will detect the crop rows and drive the robot using the IBVS(image absed visual servoing) controller [24]. They identify the central crop row using least square fitting of crop centers detected using a colour based segmentation approach.…”

“…They have highlighted the shortcomings of previous colour based crop row detection algorithms placing them at the state-of-the-art in colour based crop row detection. The baseline algorithm was tested using the codes provided by the authors of [21] on our dataset.…”

BonnBot-I: A Precise Weed Management and Crop Monitoring Platform

“…To achieve this, plant-level intervention needs to operate in different fields with varying crops, weed species and weed distributions. To enable this, "smart farming techniques" aim to incorporate automated navigation [11], crop monitoring [6], and weeding [12]. One of the key elements to achieve precise weeding is plant-level treatment, where the treatment of each plant is dictated by its species, size and its impact upon not only the crop but also the environment [13].…”

Explicitly Incorporating Spatial Information to Recurrent Networks for Agriculture

“…Arable Farming Sugar Beet (SB20), first introduced in [8], was captured at a sugar beet field in campus Klein-Altendorf (CKA) of the University of Bonn using an Intel RealSense D435i camera with a nadir view of the ground mounted on BonnBot-I [31] driving at 0.4m/s. Sequences contain robot wheel odometry and RGB-D images of crops and 8 different categories of weeds at different growth stages, different illumination conditions and three herbicide treatment regimes (30%, 70%, 100%), impacting weed density directly.…”