Robotic weed control has seen increased research of late with its potential for boosting productivity in agriculture. Majority of works focus on developing robotics for croplands, ignoring the weed management problems facing rangeland stock farmers. Perhaps the greatest obstacle to widespread uptake of robotic weed control is the robust classification of weed species in their natural environment. The unparalleled successes of deep learning make it an ideal candidate for recognising various weed species in the complex rangeland environment. This work contributes the first large, public, multiclass image dataset of weed species from the Australian rangelands; allowing for the development of robust classification methods to make robotic weed control viable. The DeepWeeds dataset consists of 17,509 labelled images of eight nationally significant weed species native to eight locations across northern Australia. This paper presents a baseline for classification performance on the dataset using the benchmark deep learning models, Inception-v3 and ResNet-50. These models achieved an average classification accuracy of 95.1% and 95.7%, respectively. We also demonstrate real time performance of the ResNet-50 architecture, with an average inference time of 53.4 ms per image. These strong results bode well for future field implementation of robotic weed control methods in the Australian rangelands.

No abstract

No abstract

Sedimentation is considered the most widespread contemporary, human-induced perturbation on reefs, and yet if the problems associated with its estimation using sediment traps are recognized, there have been few reliable measurements made over time frames relevant to the local organisms. This study describes the design, calibration and testing of an in situ optical backscatter sediment deposition sensor capable of measuring sedimentation over intervals of a few hours. The instrument has been reconfigured from an earlier version to include 15 measurement points instead of one, and to have a more rugose measuring surface with a microtopography similar to a coral. Laboratory tests of the instrument with different sediment types, colours, particle sizes and under different flow regimes gave similar accumulation estimates to SedPods, but lower estimates than sediment traps. At higher flow rates (9-17 cm s -1 ), the deposition sensor and SedPods gave estimates[109 lower than trap accumulation rates. The instrument was deployed for 39 d in a highly turbid inshore area in the Great Barrier Reef. Sediment deposition varied by several orders of magnitude, occurring in either a relatively uniform (constant) pattern or a pulsed pattern characterized by short-term (4-6 h) periods of 'enhanced' deposition, occurring daily or twice daily and modulated by the tidal phase. For the whole deployment, which included several very high wind events and suspended sediment concentrations (SSCs)[100 mg L . For the first half of the deployment, where SSCs varied from\1 to 28 mg L -1 which is more typical for the study area, the deposition rate averaged only 8 ± 5 mg cm. The capacity to measure sedimentation rates over a few hours is discussed in terms of examining the risk from sediment deposition associated with catchment run-off, natural wind/wave events and dredging activities.

Laminated sediments in Lake Ohau, Mackenzie Basin, New Zealand, offer a potential high‐resolution climate record for the past 17 kyr. Such records are particularly important due to the relative paucity of detailed palaeoclimate data from the Southern Hemisphere mid‐latitudes. This paper presents outcomes of a study of the sedimentation processes of this temperate lake setting. Hydrometeorological, limnological and sedimentological data were collected over a 14 month period between 2011 and 2013. These data indicate that seasonality in the hydrometeorological system in combination with internal lake dynamics drives a distinct seasonal pattern of sediment dispersal and deposition on a basin‐wide scale. Sedimentary layers that accumulate proximal to the lake inflow at the northern end of the lake form in response to discrete inflow events throughout the year and display an event stratigraphy. In contrast, seasonal change in the lake system controls accumulation of light (winter) and dark (summer) laminations at the distal end of the lake, resulting in the preservation of varves. This study documents the key processes influencing sediment deposition throughout Lake Ohau and provides fundamental data for generating a high‐resolution palaeoclimate record from this temperate lake.