Ultrasonic Arrays for Pasture Biomass Estimation and Grape Occlusion Detection / Detecting and Counting Sheep in UAV Videos
Room: Room ALR6/421W-501, Level 5, Bldg: Architecture and Planning Building (421), University of Auckland, City Campus, , Symonds Street, Auckland, Auckland, North Island, New Zealand, 1010, Virtual: https://events.vtools.ieee.org/m/315457Ultrasonic Arrays for Pasture Biomass Estimation and Grape Occlusion Detection Dr Mathew Legg / Baden Parr The ability to accurately measure pasture biomass can have a significant effect on the profit farmers can achieve from their pastures. One method previously used to estimate pasture biomass has been to measure pasture height using ultrasound. A transducer is fixed to a farm bike and the distance to the top of the grass is measured. This method assumes that the ground is a set distance below the transducer. However, errors occur when the bike tilts and bounces and when pasture density varies. This talk will describe the development of a novel air-coupled ultrasonic array system for pasture biomass estimation developed by Mathew and Stuart Bradley. Improved biomass estimation is achieved by imaging through the grass to estimate both pasture density and ground location. It is the first pasture biomass estimation system to have reported these capabilities using proximal (remote) sensing techniques. This talk will also describe a study performed by Baden and Mathew using a high-resolution ultrasonic array to image through leaves to detect occluded grapes. This is the first-time ultrasound has been used to image fruit occluded by leaves. This has the potential for helping to improve automated grape yield estimation. Detecting and Counting Sheep in UAV Videos Farah Sarwar The focus of this research work is to detect, count and monitor livestock in a paddock using an unmanned aerial vehicle (UAV). It involves no disruption to the farm animals as the UAV will barely be noticeable from the chosen flight altitude. Recorded images and videos are processed using a combination of deep learning and machine learning algorithms to perform the aforementioned tasks. The data set was designed after collecting data from different paddocks to cover a variety of background, weather and paddock conditions. The achievement of 98% recall for object detection and high accuracy for multiple object tracking suggests that this research will be beneficial for farmers in terms of saving time and reducing costs. Speaker(s): Dr Mathew Legg, Dr Farah Sarwar Agenda: 12:00 Dr Mathew Legg/Baden Parr: Pasture Biomass Estimation and Detection of Occluded Grapes using Ultrasonic Arrays 12:30 Farah Sarwar: Detecting and Counting Sheep in UAV Videos Room: Room ALR6/421W-501, Level 5, Bldg: Architecture and Planning Building (421), University of Auckland, City Campus, , Symonds Street, Auckland, Auckland, North Island, New Zealand, 1010, Virtual: https://events.vtools.ieee.org/m/315457
NZCS committee meeting
Monthly meeting of the NZ Central Section committee. IEEE NZCS Committee Minutes 2022-06-02