Overview and details of the sessions of this conference. Please select a date or location to show only sessions at that day or location. Please select a single session for detailed view (with abstracts and downloads if available).
Design and fabrication of an auger type applicator for 50-ohm technology based inline RF disinfestation of stored grains: A technical report
Roland Jampit Macana, Tolen Tumbong Moirangthem, Oon-Doo Baik
Chemical and Biological Engineering, University of Saskatchewan, Canada
Radio frequency disinfestation is an alternative method for the current disinfestation of stored grains using fumigants and pesticides. The advantages of this method are volumetric, fast, and chemical free. However, the non-uniform heating of the system that makes the insect pests migrate to the cold spot is a major challenge for RF heating assisted disinfestation. Some studies suggested that the insect pest’s migration during RF heating could be minimized by putting physical barriers in the heating system and by using a high powered RF heating system to reduce the time of moving of the insect pests going to the cold spot. Thus, this study dealt with the design and fabrication of a continuous auger type applicator which is expected to improve the heating uniformity of the 50-ohm technology based RF system. The applicator is connected to the automatic matching network (AMN) and AMN is connected to the 15 kW and 27.12 MHz generator using 50-ohm coaxial cable. The applicator has an auger that stirs and conveys the grain continuously until it passes between the two electrodes (hot and ground) attached to both sides of the channel. The technical details of the applicator are: capacity = 45 kg (wheat) and 40 kg (canola) at 0.3 m channel diameter, rpm (motor) = 1-200 rpm, and electrode dimensions = 0.7 m length and 0.3 m width. The materials are aluminum (electrode), polypropylene (tubular channel and auger), copper (welding cable connector between AMN and applicator), and expanded metal (casing for the applicator).
1:40pm - 2:00pm
The effect of compaction on pore structure in bulk grains
Charles Nwaizu, Qiang Zhang
University of Manitoba, Canada
Experiments were conducted to determine the effect of compaction due to grain depth on porosity, tortuosity and connectivity of pore structure within bulk grains. Wax was used to “freeze” the pore structure to allow the grain samples to be cut and the internal structures to be imaged. The acquired images were analyzed by using a 3-D image reconstruction technique to determine various pore structure parameters. Results indicated that porosity decreased with the grain depth (compaction pressure), with the initial porosity being 0.41 and decreased to 0.34 at a pressure equivalent to a 5 m grain depth. Tortuosity increased with grain depth from 1.17 initially to 1.57 at a 5 m grain depth, or 34%. Models were developed based on 3-D experimental data to predict variations in porosity and tortuosity as a function of grain depth. Model predictions were within 5% of experiment data.
2:00pm - 2:20pm
Implementation of Computer Vision Algorithms to Monitor Bees & Predators Surrounding the Apiary
Cyrus Shafai, Colin Gaudreau, Miroslaw Pawlak
University of Manitoba, Canada
Beekeepers often have their apiaries spread across a large geographical area. As a result, remote sensing technology has become increasingly explored as a method to monitor apiaries. While acoustic, temperature, humidity, and weight sensors have all been explored, little work has been done incorporating cameras. We present a system for detecting bees and their natural predators for the application of hive activity and intrusion monitoring. Our system uses computer vision methods to detect and localize bees/predators present in an image. We implemented three state-of-the-art object detection algorithms – fast region-based convolutional neural networks (FRCNN), you only look once version 2 (YOLOv2), and the single-shot multi-box detector (SSD) – and compared their performance on a dataset of bee and predator images. All three methods performed comparably well in classification and localization in general, while FRCNN worked best at localizing small objects. However, they all struggled with localizing very small objects. It is possible to improve on this by applying the trained models to larger images, since each approach is fully convolutional, however this is at the expense of detection speed and is restricted by memory constraints. With a detection speed of approximately 10 images per second on an Nvidia GeForce GTX 960, both YOLOv2 and SSD were faster than FRCNN by a factor of 10. That said, we found FRCNN far easier to train.
2:20pm - 2:40pm
3D Imaging of a Grain Drying Front in a Scaled Storage Bin
Colin Gilmore1, Kyle Nemez1, Mohammad Asefi1, Jitendra Paliwal2, Joe LoVetriJoe.LoVetri@umanitoba.ca2
1151 Research Inc, Canada; 2University of Manitoba
Post-harvest, grain is stored and commonly dried in on-farm bins. The careful control of drying and storage of the grain has significant effects on its quality. Recently, we have introduced a 3D imaging technology that creates an image of the permittivity and conductivity of the bin contents e.g., [Asefi, et al. Computers and Electronics in Agriculture, vol. 119, 2015] [Gilmore et. al. Computers and Electronics in Agriculture, vol. 136, 2017]. This imaging technology has applications for both monitoring and controlling of the drying process, as well as for long-term grain monitoring to ensure no losses occur during storage.
In this presentation, we show the images of a grain drying experiment carried out on a small-scale storage bin (approximately 200 L of storage). Very wet wheat (>22% moisture content, wet basis) was dried with ambient air to safe storage levels (~14% moisture content) over 7 days. The contents of the bin were imaged every 4 hours using custom electromagnetic imaging equipment. The drying process was also monitored with in-situ humidity/temperature sensors. These results are the first ever images of in-situ grain drying and establish that the imaging technology is capable of monitoring and controlling industrial in-bin drying processes. The imaging results are in agreement with previous grain drying studies and will be discussed in detail.