Vol. 86, No. 1
Abstract
Development of a Technique for Sweet Corn Tassel Flowering Status Detection Using Unmanned Aerial Vehicle
[Keywords: sweet corn, aerial image, UAV, object detection, YOLOv5, artificial intelligence]
Novel object detection technology to discern the growth stages of sweet corn tassels was developed. Training of two models, the Coloring-model and the Flowering-model, was performed using YOLOv5. The models detected the growth stages of sweet corn tassels in three categories each: 1） Non-flowering, 2） Early stage of flowering and 3） Late stage of flowering for the Coloring-model, and 1） Non-flowering, 2） Partially flowering and 3） Fully flowering for the Flowering-model. Input images were taken by unmanned aerial vehicle （UAV） then resized in 640 or 1080 pixels before training. The results showed that the mean average precision of the Coloring-model and the Flowering-model are 0.71 and 0.61, respectively. In conclusion, these models were able to predict the expected harvest maturity.
Driving Stability of an Agricultural Articulated Vehicle based on the Closed-loop System
[Keywords: agricultural articulated vehicle, Jack-knife, closed-loop control system, preview distance, damping capability]
In stability analysis of systems, it’s common to consider the open-loop control based on two-dimensional motion equations. However, actual driving involves steering control through the closed-loop system, and accidents occur in the process. Therefore, stability analysis was conducted based on the closed-loop control system. The results of simulation confirmed that although the open-loop system can become unstable, it can be stabilized by the closed-loop control system. However, as the driver's preview distance reduces, the damping capability of the vehicle following a target course deteriorates and becomes unstable. Theoretically, it was clarified that even small-sized tractors used in Japan can exhibit unstable behavior such as jackknifing at high speeds, which can lead to accidents if the driver's preview distance becomes small.
Effects of Weed-detection Maps on Labor-saving during Weed Spraying
[Keywords: GNSS，weed-detection maps, weed spraying, Rumex obtusifolius L., labor-saving effect, decision support system]
This study aimed to measure the labor-saving effects of using detection maps during weed spraying. Average travel distance and operation accuracy were established as the labor-saving indicators and monitored. The use of detection maps improved operation accuracy and reduced the average travel distance by 9.61 % and 8.31 %, respectively. Compared with non-informed workers （without detection maps） who adopted the overall-path strategy, informed workers who adopted the shortest -path strategy had similar operation accuracy and 34.7 % lower average traveling distance. The labor-saving effect of the detection maps was significantly high in grasslands with a weed density above 0.117 /m2. Our study demonstrated the labor-saving effects of detection maps and the optimum conditions for their effective usage.
Development of a Small Multi-Crop Combine Harvester
[Keywords: multi-crop combine harvester, cost reduction, mountainous area, miniaturization, weight reduction]
To reduce the production costs for harvesting paddy rice, wheat, and soybeans in unfavorable terrains such as mountainous areas, a small multi-crop combine harvester was developed. The developed combine was about 10 % smaller in length and width and about 20 % lighter than conventional multicrop combines by introducing new technologies（baffle-plate control, water-repellent oscillating separator, and narrow pitch cutter bar). The developed combine was tested on paddy rice, wheat, and soybeans to study the effects of the new technologies introduced for miniaturization and the harvesting performance. These tests indicated the practical harvesting performance of the developed combine for paddy rice, wheat, and soybeans. The problems in its practical application were also investigated.
Verification of Detection Errors According to the Density of Rumex obtusifolius L.
[Keywords: UAV, weed detection, Rumex obtusifolius L., neural network, grassland, vegetation]