Development of 3D image acquisition system and image processing algorithms for the characterization of the ejection parameters of fertilizer granule
Development of 3D image acquisition system and image processing algorithms for the characterization of the ejection parameters of fertilizer granuleDevelopment of 3D image acquisition system and image processing algorithms for the characterization of the ejection parameters of fertilizer granule - Department of Biosystems Engineering - Ghent University
Development of 3D image acquisition system and image processing algorithms for the characterization of the ejection parameters of fertilizer granule
Summary
Goal
Centrifugal spreaders are used to distribute fertilizer grains as correctly and precisely as possible in the field. This limits losses, which spares the environment and enhances the production efficiency. But delivering the exact quantity at the desired location is still a challenge. The farmer does not have a tool to accurately measure the spreading pattern in a quick and simple way. Even with a spreader that is adjusted properly, a perfect spreading pattern is not guaranteed: differences in the properties of the fertilizer grains, wearing of the spreader, the operator, or even external factors like wind or the ruggedness of the terrain can influence the result. The aim of this research was to measure the spreading pattern of the spreaders during operation in real-time to enable adjustments of the spreader and to achieve a better result.
Method
When ejected from the spreader, fertilizer grains can be considered as projectiles. To research their spreading pattern on the ground, we made a ballistic flight model. The speed of the grains and their trajectories were estimated to calculate their final destination. Both parameters (speed and trajectory) were estimated in two ways. The first method, using 2D camera images and a motion estimation algorithm, did not seem suitable for spreaders with a concave disc. The second method used 3D stereovision and a zone based matching algorithm, combined with an additional algorithm where movements were estimated based on 3D images. This seemed appropriate for any type of spreader. The efficiency of the developed algorithms was validated based on the real spreading pattern that was measured using a collector specially developed for this purpose.
Results
In this project a technique was designed to measure the spreading pattern during spreading with high precision. In the next phase, the pattern could be adjusted in real-time. In that way the farmer would have an extra tool at his disposal for an optimal fertilization of his crops. This would not only improve production efficiency, but it would also take better care of the environment.
