Abstract

During on line yield mapping of potatoes with machine vision on a potato harvester not all potatoes will be free from dirt tare. Proper yield estimation requires a good estimation of the size of these potatoes. The objective of the work described in this paper is the development of a method to assess the volume of these potatoes. The potatoes are approximated by ellipsis. This enables the reconstruction of the invisible part of the contour. The center of the ellipsis (potato) is determined with three tangent lines at three points on the contour of the visible part. When the center is known, other parameters required to estimate the volume of the potato (width and projected area) can be determined. The average error in the estimation of the potato center is 2.6 mm (8.4%). The average errors and standard deviations for the major axis and the minor axis are respectively 6.1% (4.4%) and -4.3% (3.9%). The approximation of the projected area by an ellipsis shown an error of 5.1% (st.dev. 4.6% ) when the real data from the potato are used. This error increases to 6.4% (st.dev. 6.2%) when the estimated values for the major and minor axis are used. Finally the average error in the volume estimation is 1.0% (st.dev. 9.2%). The model with data of completely visible potatoes gave an estimation error of 0.27%. Approaching the potato by an ellipsis based on a visible part of the contour gives an increase of the error but the increase is relatively small. This means that there are good perspectives to estimate also the volume of potatoes partly covered with dirt tare.

Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call