Improvements and evaluation of an infield bin filler for apple bruising and distributions

摘要

Automatic bin filling is critical for the new apple harvest and infield sorting machine developed by our group recently. A commercially viable bin filler for infield use should be simple, compact and low in cost, and be able to distribute apples evenly in the bin without causing bruising damage. An innovative bin filling technology was developed for incorporation with the new apple harvest and infield sorting machine. Field tests of the first version bin filler in the 2016 harvest season showed highbruising rates and uneven fruit distributions. Subsequently, a second version bin filler was constructed with several major improvements, including the addition of a new pair of foam rollers for better controlling apples exiting from the sorting system and avoiding fruit collisions during free falling, and installation of an improved pinwheel with nine longer soft pads, instead offour short ones with the original version, coupled with fruit moving gitides, for avoiding apple collisions and reducing therolling speed of fruit from the pads into the bin, with better fruit distributions. Field tests conducted in the 2017 harvest season showed that the improved bin filler achieved superior performance in reducing bruise damage, with 99% 'Gala' and 98% 'Blondee'apples being graded 'Extra Fancy'. Furthermore, a depth imaging method, using Kinect-v2 camera, was proposed to quantitatively compare the performance of the two bin fillers for distribution of fruit in the bin under the uniform and non-uniform feeding conditions. Analysis of the fruit height data showed that the apple distributions were not significantly affected byfeeding method for both bin fillers. Overall, the second version bin filler resulted in better distributions of apples in the bin, compared to the first version, and uneven distributions mainly occurred in the corner areas of the bin, which could not be reached by the bin filler's pinwheel. The improved bin filler has met the requirements for apple harvest and infield sorting, and it has potential for use with other harvest platforms.