Loading Condition Detection of Grain Transportation Vehicle Based on Visual Image

To solve problem of grain loading reliably and accurately during master-slave collaborative operation of harvester-grain transportation vehicle structure, the detection methods of loading status based on visual image are proposed in this paper. The original image of grain bin is acquired by in-vehicle. After noise is eliminated by Gaussian filter, edges of binary image are recognized with Sobel operator, and gradient map of grain and bin is obtained by non-maximal-suppression method. Straight segments of grain bin edge are acquired using line detection algorithm of randomized Hough transform, and then region of interest (ROI) is obtained through pairwise intersection points of four straight edges. Grain region is recognized through colour space conversion and threshold segmentation. Loading status of grain bin is identified by the distribution state of grain with the convex points of grain region. The system extracts clear and complete edges of grain bin and monitors grain loading status real-time and dynamically, which avoids grain scattering outside and improves master-slave collaborative operation efficiency and monitoring capability of harvester-grain transportation vehicle. © Published under licence by IOP Publishing Ltd