Images of a disposal site, called a node, is analyzed to locate the three bins that might have been shifted, rotated and occluded by trash. The strategy is to identify each bin by its opening; an area surrounded by four corners. . The proposed system is trained and tested using a custom database that contains 200 images where half of the images are used for training and the other half for testing. The waste bins in the images can be shifted, rotated and occluded.
Funding
High Impact Research Grant Scheme (D000016-16001) under the Ministry of Higher Education, Malaysia.