In the mining and quarry industries, rock fragments are often transported by conveyor belt. To estimate the qualities of the rock fragments (i.e., size distribution and shape), an operator must take samples from the conveyor and manually perform a sieve analysis, which is time-consuming and tedious. Alternatively, an image-analysis system can be used to monitor the fragment quality on a fast moving conveyor belt. However images of fragments on a fast moving belt vary to such a degree that the quality of any two successive images are not the same, and some of the images are of such poor quality that they are unrecognizable by human eyes. If poor-quality images are directly analyzed by a segmentation algorithm, the analysis will not be satisfactory. To resolve this problem, it is necessary to use an image-classification pro-gram for the automatic selection of fragment images with good quality. Because of this need, an image-classification program was devel-oped and tested both in the laboratory and in the field. Testing of the system on different types of fragments showed that the system works reasonably well.
展开▼