Real-time nondestructive inspection of chestnuts using X-ray imaging and dynamic threshold

Quality defects in chestnuts seriously affect their market value. The defects in nuts covered with hard reddish-brown shell are very difficult to inspect using conventional nondestructive inspecting techniques. In this paper, the possibility of using X-ray imaging to inspect defects inside chestnuts was explored. To accurately evaluate quality of chestnut, a dynamic threshold segmentation algorithm was proposed and a defective area threshold was set to judge chestnuts´ quality. The result showed that the correct recognition ratio of total was 89.1%. One chestnut had been inspected about 27 ms, which is fast enough to be implemented to real-time inspection of chestnuts.