An Efficient X-Ray Projection Simulator of 3D Fruit Shapes for Use in Non-Destructive Internal Quality Inspection

To preserve the quality of fresh apple and pear fruit after harvest, a controlled supply chain is necessary. However, so far detection of internal disorders has been difficult non destructively on sorting lines (e.g. internal browning, water core, bitter pit, cavities). A combination of prior knowledge based 3D shape modelling and X-ray imaging has been proposed to successfully detect internal disorders in apple fruit. However, to reach commercial grading speeds, advanced image processing algorithms are required. In this contribution, we present how 3D shape measurement, parameterized shape models and a forward projection simulator can be combined effectively to simulate an X-ray image of a random shaped fruit in any orientation. In the inspection system, this simulated image can then be compared with a measured X-ray image and internal defects can be detected based on the difference between these images.

Publication year:2019