High Speed Vision-Based Quality Grading of Oranges

We describe a novel system for grading oranges into three quality bands, according to their surface characteristics. This processing operation is currently the only non-automated step in citrus packing houses. The system must handle fruit with a wide range of size (55-100mm), shape (spherical to highly eccentric), surface coloration and defect markings. Furthermore, the point of stem attachment (the calyx) must be recognised in order to distinguish it from defects. A neural network classiier on rotation invariant transformations (Zernike moments) is used to recognise radial colour variation, that is shown to be a reliable signature of the stem region. This application requires both high throughput (5-10 oranges per second) and complex pattern recognition. Three separate algorithmic components are used to achieve this, together with state-of-the-art processing hardware and novel mechanical design. The grading is achieved by simultaneously imaging the fruit from six orthogonal directions as they are propelled through an inspection chamber. In the rst stage processing colour histograms from each view of an orange are analysed using a neural network based classiier. Views that may contain defects are further analysed in the second stage using ve independent masks and a neural network classiier. The computation-ally expensive stem detection process is then applied to a small fraction of the collected images. The succession of oranges constitute a pipeline, and, time saved in the processing of defect free oranges is used to provide additional time for other oranges. Initial results are presented from a performance analysis of this system. Most of the processing of fresh fruit in packing houses is highly automated. Machines are used very eeectively for operations such as washing, waxing, sorting by size and/or colour, and packing. However the most important step in the process, namely inspection and grading in quality, is still, with very few exceptions, performed manually throughout the world. This step has not yet been fully automated in production systems, as it requires fast and complex image analysis. Although grading of fruit shares several common features with more classical automated inspection of manufactured goods, this problem is signiicantly more diicult due to the wider range in variation found in natural products. Automation of the grading process is expected to reduce the cost of this important step and to lead towards the standardisation of grades of fruit that is desired by international markets. The inspection process must grade individual oranges into a small number of qualities …