Shape Sorting Adds a New Dimension to Performance07/01/2011 Shape sorting is one of Bühler Sortex's core technologies. It refers to the detection of defects either by shape alone or by some combination of shape and colour. The technology is available on the company's full range of bi-chromatic optical sorters. Basic concepts
In most applications, objects are sorted simultaneously by colour and shape. During shape sorting each object is classified good or defective by its silhouette which is its two dimensional shape. So a round object, such as a pea, is seen as a circle rather than a sphere.
Two points must be checked in preparation for shape sorting. The first is the product stream's feed quality. Product flow past the sorter's camera should be uniform, ideally with each product item separated from the rest. In practice, some touching product will occur. If product is approximately round in shape, the object separation technique can be used to separate touching product. Secondly, check that the product silhouette is a true image. Misleading results can occur if the sorter's glass is not kept clean or if there is excessive water in the product flow.
Case study 1 – Stalk and stem removal
Stalks on green beans are detected as objects or parts of objects thinner than a set width. An additional benefit of shape sorting is that the sorter ejectors are aimed at the entire object and this more precise aiming results in a more concentrated reject. Figure 1 shows the concentrated reject after shape sorting of green beans.
Case study 2 – Pods from peas
Shape sorting is commonly applied to removing pods from peas. Pod shells, often the same colour as peas, can be identified by their triangular or rectangular shape. Peas are processed at high throughput and as this increases so does the percentage of touching peas. Object separation works well for round product such as peas and has led to a significant increase in typical throughput. A highly efficient level of pod rejection with minimal loss of good product has been maintained.
Case study 3 – Larvae in berries
An important application is the removal of larvae from soft fruit red berries. In the field, there is a clear colour difference between berries and larvae. In the processing line, however, larvae are stained by red berry juice. As the berries reach the sorter there is little colour difference between them. But the 2D shape of berries is circular and that of larvae, elliptical. This difference is an adequate visual trigger to allow the removal of larvae by shape sorting alone.
Case study 4 – mis-shapes and pac-men in carrot slices
In processing sliced carrots, mis-shapes are defined as items too long to be a single carrot slice and pac-men are identified as having a high curvature as shown in Figures 5 and 6. The operator fine tunes sorting performance to deal with these via two sensitivity controls and a third is deployed to deal with colour defects. By fine tuning each separate sensitivity control an ideal balance can be maintained between maximising yield and minimising rejection of good product.
The sorter's GUI includes an 'Activity' value for each sorting criterion which is updated automatically every 10 seconds with an estimate of the percentage of items rejected under each criterion. As a result the operator can see at a glance which sensitivity control requires fine tuning. Consequently, the Activity values make the sorter easier to adjust.
It is clear that shape sorting technology deployed in the Bühler Sortex range of bi-chromatic optical sorters delivers many performance enhancing benefits.