Bottle type detection for returned empties

The demands made on returned case inspectors are constantly increasing. New bottle sizes and bottle shapes are being designed almost on a daily basis. The companies filling returnable containers get this assortment of containers back thoroughly mixed up in the production cases. It is necessary to sort these in the case area in order to provide the filling lines with a sufficient number of empty bottles. Foreign bottles must already be sorted out here.

A case inspection has to be equipped with the most varied detections adapted to the container material of the production container in order to solve this task.  The height of the bottles is usually measured using several ultrasonic sensors and the colour of the bottles identified with a colour camera.  Even the closure can be an important detection characteristic and is evaluated by means of camera and image detection.  A camera picture of the neck ring is analysed as well in the case of PET bottles.  Glass and PET bottles can also be differentiated by means of X-ray measurements.

Finally it has to be determined whether the bottle is suitable for production or not on the basis of all the information obtained from these detections.  Furthermore there is a tendency to repack the non-production bottles according to their type.  Today the HEUFT LGX returned case inspector therefore no longer only identifies the currently produced container but it also identifies a multitude of different containers and can emit a special signal to a downstream packer for each container type.
That this is possible in the somewhat difficult conditions in the returned empties section is achieved by HEUFT SYSTEMTECHNIK GMBH by means of a cleverly devised combination of different measuring methods.  The result of each individual detection is evaluated and an overall picture defined.  The degree of probability that this is a container of the type A is derived from this.  The colour detection will produce a false result if the container has spent three weeks on a building site and is covered with a fine layer of plaster but if the neck ring, closure and height clearly argue in favour of it being a container of the type A then it will be identified as such.
A decision which is made by a human by evaluating different criteria is copied by a machine in this case.  The result is an amazing detection reliability which makes a cost-saving efficient sorting process possible.