Cloud-based OCR: A sortation solution that increases profits and capacity

Manual intervention is required for 5 percent of the parcels in the sortation system – a 30-second procedure that is both a drain on finances and labour – but now CEP operators can change the narrative with a Cloud-based OCR solution.

By Søren Thing Pedersen

During a busy period, scanners on a parcel sorting system might read as many as 60,000 barcodes an hour.

But for every thousand, there are always a few unreadables. It could be the result of plastic wrapping or a wrinkled, folded or missing label obscuring the barcode.

New Cloud-based technology

The new system uses Cloud-based Google Vision AI technology, which has many advantages:

  • Uses market-leading text detection algorithms
  • Tech is constantly updated by Google
  • No issue with low DPI, image resolution or misaligned images, as Google Vision AI will realign
  • Only charges for what is used
  • Uses fewer manual encoding resources
  • Improves the system capacity
  • Fast enough to be integrated with line sorter technology

The Cloud-based OCR uses a multiple camera system to quickly identify delivery destination data on an average of 60 percent of the ‘unreadable parcels’, resulting in an overall sortation rate increase from 95 to 98 percent in the process.

This 3 percent gain is a huge advantage, as operators only pay for what they use once they have taken care of a small installation fee. It delivers almost instant savings – not only in terms of saving on labour costs, but also by eliminating bottlenecks of capacity.

Any reluctance to invest a lot of money in the Cloud-based OCR technology will quickly dissolve in the light of the low cost of entry and fast return on investment.

Quick, efficient Cloud solution

The Cloud-Based OCR solution enables successful sortation in just three to four seconds.

The Cloud-based OCR solution utilises a trained neural network to detect and crop images of the correct label, which is then sent to the Google Cloud.

Cloud-based Google Vision AI technology is used to extract the readable text from the labels, which is then matched with sort keys used in the sort plan. Sortation logic then completes the process. In total, it takes 3-4 seconds.

It’s so fast that parcels don’t need to recirculate in a loop system. Furthermore, at this speed it also makes it possible to apply Cloud-based OCR to line sorter-based systems.

Perfect fit for wide range of customers

The Cloud-based OCR also increases capacity. Parcels with unreadable barcodes can end up remaining on the sorter for long periods – literally lost in the loop.

And the Cloud-based OCR is also a great fit for companies that have huge fluctuations in their parcel sortation rates – for example, mid-January compared to the run-up to Black Friday or Christmas – which makes it a win-win for both peak and low capacity requirements situations.

During peak scenarios using Cloud-based OCR, the parcel is only on the system once. If Cloud-based OCR isn’t used, there’s a risk it gets thrown into the rejects for manual handling, but as soon as the parcel goes into the system twice, then capacity is lost.

Meanwhile, in the low throughput scenarios you don’t need to have a lot of people on standby in case there is a parcel that cannot be read.


The beauty of the Cloud-based OCR solution is that it is giving both large and small CEP operators the opportunity to save money without the need for much investment – at a time when costs are increasing in almost every component of the business.

Passing on the cost of handling the parcels to customers is not a viable option. Only an internal optimisation of the sortation system will increase the profit per parcel.

It really is a case of ‘every little thing helps’. It might look like a small matter on paper, but as a part of the bigger picture, it will decrease the cost of operation whilst adding to the profit.

Furthermore, using the Cloud-based OCR is a smarter use of resources at a time when staff are becoming harder to find.

Subscribe to our newsletter