You are here

A new approach to potato defect analysis

21 June 2013

In the first of a series of Grower Gateway articles, AHDB Potatoes’s pathologist Glyn Harper discusses a cutting edge research project, on defect analysis, taking place at the Sutton Bridge Crop Storage Research (SBCSR) this year. 

The GB potato industry delivers a sales value of £731m at farm gate and £5.7bn at consumer level, so maintaining the health and quality of stored potatoes is paramount for growers, processors and consumers. 
As the food industry develops, fast and accurate quality control (QC) testing has become increasingly significant at both the grower and purchaser ends of the supply chain.
“QC testing is an important decision tool,” notes Dr Harper. “For example, it is used to check that a crop is suitable for storage and it can help decide which market sector to target. 
“For potato buyers, raw material intake QC decisions are crucial for food safety, ensuring the potatoes delivered can meet the pre-defined purchasing criteria, or define the grading required to meet and deliver on pre-defined contract tolerances or specifications.  This helps the plant or pack house to generate a product efficiently, meeting the needs of both its customer and end consumer.”
New automated QC technology
The University of Lincoln, in conjunction with the team at SBCSR, used AHDB Potatoes funding to develop a prototype system for automatic identification and quantification of potato defects to aid potato QC testing. 
This ‘machine vision’ technology uses artificial intelligence algorithms that automatically learn visual features including size, colour, texture and shape, and can detect unwanted blemishes and defects of potato tubers in real-time. It is highly adaptable to very different situations including QC during food packaging or potato and vegetable sampling and, as a low cost and off-the-shelf system, has good commercial potential. 
Preparing for the commercial environment
The Technology Strategy Board (TSB) is now funding a project to develop commercial applications of the technology. The project is led by the commercial partner Ishida (Europe), in collaboration with Branston, and the academic leads are the University of Lincoln and SBCSR. 
“At SBCSR we are excited that TSB funding will enable testing and development to get the equipment into the commercial environment,” explains Dr Harper.
“Traditionally, quality control assessments are done subjectively, by hand and eye, but our prototype is a low-cost solution which could provide better consistency and reliability whilst speeding up the whole QC process.” The system has been shown to reliably identify diseases on washed samples such as black dot, silver scurf, common scab and defects like greening.”
“The potential speed means that QC results at the purchaser end could be instantly reported back to the supplier,” notes Dr Harper. “This means harvesting or grading changes could be made immediately, so that specifications for further deliveries can be tweaked. This might mean increasing or decreasing defect levels, so they are closer to buyer tolerances thus minimising load rejections or unnecessary grading wastage.
“We will also be evaluating the equipment for production line scanning. For example, for potato or other crops, QC results could be produced from a small sample taken at intake, but the system could also potentially provide a detailed analysis based on a whole load by scanning it on the production line.” This can feed back to the grower to inform returns plus future grading or growing decisions. 
The results would be automatically recorded into a database for record keeping. Dr Harper goes on to say “The database is a key benefit and the saved images are valuable for traceability which could be linked back to field records for agronomic and crop rotation analysis.”
The equipment initially needs a human expert to ‘teach’ the system to recognise healthy potato as well as each of the specific defects, blemishes and diseases to be assessed. Therefore a supply-chain could set the system up to customer specifications, tuber quality limits, contract payment criteria or any other bespoke requirements.
“We will be testing different image types and other sensor data including example colour images, polarised light, laser scatter, X-ray and 3D scanning to extend and develop the range and scope of defect and disease identification,” adds Dr Harper. 
The results from this cutting-edge study will be communicated to levy-payers on completion of the projects. Dr Glyn Harper will be on the AHDB Potatoes stand at BP2013, Harrogate, 27 and 28 November to discuss the projects’ progress and storage best practice with levy payers. For more details about BP2013 go to
In our next issue, Dr Harper will discuss a further project researching the use of volatile chemical profiling equipment to remotely detect disease and provide an early alert on potential storage losses before even the most attentive store manager knew there was a problem. 
AHDB Potatoes provides a free storage advice service to levy payers via its team at SBCSR. Tel: 0800 02 82 111.
How useful did you find this information?
Only logged in users can vote. Click on a star rating to show your choice, please note you can only vote once.
No votes yet