+7 8202 20-16-35
Mon-Fri 9:00AM – 6:00PM (MSK)

Home Products Machine vision for industry VISCONT.SugarBeet


The system for assessment of the sugar beet quality at the produce delivery stage

VISCONT.SugarBeet is a system for assessment of the quality of the sugar beet located in the open bed of the delivery truck. With the use of the neural networks the system locates the contents inside of a truck bed and detects the presence on the dirt beet tops, chipping and snow on the sugar beet roots. Based on the collected data, the system classifies the delivered produce into different quality categories. The system has the recognition accuracy of above 90%.

Ask a question

During the sugar beet harvesting period the personnel at the agricultural production facility assess the quality of received produce inside of every delivery truck bed and make decision either to store the delivered produce or to processes it right away. The cost of the mistake at this stage is high since wrong decision on the quality of the sugar beet, and, particularly, when storing low quality sugar beets long term, could lead to a spoilage of large quantity of raw products, which, in turn, causes substantial financial and economic losses. The accuracy of manual assessment of the quality of delivered sugar beets has usually been recorded in the range of 60-70%. The developed system that utilises artificial intelligence, however, performs this analysis with the accuracy of above 90%.

VISCONT.SugarBeet is an intelligent system developed by Mallenom Systems, which performs the assessment of the contents inside of the delivery truck bed. With the use of the neural networks the system analyses the delivered produce based on the following attributes:

Number of chipped sugar beet roots Presence of sugar beet tops (and other weeds)
Presence of dirt on the sugar beet roots Presence of the frost damage and snow

To derive the aggregate result for the quality of received sugar beets, each of these four quality indicators is assigned with its own weighted coefficient. Then, the aggregate score is compared against the dynamically defined threshold level by the system which does depend on the utilisation level of production facility. Overall, the system makes a dynamic decision either to Store or to Process the received shipment.

 Purposes of the system implementation:

  • Improvement of the accuracy and reducing the influence of human error for the designation of the delivery vehicles (i.e. the contents that are inside of delivery truck bed) at the delivery stage of the produce.
  • Creating of a planning system for the feeding of the sugar beet to the production facility (by an integration with the corporate information system)

Stages of system implementation:

  • Installation and calibration of the equipment on site
  • Installation and calibration of the system libraries and interfaces to meet the individual demands of the client
  • Integration with the corporate information system
  • Collection of the further data to perform additional training of the model

The system has been implemented at the sugar factory in Tambov region (at RUSAGRO-CENTER LLC)

For the inquiries regarding purchase, please feel free to contact:

Andrey Kalinyak
Manager for international development