SproutCam
Advanced camera system with image processing based on Deep Learning.
Do you recognize the following limitations of current optical roller sorting systems?
- Lots of manual labour is needed after optical inspection because only a fraction of the defects is removed;
- Rejection of 2-5% of the good sprouts reduces the economic return;
- Limited belt speeds due to use of older computer systems;
- Hard to adjust machine settings and unclear whether a change results in an effect;
- Hard to sort according to customer needs;
- Occurrence of double sprouts in a roller position;
These limitations formed the motivation to develop the SproutCam: a new advanced camera system with many viewing directions and an image-processing architecture based on deep-learning.
Highlights of SproutCam:
- 50% less good sprouts are rejected, hence more sellable sprouts
- Labour saving due to removal of twice as many defect sprouts, up to 95%
- 100% omnidirectional view on the sprout from two sideviews and a topview camera
- 16 colour pictures per sprout!
- 30% more capaciteit due to an increased belt speed
- User-friendly control
- Attractive pricing as compared to belt sorting systems
- Automatic speed control of the infeed to avoid double sprouts
- Short return on investment
Customer experiences with SproutCam:
“SproutCam removes much more thrips damage without removing any good sprouts”
“If I get to sort a bad field then I used to operate slowly in order to remove the remaining defects by hand. Now, SproutCam removes all of these and additionally my capacity increases!”
“During periods with good prices I can deliver much more sprouts due to the increased capacity”
“My employees adjust the sorting settings themselves with little training. In the past they were reluctant to adjust the machine in order to avoid bad sorting settings”
“The sorting statistics give insight in my daily sprout quality, which allows me to anticipate more and deliver good quality sprouts from risky fields”