This story is part of the 2020 Harvest Value Stories Series. This series includes various real world examples of challenges growers face each harvest season. We want to show the impact that understanding harvest loss differently can have on an operations profitability and performance.
During the 2020 harvest season, the Farmwave Harvest Vision System demo version covered almost 40,000 acres. Growers experienced incredible moments firsthand. For the first time, growers saw loss from the header, providing them with the data to adjust settings in real time to reduce actual loss from their combines. In 2020, our visits to farms meant the FHVS encountered a variety of environments, terrain, and crop conditions. The variance in crop conditions we experienced were very wide. Crop conditions ranged from 9% moisture corn, showing greater than ten bushels per acre of loss, too late plant soybeans at 18% moisture showing a bushel per acre loss. Most importantly, our field-tested data allow us to build resilient, accurate, and trustworthy harvest loss AI models for real world decision making.
- Crop: Corn
- Crop Condition: Good, very dry
- Grain Moisture: 9%
- Location: Southwest Iowa
- Terrain: Flat Land
- Weather: Overcast and Cold
- Machinery: John Deere S680
- Header: John Deere 612C
- Field Acres: 1000
It was a cold, wet day in late October. The corn harvest had not begun yet due to the crop being very damp. The field was about 1,000 acres with long runs just under a mile long. A quarter of the field was already harvested from the previous days. The operator in the first portion of the field thought yields were good but he could not accurately identify how much loss was being seen from the machine vs the header. Given the low moisture content, he knew there was header loss but wanted to be able to quantify his header loss vs combine loss.
We began the runs with three AI modules on the header to get the best understanding of what was happening. It was very clear early on that there was a tremendous amount of shatter loss at the header. The system was seeing 17-20 kernels per square foot, or roughly ten bushels per acre being left at the header alone.
The operator closed the deck plates down as tight as the headers would allow to try to lessen the shatter loss. After doing so, we were pulling more material into the combine but were able to reduce the header loss by about three bushels per acre.
The next focus we took with the vision system was the rear of the machine. We mounted two cameras up under the spreading unit to see the kernels in flight as they left the machine. The combine itself was performing pretty well with the system showing about three bushels per acre of corn out the back of the combine. Once the operator was able to see a more accurate loss calculation, he began to make changes to the machine settings, which resulted in additional retention of about two bushels per acre. This brought the loss down to just under a single bushel per acre being seen out the back of the combine.
Once our time in the field was completed, we were able to show the grower an instant savings of roughly five total bushels. The breakdown of these savings? About three bushels per acre at the header and two bushels per acre out the back of the combine. At today’s corn prices of $5.25 per bushel, savings were about five bushels per acre total for a total savings of over $19,000 for the remainder of the field.
The other step the grower was potentially going to take to retain more grain at the header was to add a gathering chain brush system. We were not able to get back out with this grower once the gathering chain bush system was installed.
"The technology is incredible and gives us an insight into our harvest process that we have never had before. The ability to understand loss at the header vs the combine is a major step forward for our operation in improving yields and margins." - Southwest Iowa Grower