|
Cedar Rapids, Iowa | The InfoAg conference was last week in St Louis, here are some things I found note-worthy.
What I remember about the conference last year was that drones were the hot thing. It seemed that you had to have a drone offering to be cool. This year there were still a few drones, but they weren't the new cool thing. I did see some new image-stitching solutions for the pictures captured by a drone. One option required uploading all the pictures to the Internet so a server can stitch the images together. That's going to be a very bandwidth intensive operation, but if we all had gigabit Internet access it would be ideal.
The new hot thing appears to be data management and storage. Everyone has some data management solution in the cloud where you can upload all your yield data, soil sample results, planting maps. There are some that have the ability to do unique stuff like live crop-scouting reports, rainfall monitoring, and a pile of other things. Of course everyone says you still own the data. The services that also aggregate the data in a larger scale claim that yours will be anonymized, but someone pointed out that removing a farmer's name doesn't matter when all the data is geo-located. If you know the lat/lon of a data point, you can figure out whose it is. If you truly make this data anonymous, there will be nothing left.
CAN bus is becoming a topic of heated discussion. Farmobile had a presentation on how they're doing automated yield data logging directly from a combine by sniffing the CAN bus and listening to the sensors directly. They're bypassing the display in the cab. The OEMs don't publish the protocols that the CAN sensors use, so it took a while to reverse-engineer the messages. That spawned the question of who has ownership of the data - the farmer or the OEM - if the data was obtained from proprietary sensors in an unpublished format. In my opinion, the answer is simple: if you are in control of the hardware, you own the data that hardware produces. This is the way it works with copywright of camera images - whoever clicks the button to take the picture owns the picture. The camera manufacturer doesn't own the pictures you take with a camera they made.
In my opinion, the OEMs are doing a poor job of data management, and because they want to be all things to everyone, they won't publish the sensor protocols to allow third parties to do data management. This stifles innovation and is partially why ag data is such a mess. As an example, look at how hard it is to transfer an AB line from one monitor to another. The only data needed to do this is lat, lon, and heading. Yield monitor data is far more complex. We're a long way from ag data being standardized.
As for data logging, the best line I heard was "Never trust the humans". Basically it means if you're collecting data, you need to be working with data only from sensors. If a human has the ability to do something with the data, you have to assume they did something to mess up the data.
-Lance | |
|
InfoAg 2015 Comments
