I recently put my Bushnell Aggressor Wireless Trail Cameras back in service after giving them almost a year off. The cameras were temporarily put into semi-retirement while I attended to a few other pressing matters. But now they are back online and ready to record more excellent wildlife photos.
For those who are not familiar with these cameras, the Aggressor model trail camera comes with a built in cell phone. When the camera records an image in the field, it can then use the cell phone to upload the picture to the cloud. Once the photo is in the cloud, you can easily view it in near real-time by simply logging into your online account.
The remote access feature of this camera is invaluable in and of itself. The ability to review the pictures your camera has recorded in near real-time allows you to quickly determine whether the spot you chose to monitor is likely to result in a productive set or not. Further, checking your camera remotely, prevents disrupting the site, so the animals remain comfortable returning again and again.
When I fired up my Aggressors for the first time in months, I noticed right away that Bushnell had upgraded their cloud app with some cool new features. Most notable of these is the auto tagging of pictures as they are uploaded.
This is a potentially powerful new capability. Bushnell is using some advanced AI to analyze each picture with the objective of identifying and categorizing what has been photographed. As the picture is processed, the system tags your picture with its best guesses. The technique is likely very similar to what is described in this article: What You Need to Know About Image Auto-tagging.
You can read more about a similar technology implemented as a service designed specifically to manage trail camera pictures by following this link: Trail Camera Photo Recognition and Auto-Tagging Now Available!
Companies like Facebook and Google and others have been working to perfect photo recognition and auto-tagging technology for years. Their engineers are starting to get the hang of it. Tagging accuracy is reported to be very good. Some where around 90% or so for easily identifiable animals like deer. Here are some examples from a few test runs I did. I was surprised to see how well the technology worked.
I was especially eager to see how the system would do identifying the animals I recorded when those pictures started coming in. As you can see, the initial results were pretty respectable.
But with animals not a easily recognizable as deer, the results can be a little mixed. There is definitely room for further refinement.
For several years now I have been talking to friends about how cool and useful it would be if trail cameras could be set to only record pictures of certain animals. No more wasted disk space on false triggers. No more sorting through dozens–or even hundreds and thousands–of unwanted pictures, just to find the one or two you are really interested in.
With this powerful new tagging technology, the day trail cameras can filter out unwanted images may soon be here!