A Platform to Build On
Our platform provides the basis to build additional functionality.
AI Fire Detection
We have recently been working with a South Korean company, Alchera, on a pilot using 200 of our cameras in the USA.
In recent months they have made huge strides in the accuracy of their detection. This means that they are detecting most fires and they have very few false positives.
We show some examples below where the human eye would find it very difficult to detect the fire.
This is a really exciting development and is a genuine breakthough. Every other AI vendor that we have worked with has failed to produce an acceptable result, despite some of their claims. We have worked with a lot of vendors.
You can watch their detection in real time on this link.
Calibration and Fire Ground Mapping
We are working with Simtable to calibrate each of our cameras so that each pixel will have a lat long associated with it.
This will alsoow us to provide a ‘hot cursor’ where the operator, or a member of the public, can click on the fire and it will show the lat long. This allows the fire location to be accurately reported.
At present the experts tell us that only about one fire in five is accurately reported. This means that the fire tankers are sent to the wrong location more often and not, and waste time trying to find the fire.
Simtable achieve this by doing celestial mapping of the imagery. They control teh cameras at night and match the location against the stars that are visible. Great tech and a great move forward.
Fire Perimeter Mapping
The ability to determine the location of each pixel enables Simtable to do fire perimeter mapping. In this example below you can see the fire perimeter being mapped on the Maria Fire:
This capability can be married to fire progression analysis.
See also the post on Simtable’s mapping of the LNU Complex Fire.