The count of the detected with trueGeiger (and any other radiation detector for that sake) events is following Poisson distribution. For the Cs-137 experiment described on the Testing page, we got the following distribution:
The histogram shows results of N=71 tests, 50 still frames each. The histogram of the number of events probability (bluish grey bars) is plotted next to Poisson distribution (light sandy bars) with the same mean value = 0.17 that we obtained in this experiment.
Any radiation detection is probabilistic, and with low events count we obtain lower accuracy. So trueGeiger estimates the probability of presence of radiation level additional to the normal background level basing on the number of (usually very few!) detected events. Hence the accuracy usually is not high, and that’s the good news. Any value greater than 1.0 may (but not necessarily would) indicate elevated level of radiation. Higher number of frames shot during single measurement increases accuracy of the method. However, if you get readings greater than 1.0 reliably and reproducibly with trueGeiger in ten-fifteen seconds of exposure, then your got a real source of radiation. It’s intensity will be on average proportional to the readings value.
Reading of level = 1.0 means that the combination of the camera and the app didn’t detect anything beyond normal background level, which doesn’t guarantee that the radiation level is normal or even safe!
Calibration (i.e. graduating the method in meaningful units) of trueGeiger presents considerable challenge. Primarily we used several dental x-ray samples to set the main reference point for readings from the camera to the known radiation level.
In experiment with the ion-based smoke detector and front camera trueGeiger algorithm gives readings at about 3-6 times above the normal background radiation with the devices we tested. These readings are based only on few decay events detected. Some runs didn’t detect any elevated levels of radiation at all due to the randomness of the decay and the detection processes. However estimated levels at about several times higher than the normal are reasonably close to the expected values.
Again, we’d like to remind that trueGeiger is much more sensitive to piercing radiation like gamma- and x- rays than to alpha or beta radiation. For example it fails to detect any additional radiation from Potassium-40. This isotope is contained in the so called “No-Salt” product from grocery store used as table salt replacement. It should emit certain amount of beta radiation. Elevated radioactivity of that product can be easily detected with regular Geiger counter but not with trueGeiger.
Bottom line: if you consistently get 1.0 then it doesn’t mean much, it doesn’t even guarantee a safe level of radiation (i.e. we believe there are no false positives). If the readings, obtained when following suggested measuring techniques, are consistently higher than 1.0 then you got some indisputable source of radiation!