I have been using the AS7265X Triad, along with some AI models to try and detect certain chemicals. However, after training my models, my data suddenly has changed on some channels by over 1000, doubling in some cases, and decreasing in others. I have an enclosure designed to reduce interference from ambient light. However, the data has abruptly changed multiple times, even with the exact same sample. Can you please help me with standardization. Is it possible to work with my old data with some editing, as I am short on time?
I would take a look at maybe redesigning the enclosure…and/or run a calibration routine on some known-wavelength samples
I’m not sure if it is being moved around when in-use but if so that might affect readings…try keeping it stationary to minimize outside light as a variable
1 Like
What might I use for calibration? How would I implement this?
You’d just collect data from a known source sample for light character and sample purity (while controlling the variables as much as possible) and gather data across several points in time…and you can adjust the data in post-processing to ensure your readings are consistent
Grabbing a few decent LED sources (small range for their wavelength output) and a couple metal samples worked well for me
Is it okay if I just use a piece of cardstock and the built-in illumination?
Sure - it all depends on your needs