Post Reply 
 
Thread Rating:
  • 0 Votes - 0 Average
  • 1
  • 2
  • 3
  • 4
  • 5
Automated Water Quality Tests
01-27-2012, 04:54 PM
Post: #1
Automated Water Quality Tests
I know you guys are really busy right now with Frag Swap but I thought I'd post another suggestion.

All my previous suggestions haven't been novel since you are already working on them. This time, I think I have a doosey. It actually might require a separate app though.

Here goes. Using the iPhone camera to take a picture of a water test result, have the app interpret the color information and compare against a preset standard curve. This will allow interpolation between the sometimes very large gaps in color chart concentrations. This would require controlling the color balance on the camera and probably some sort of simple DIY template to position your sample vial such that the picture is relatively constant. Consistent lighting will be the biggest issue, however, all iPhones have a flash. Thus designing a template which uses the built in flash should keep things relatively consistent. The DIY template could be a simple thing to print out one time and fold on the dotted lines. Place vial in circle on the paper, place camera with flash on left/right/top etc. at the thick bold line. Then snap picture via the app. The app could control exposure, color balance etc. assuming this template. The resulting picture should have a consistent color spectrum that would be interpretable by the app. You guys could build an initial calibration curve by mixing standards of NH3, NO2, NO3, PO4 what have you. Then the color sampled is compared with the curve and the ppm estimated.

For info on color change values I found one app called Delta E which might give you some info on calculating change.

The preset curves could be simple at first and expanded slowly by you the developers and by users who could submit them to you. At first the most common test kit like API saltwater or freshwater could be done.

This could potentially be huge. I know many who hate trying to decipher between some of the colors.

So there you have it. My automated spectrophotometer idea for water quality test kits.

Now you can feel free to tell me how extremely high my expectations are and how crazy I am. Cool
Find all posts by this user
Quote this message in a reply
01-27-2012, 05:54 PM
Post: #2
RE: Automated Water Quality Tests
Hello VisualPhD,

I also routinely take photos of my test kits with RTP and sometimes compare them. There is a new feature coming in Stats and Graphs that allows the data behind the graph to be displayed in a scrollable list. This feature will provide the ability of displaying a bunch of entries of the same type and if photos of test kits were taken, then you can see color changes over time. This isn't as advanced as what you have mentioned. In the future we are hoping to implement a search (another suggestion you've also mentioned) that would perform basically the same type of data aggregation as Stats.

I have sometimes wondered if there might be a way to auto determine value from the color of a test kit. I have an app that can check my heart rate using the iPhone flash light and camera and I'm amazed by the accuracy. I haven't given this as much thought as you have. It's definitely a good idea.

Presently all the tests in RTP are designed to be extensible in order to account for different types of tests. Probably the most difficult part would be handling the data entry of the color spectrum and calibrating it to corresponding values. We are considering for a future update the ability to download entry packs for the application. This type of functionality could provide the ability to install (for lack of a better term) already calibrated test kit color cards (API would definitely be first to be included).

It is possible to overlay a template within iOS when taking a photo that could represent the shape of a test tube. This would ensure that the photo was lined up correctly.

I imagine that this may be possible but a proof of concept would need to be put together to determine how feasible and accurate the numbers would be. As you spoke of the consistency of lighting would be important. I imagine that we would attempt to go by the average HSL (hue, saturation, lighting) for a given rectangle within the photo. HSL would be better than RGB (Red, Green, Blue) in terms of being able to match the actual coloring (hue) of the spectrum being analyzed.

I'm still working on just a few things in order to get the iPad functionality included within the app. There have been quite a lot of changes in Stats, a few minor bug fixes, and we're adding the ability to view and zoom photos in the very near future.


Thank for the suggestion,

Eric
Developer of Reef Tank Pro
Find all posts by this user
Quote this message in a reply
Post Reply 


Forum Jump:


User(s) browsing this thread: 1 Guest(s)