|
| I have worked quite a bit with trying to use remote contorlled airplanes to get NDVI images. The first problems is the camera. You must have a camera that can record red (which most cameras do) and NIR (near infrared - to calculate the NDVI ratio), which most cameras don't. There is a way to get around this. Most cheap and in-expensive cameras can do this, but you have to break into the camera and remove a small filter behind hte lens that most manufacturers use to knock out the NIR readings (because they don't want this part of the spectrum, causes glaring, etc). The camera will then see the NIR part plus the standard red, green, and blue parts (in a combined fashion, the NIR is on top of the blue, green, or red reading). Now to get the original red reading, you need to have a second cmaera that is the same as the first, then you can record the normal red with one camera, and the red +NIR with the second camera, and use some subtracting schemes in imaging software to get the correct values to calculate NDVI, so your plane has to run two cmaers side by side (there are expensive cameras taht do this automatically with one cmaera, but they are usaully too heavy (Duncantech) or too expensive (Tetracam)). The next problem has been taht to typically get an entire field, you must fly high enough get refernce points (intersection around the field, trees, etc.) that will be used by the GIS software to geo-rectifiy the image. This causes problems with RC airplanes also because most of the time you must fly around 10,000 feet or more jus to get a 100 to 200 acre field in one image (way to high for most RC airplanes which usually only fly between 500 to 2000 ft. high). So for this reason some way is needed to take lower, samller area iamges and be able to put them together, which is where the big cruch is. Most of the time, in the middle of the field, there arn't any refernce pioints in the image (only rows of crops that look like every other image), so you must now have some tpye of goerefenceing system and some tpye of software to put it back together. This really doesn't exit, which is what i have been working on. I have been able to succesffully construct a VB program taht takes the two images and recombines them to put out an NDIV text file based on pixels location, but there are still some problems with multiple images, in that the constrast doesn't always stay the same from photo to photo (since the plane is alwasy titlting during turns and sun blinds the camera sometimes). | |
|
Best Infrared Imaging....
