I like to use a small uav that is equipped with a near infrared camera to take images of various crops showing stressed vs non stressed vegetation (Infrared is good for this purpose). The camera I found is only 7 onces (about 200 g) and it calls for GPS intput to be a true +/-RS232. Additionally, I would like my UAV to be independent in term of missions so that a minimum work is required in the field...Ideally, I would like it to fly itself; so that I can make a file (GPX or shapefile) that shows the exact location of the field/ground where the picture will need to be taken. My last concern is about the stability of the drone/uav and whether images taken with this uav would be clear enough to distinguish between stressed and non stressed crops. Field size I am looking to use it for is 100 acres. By writing my thoughts here, I am hoping that someone in this forum has done something similar and I would very much appreciate the advise...Otherwise, do you guys think I can custom make such a UAV... or does it require advanced skills...
I'm looking for a camera exactly the same as the one you discribed.
Would you mind telling me the make & model of it?
I have not purchased a camera yet. However, I have been doing research and reading about the available options. It seems that some people have been successful using a modified regular camera; mostly Canons (for remote triggering). These modified cameras try to remove the infrared filter that comes pre-installed in all regular cameras, and instead they add a red filter (internal or external) which would block red light and register the infrared signal instead in the red channel. There are a couple of places that offer to modify any camera including lifepixel.com (or on ebay you can find this service for a cheaper rate). Otherwise, you can purchase a "NDVI" camera from maxmax (basically a modified camera like I said earlier). I have not tested any of these options and I am not sure they would do a good job for detecting vegetation stress but I am hoping to test a couple this Spring.
If your budget allows it, then I would recommend looking into the tetracam ndvi cameras as they seem to have been used by scientists for this purchase; they are more expensive but the sensors are built for vegetation stress detection. I hope this is helpful!
I have been reading about the steps involved in building an drone/uav. For my purpose of taking infrared photos high in the sky to see stressed vegetation, I think a number of factors are going to be important. I need the drone to be stable when taking photos so for this reason it needs to be not too light (weight wise), and will need a on-board stabilization system to help. Carbon frames seem to be the most stable from my reading (shock buffering). I am planing to use a small digital camera so it should weigh less that 10 oz. If we assume a 4-5 pounds system is going to be stable at 400ft altitude then I need to come up with the proper frame size, motors, speed controllers, flight controller (it seems that AMP 2.6 has more online help material than the new pixhawk; not sure about naza for autopilot and multi waypoints). The last thing is the batteries and I need to be up there for 20 minutes to fly a 120 acres and get multiple shots before landing. I appreciate everyone's inputs. By the way, I think a quadcopter is going to be more reliable and stable than a hexa or octa. Thanks
Many thanks Essam, I had also read about modifying Canan cams, but most comment that the result is far from optimal.
Apparently the blue filter lets through the red light in a small quantity, but enough to spoil what we're looking for.
I guess the TetraCam, FLIR or FluxData are the only solutions, but prices start at USD 10k. We'll have to do a lot of business, to pay for them.
Try Multispek http://multispek.myshopify.com/
Cheaper, VIS-NIR images at once from a single Gopro like camera. Dustproof, custom optics. For the money best buy.
I have modified Canon SX260 by maxmax.com and the quality of their service is shit. The resulting images are always blurred, not worth a penny.