This magazine isn't online, but you can get it on your iPad, Kindle or Android tablet via NextIssue. The article is a pretty basic overview, with examples of using a Rite Wing Zephyr and a hexacopter along with a Canon s100 modified with the IR filter removed. But it's notable that drones are getting this kind of attention in agriculture.
Comments
Thanks for the info guys.
One problem i foresee with getting the maxmax conversion is that a true NDVI requires NIR light, and this starts at 760nm. The spectral intensity graphs shows that the conversion is not actually capturing any part of this spectrum. It is in fact a part of the spectrum inbetween red and NIR. Is this part of the spectrum actually good for looking at plant health?
If a cheap camera existed which would measure R G B and NIR (760-1100nm), it would be perfect. The tetracam does 760-900nm in the NIR spectrum with R G NIR1 but is expensive and crap resolution. The maxmax conversion maxes out at 760nm, so im unsure how useful this will be for plant health detection.
Does anyone know of any information regarding using these bands for crop health monitoring?
Mark , yeah, I am not surprised at the image quality of the TetraCam. I have not talked to a single person who has purchased one who is impressed with the image quality. The term I often hear used to describe the image quality is "adequate." The comparison between the Tetracam and the MaxMax images is night and day.
This image shows sensor characteristics from a MaxMax-converted S100. It was done using a diffraction grating and halogen light. The spectral calibration was done using a mercury lamp. It should be fairly accurate in terms of wavelength, but the peak magnitudes are not calibrated for the response in sunlight.
I expect you will be happy with the results from a MaxMax camera. They use filters that preserve the ability of the camera to auto focus correctly, and you will not have hot spots or other artifacts.
Many people use the SX260 with good results. I like the S100 (you have to call them to order one). It also has GPS, but it has a bigger sensor, exceptional dynamic range for a point-and-shoot, and it has "DSLR-like" manual controls. Some people have experienced lens vibration wit the S100, but it only seems to be an issue on some aircraft. Probably due to the prop vibration frequency resonating with the floating lens.
I read through the maxmax website and i think i understand it now. Are there any spectral response graphs of these maxmax conversions? I am trying to look for stress in wheat maize rape and peas. Will this converted camera give me the data i need to detect a good NDVI picture? Also, any suggestions on cameras to use. I liked the look of the Canon SX260 as it can be hacked with CHDK and has GPS. Are there any other cameras which are better?
Cheers for your replies btw, i appreciate the help
The NIR+G-2B/NIR+G+2B algorithm was actually created by Dan at MaxMax.com.
I created two related algorithms, one of them has a slight advantage over the MaxMax algorithm in most experiments: (NIR-G-B)/(NIR+G+B)
The algorithm I designed outperforms the MaxMax algorithm because the blue and green bands are strongly correlated with each other, and combining their information usually results in a slightly better dynamic range compared to using the MaxMax algorithm or an NDVI based on a single visible band.
With all that said, however, the difference in performance between the algorithms is usually not significant enough to be a real practical concern.
To produce an NDVI you need at least one visible band and the NIR band. The visible band that is least affected by atmospheric scatter is the red band, which is why it is generally the best band for high altitude aerial photography. Because we fly so low, however, the blue and green bands are useful too. In our experiments the visible bands perform very similarly, but the red band has a slight edge under some circumstances. The quality of the camera and the quality of the conversion are probably more important than the specific visible band you use.
To calculate NDVI you are meant to use the Red and NIR channels to do the calc. http://earthobservatory.nasa.gov/Features/MeasuringVegetation/measu.... I thought i understood all this, but its one of them "the more you know, the more you realise you dont know".
There seem to be several types of conversion you can do. Which one is the most helpful?
Tetracam does a traditional red NDVI, not a blue NDVI.
Does anyone know how the likes of tetracam are doing the conversion of the blue channel to NIR1?
-
1
-
2
-
3
-
4
-
5
of 6 Next