OCN Filter Improves Results Compared to RGN Filter
With our RGN (Red+Green+NIR) filter being the most popular model sold for Survey3 we decided to work to improve its results. We ended up with our new OCN (Orange+Cyan+NIR) filter. The following discusses how the OCN filter will improve your results when compared to the RGN filter.
First let's look at the differences in the filter transmission of light that the sensor collects:
First thing to notice is that all 3 of the bands are different, even the near infrared (NIR). Instead of red light it captures orange, instead of green it captures cyan (blue-green) and instead of 850nm NIR it captures 808nm NIR. The band widths are also slightly wider to provide more contrast.
Speaking about contrast, one of the biggest issues with capturing light to compute the NDVI index is that when you're also capturing soil you get a lot of noise in the pixels. Soil has a lot of red light in it so we shifted to capturing orange light instead to reduce the soil noise. In the below comparison photos notice how the soil (colored red) seems to always be surrounded by yellow. That's soil noise pixel blurring/cross-talk. Notice on the OCN image how there is a much more defined transition from red to green pixels, just like sparse vegetation in soil should be. Also notice that the RGN filter blurs/removes some of the smaller vegetation that the OCN filter was still able to capture from 400ft (120m).
You can also notice that within the green vegetation regions that the RGN has less contrast, meaning how well defined the dark to light green regions are shown. This has to do partially with how the color lut (green to yellow to red) is adjusted, but since we've tuned the soil to be red you can see you lose contrast within the vegetation itself. The distribution of dark green pixels in the RGN image is likely the shaded vegetation, whereas with the OCN filter it likely means the vegetation in the darker green areas is more reflective to NIR light (so more healthy/vigor).