I would like to compute the vegetative index named NDVI at small scale (on a green roof) and study the evolution of this index over a year. This index is calculated from both the near infrared band and the visible red band. Ideally, I would like to make automatic acquisitions (let say once an hour).
How can I do that ? What kind of sensors and materials shall I buy ?
You will need a multispectral camera collecting data at least for red and near infrared spectral bands.
I think you need a multispectral camera that collects data, especially red and near-infrared spectral band data. And link the camera through the web to transmit and receive pictures and use the pictures to study the NDVI at least every hour
You can use two cameras NIR and Visible red to capture simultaneously the same location periodically (every hour). These images will be stored in a computer with enough storage capacity. Then an automated application can process images and calculate NDVI.
Another suggestion is to use the method of Helen Anderson et al. in " Using Ordinary Digital Cameras in Place of Near-Infrared Sensors to Derive Vegetation Indices for Phenology Studies of High Arctic Vegetation " where authors use only visible camera to calculate GRVI (Green Red vegetation index) (GDN − RDN)/(GDN + RDN) which showed similarity to NDVI for some plants.
Thanks a lot to everyone ! I found that Campbell Scientific is selling such a camera : https://www.campbellsci.com/ccfc
I do not know how much it costs, but this solution seems to be relevant.
Get a raspberry pie 4 and an NIR camera to interface it with. Second you can write a simple python script in Raspberry Pie and with some basic equipment like power source for Raspberry Pie you can monitor your area of interest with the regular intervals. Raspberry Pie support large memory cards which to be sufficient for you to capture your data.
Your data requirement as you mention is 24 x 365 = 8760 NDVI images.
Suppose each NDVI signature takes 4 Mb, which is just an overestimate, the total space required for 1 year will be approximately 35 Gb, while Raspberry pie 4 supports 64 Gb of memory cards.
Regards
Have you considered setting up a spectral radiometer to take readings every 15 minutes (or what ever time interval you want --- 5 minutes). I used an Exotec radiometer (1980s and 1990s vintage) with 4 bands similar to the Thematic Mapper and had it pointed at the water in the Colorado River at the bottom of the Grand Canyon for 1.5 years taking readings every 15 minutes. The values were used to look for correlations with sediment loads computed from samples collected twice a month. This included looking at various ratios (the NDVI or any other index can be computed from the data collected).
Pat
- Badges
- Science topic
- Similar topics
- Geoscience
- Asia
- India