UAVs and STARS: The adventures of a master student
One of the main goals for my master thesis was to gather my own data in another country. I would never have guessed that this would turn out in flying an octocopter for the STARS-project in Mali.
I am a student in Geo-Information Science and Remote Sensing at Wageningen University in the Netherlands. Therefore I am highly interested in the remote sensing part of the STARS-project. Crop recognition with Unmanned Aerial Vehicles (UAVs) and satellites seemed to be a doable and very interesting thesis subject. However, my two months visit to Mali learned me that it is not that simple… To be honest, a lot of thing that seem easy aren’t that obvious in Mali.
My first week I was ‘dragged’ into the STARS-project by going on a trip to Sougoumba (Koutiala district). In short, we had some interesting discussions with local authorities, locals and AMEDD about the progress and the use of the project. In addition, we visited the farms and had a data collection demo. In the meantime of this demo, I looked at the crops which I eventually have to recognize with satellite imagery. Within a single field crops can look totally different: small plants, big plants, dark green, light green etc. For remote sensing in the STARS-project the appearance of crops is essential, as we are working with passive sensors that measure reflected light. After this visit I knew the crop classification would not be a piece of cake. Also, I must say that in this week I really learned that my French was not sufficient for Mali. Did I say I thought this would be easy?
Back in Bamako I was placed in an apartment at Sebenikoro with some fellow researchers. At ICRISAT I worked with a great team and had a superfast computer. After solving some troubles with my equipment I started flying and collecting data. UAV imagery has a very high spatial resolution, but the octocopter can only fly for 12 minutes. Which means it cannot cover a large area and it is thus not applicable on the whole area of Sougoumba. Therefore the octocopter was only flown over the test fields in Samanko. The idea was to gain knowledge on small differences between similar looking crops and to find the spatial, temporal and spectral requirements necessary for an accurate crop recognition.
The Samanko team! (Kone missing)
Besides my data collection I trained Kadiatou as an octocopter pilot and yes, I heard that just after I left she flew the UAV alone for an important visitor. Mission accomplished! The best events during my own data collection: The enthusiasm of Oumar and Kone in helping me with the execution of the flight, a grownup hiding behind a tree when the octocopter came close, a bird of prey attacking the UAV (quick landing solved the problem) and the interest of locals in the results during the visitor day.
Although my visit to Mali was mainly for data collection, I also started some analysis. I knew the recognition of crops would not be easy, but with the UAV data some interesting results showed up. It is indeed not possible, to recognize the crops with a computer based on a single image. Therefore I created an extensive dataset over time with the octocopter. It was only possible to monitor the crop senescence, as I arrived in October, but similar looking crops (millet, sorghum, maize) had a different temporal profile. Millet decayed faster compared to sorghum, which might be a useful characteristic in separating these two crop types!
Temporal NDVI trends
Next to doing my research I have to say that I tried to use my spare time well to see Mali. With Cass and Frederic we climbed the ‘mountain’ close to our apartment, visited the famous markets, went to the zoo and had a nice trip to Siby. On those trips I learned that the Malian people are friendly, hospitable and … easy going? Of course you had to break the ice first, we still are a couple of ‘toubadous’ (mzungu/white man). But playing some cards and talking some faulty French was enough to have a nice conversation.
The 'mountain' close to our apartment
All in all, I had a great experience in Mali. I hope I can help improve the information system around smallholders farms with my research. And who knows, maybe I will come back, I still have to see Timbuktu!