100KM

Remote sensing with UAV for biomass estimation

3689536775?profile=original

We have just completed a pioneer project with Dr Chue Poh Tan from ETH university, Zurich for her postdoctoral research on tropical rain forest biomass estimation with low cost UAV.

Flying in the middle of dense tropical rain forest is very challenging. Looking for a suitable takeoff and landing site is almost "mission impossible". But we are blessed with a team of highly knowledgeable forestry department staff lead by a senior officer. Finally we have two site to lunch the UAV.

I'm using my Skywalker base UAV with the same configuration as this, But I have upgrade it with the APM 2.5 and ublox GPS and 2.73 firmware. The new navigation controller is really amazing. My plane now trace perfectly along the grid line. It also hold altitude better than previous version. I'm totally happy with my current set up and a big thank you for the arduplane development team.

3689536888?profile=original

The real challenge of flying in forest is definitely landing. Although we may have a suitable site for landing but it is all surround by tall trees. The window for approaching is very narrow. Most of the time we have to do a steep dive to lose altitude once the plane pass those tall tree. 

We fly a total of 8 mission in three difference area including one using a modify near infra red camera. There are a few thousands geo-ref image waiting for the Doctor to analyse. I think she will be busy for sometime.

Here I share the tlog from one of the mission and you can see how well the new controller is.

2013-07-18%2010-26-32.tlog3689536913?profile=original 

E-mail me when people leave their comments –

You need to be a member of diydrones to add comments!

Join diydrones

Comments

  • I want to know about the processing for individual tree crown area detection and delineation. 

    How are you performing this process?

  • 100KM

    Hi Uttam

    Have a look at here, It may be your solution for accuracy provided you have the budget. 

  • Dear Keeyen ,

    Thanks for the info.I have recently completed a project : Generation of High Resolution DSM using UAV images.(http://www.slideshare.net/Uttampudasaini/generation-of-high-resolut...).Sony Nex Series was used and the external orientation parameters as you said the GPS info were provided.But i am afraid they have an accuracy of about +/- 10 m .So how can we get good results at such case?Even if we establish ground control points how is that possible in ares covered with dense trees???

  • 100KM
    Hi Uttam

    We use Canon S100 camera with build in GPS. Every images it take had GPS info in it's Exif file. I am using Agisoft with good result. If you want sub meter accuracy, then ground control point will be needed.
  • Hi Keeyen,

    This looks interesting.But how you guys did georeferencing of the images?As per my experience no commercial software provides good results for image matching and georeferencing in areas covered with dense vegetation.Can you explain it more?

  • 100KM

    Hi Jul

    Nice to meet you here. You can contact me via e mail intrajasa@gmail.com. Good to know someone here in Sabah also using UAV.  

  • Hello Keeyen...i am love to see ur plane and the setting...Im doing a little bitresearch also in UAV and my location is Kota Kinabalu..im and my  research team also doing a remote sensing using small UAV MR ..some area in West Coast .. hopefully we can meet someday and and cit cat about ur UAV .. Abdullah Ghani, i see your profile and ur are from membakut...still in Sabah? or already back to ur University?

  • Hi keeyen,

    Wow..good, I hope my locations of interest are within your rgb data. I know Dr.Tan, I hope she still can manage to derive the LiDAR data. They need some kind of permission to fly. Thanks for your email. I will contact you via it...and I'm so pleased to stumbled upon your blog and meet Sabahan who doing this (uav monitoring)...well done.
  • 100KM

    Hi Abdullah

    Sepilok Forest Reserve is one of the study area of Dr Tan. I have a set of RGB aerial images of the 160 Ha study plot. Currently our drone can only carry a normal point and shoot camera, in my case, a Canon S100GPS which is about 200 gram.

    Dr Tan also want Lidar data for that area but the cost to acquire Lidar data is very expensive ( few hundred thousand RM ) so we are looking into the possibility to use light weight Lidar sensor, put it in a multirotor and fly the area. The sensor we are looking at is  http://www.velodynelidar.com/lidar/hdlproducts/hdl32e.aspx

    You can contact me via e mail, intrajasa@gmail.com. Thanks

  • Hi keeyen,

    Do you think that you can fly it into some part in the middle of Sepilok Forest Reserve? I'm interested to know of your drone capability to derive LiDAR Data in some part of Sandstone hill area in SFR.

This reply was deleted.