Hi,
I've been trying to figure out what's the best way to accurately plan my flights for photogrammetry use. I'm having a bit of a difficult time calculating the necessary distances between photos at different altitudes, for achieving 70% / 40% overlap.
I have found a few simple formulas for doing this, but they all use the "format size" as data input; and basically, I don't know what it goes here for a Canon sx260 hs. I know the sensor size (6.2mm x 4.6mm), but I guess this is not what I'm supposed to use in the formula, as this would result in taking photos every couple meters, even flying at 100m or 200m.
Any help would be appreciated.
M
Replies
I have had great luck with my SX260 on a hexacopter for pictures. For video I do get jello with the sx260. I do have it hanging from a vibration isolated brushless gimbal.
Hi Miguel, can you post one of the blurred pictures? I have a setup with the same camera and I also struggled to get the results right, basically you need to set the shutter speed (I use Tv Mode), the drone flying speed, and altitude you fly at also have a play in the matter. And another thing I struggled with is the camera focus, I had to fix the focus at infinity and disable the autofocus. Thats why I am asking for one of the pics, I may be able to tell you what the issue is from the metadata and what the pic looks like.
By the way Manuel, I am in Guadalajara, where are you?
Hi @Jamie
Could you possibly explain how you fixed your focus at infinity and disabled autofocuing on your SX260-HS?
I've got the same camera running CHDK but I'm having trouble locking the focus (I've tried CHDK>Enhanced Photo Operations>Override Subject Distance, but this parameter seems to be forgotten when I run a script containing the 'shoot' command).
Thanks, Andy.
One way to do that is to use the kap_uav.lua script. That will disable the auto-focus, which reduces the time before a picture is taken, as it just sets it to infinite. IS should be turned off. That script also calculates exposure and provides better results, because sometimes exposure calculations from the camera are inaccurate when you have strong luminosity changes.
Hello Jaime,
Sorry by the delay of reply.
I now use the manual mode with nice results.
I`m from Porto, Portugal
Thank you to all you guys for the help,
Miguel
Manuel,
The calculations are a bit extensive, but Mission Planner has basically solved the problem. In the flight planner tab you can right click, select Survey (Grid), and then tell it you have an SX260, tell it what altitude you'll be flying at, and then it spits out a grid tuned to what you're doing.
If you're interested in how I calculated this stuff before MP implemented the Survey tool, here is how it works:
Sensor size is 6.2 x 4.6 mm, focal length at minimum zoom is 4.5 mm (if you take a picture then look in the metadata the focal length (f) will be there). Angle of view is calculated like this:
arctan(sensor dimension/(2*f))
Note that this is the angle away from the center of the image, not the full angle of view. This is because our trig works best with a right triangle.
Now, we need to convert this to field of view, which is just more trig:
2*tan(angle of view)*altitude
So, according to my math, if you are flying at 100 m and the camera is mounted with the long axis of the image perpendicular to the flight path (probable if you're using a copter, improbable if using a plane), you need 83 m between flight paths in the grid and 30 m between triggers. At 200 m it would just be 166 and 60.
Max
Thanks Max and Gerard, I didn't knew of this option in Mission Planer or the trigonometry formula. What I did after writing my last post was this:
I tried to create an angle view ratio of my camera, I used for inspiration the procedure on this site: http://www.paulillsley.com/airphoto/systems/balloons-kites.html which I'll try to explain here:
1) Place a large ruler or tape in front (perpendicular) to the camera, making sure that the ruler extends beyond the edge of the frame view, and measure the (exact) distance the camera is away from the ruler.
2) Take a picture of the ruler in the desired wide angle of the camera, and determine the length of the ruler that was recorded in the image.
3) Take the length of the ruler in the image and divide it by the distance the camera is away from the ruler and you will have the angle of view ratio for that camera at that wide angle setting. Mine was: Camera and perpendicular ruler at 50 cm distance, the image saw 61.13cm. My angle of view ratio is: 61.13 cm / 50 cm = 1.22260.
4) Once you have this value (1.22260) you can multiply the flying height by this number to determine the length of coverage on the ground (assuming the camera is pointing vertically downward).
5) To calculate the ground pixel resolution for a given height simply divide the length of the ground distance covered by the number of pixels covering the distance, for example, at 100m height, the ground distance covered would be 1.22260 x 100m = 122.26m now, 122.26m / 4000 pixels = 0.030565 m/p, or 3.05 centimeters per pixel.
From here you get the idea, so what I did was a spreadsheet in excel, where I just need to fill-in the desired scale, cm per pixel or height, as well as the desired endlap and sidelap percentages, and it will calculate everything for me.
Curious thing is that with this method I get a result that is slightly different from the Mission Planner (which I just successfully tried), and also a different result with the Photogrammetry utility in DJI PC Ground Station, which I have also tried.
My spreadsheet looks like this (image as attachment), it's in spanish, and the only fields you need to fill-in are the ones in white.
Manuel
fotogrametría.JPG
You could use the mission planner from apm to do this. When you plan a survey, you can input all your variables like camera, required percentage or meter overlap in the dialog box. Then it displays the lane width (to achieve a certain side overlap) and it stores a parameter "CAMERA_TRIGG_DIST" (something like that), which is the forward overlap on the apm after you save the mission. Could be that in the current version it also shows that on screen.
That should help you to verify your own calculations. I did a quick search on Google, but couldn't see a straightforward formula to use. Anyway, if you draw this out it should help a lot to figure out what the formula should be.
My focal lenght (35mm equivalent) is 25mm.
The camera is 12.1 Megapixel, 1/2.3-inch CMOS
4000x3000 pixels
-
1
-
2
of 2 Next