Double Sony ILX-LR1

[Darko]

Hi,
I am planning to use two Sony ILX-LR1 cameras equipped with Zeiss Loxia 85mm lens in simultenious triggering mode. The cameras will be rigidly mounted one in front of the other and leaned from the vertical axis for 10 deg left and right. The planned overlap between cameras will be about 10%. Is it an acceptable way of widening the field of view? In case I can use it in such a way my field of view would be 44 deg instead of 24 that has a single camera, is this OK?
Best regards,
Darko Car

[JMR]

Well I don't see major cons, but why 85mm and not a wider lens instead? Is there a need for ultra hight ground resolution? I guess it's going to be a manned aircraft borne solution. Sounds interesting.
Regards

[olihar]

It is better to create parallax instead of creating a panoramic rig.

[Darko]

Thanks, JMR,
I appreciate your opinion. The 85mm has been used for flights with manned aircraft (ultralight) and a ground resolution of 5cm. The 85 has been used to avoid large changes in ground resolution since the terrain is very uneven and it is hard to follow it precisely. Two cameras will be used just to get a wider swath with the single flight line. With two cameras I can cover about 850m instead of 475m and the distance between my flight lines changes from 150 to 250m (70% side overlap) which is significant.
What do you think about that?

[Darko]

And, my flight is 1100m AGL.. Which is much safer than 280m that we used to fly with Loxia 21mm. Results were fine but the terrain was much more even and easy to follow.

[JMR]

I understand the convenience of flying higher, but keep in mind that long lenses come with some inconveniences. More atmosphere in between the camera and the ground, and worse sensitivity to depth. So if the gsd is the same (long lens at higher altitude equals wide lens in lower), generaly one can expect a more robust bundle adjustment and better 3D accuracy with a moderatley wide lens.
Additionally, have you considered a variable angle gimbal and one single oscillating camera? the configuration of two synced cameras overlapping 10% only to widen the field of view will hardly be beneficial in any other respect, so I wonder if it is not just more expensive than a movable single camera. Have you seen the so called "smart oblique" mode allowed by DJI Zenmuse P1?
Regards

[Darko]

The complete setup is hung on the two-axis gimbal. Have no better idea of how to handle a steep climb of 30° when the plane can climb just 10°. The area that we must fly over is huge. The planned flights with a side overlap of 70% is about 4000 km. No way to use something like a small drone. We have M300 and M350 equipped with the P1 cameras and one M3E. The plan is to cover the area with the manned aircraft and later just fill in the populated places with the drones. The job is in the north Adriatic, Island of Cres, Lošinj, and a few smaller ones. I've done smaller Islands like Susak, Ilovik, etc. with the M3E with fine results but Cres is too big with no roads or places to take off with the drone and regulations are the problem as well.
Some other company have already made that aerial shooting for us on some other bigger islands and they also used a professional strictly rigid double camera flown on 1000+m AGL. The results were good. The price of one flight covers the price of all our equipment and maybe the used small airplane. For this job, we have a rented one with the pilot and I have invested in their Auto Pilot that's been installed and tested.

We are planning to make a smaller single-flight test and if that comes out fine, good. If not the option is a single camera with 50mm Zeiss Loxia lens and flight on 700m AGL.
Best regards

[Jimbob]

I have used a twin camera setup many times. It works fine if you are using long lenses (like the 85mm on a 35mm sensor). As another commenter mentioned, twin cameras minic a panoramic style, which is not as strong for DEM creation. Make sure you have good overlap (say 70-80% forward overlap) and you will be fine. One suggestion...you might want to consider a larger sensor (like Fuji gfx) instead of twin sony's. The sensor is 1.7 times larger, thus eliminating the need for twin sonys. Hope your project goes well.

[JOSEALEXGOMES]

I'm tying a sony lr1 with 85mm lens in a manned aircraft too, but my 1st test didn't work fine, the agisoft alignment solution was terrible.
Changing to a 50mm lens returns better results.
Can you confirm that the 85mm works for you?
Can you tell me what kind of two-axis gimbal are you using?

[jrp]

I'm tying a sony lr1 with 85mm lens in a manned aircraft too, but my 1st test didn't work fine, the agisoft alignment solution was terrible.
...

That's a worrying result, would be interested to know more.

[Darko]

Hello guys, thak you for your previous comments, it helped me a lot.

I did it. The project is successfully finished, and the results are perfect. I have used two Sony ILX-LR1 cameras positioned in front of each other, hinged on a single two-axis gimbal.  Each camera has been leaned 15 degrees left and right from the nadir point. I have used 50mm Zeiss Loxia lenses with the aperture set to 8, and the exposure time has been set to 1/1250. It's all been set to manual, but the ISO setting was Auto. The lens has been taped so the rings could not be rotated. It all returned fine and sharp pictures that were good for the Agi alignment procedure. Since the exposure was taken by the external timer for both cameras simultaneously, only one event was recorded for the camera pair. Some problems did appear since cameras didn't have the same number of photos (a few only) in each flight for an unknown reason. That's been solved by time comparison taken from the EXIF and events.
I have flown 700m AGL, and the flight speed was about 35-45m/s, the exposure timer was set to 1.6 sec, which gave a good overlap of 70-80%. Ground resolution was about 5cm, and the adjustment gave me errors in mm. The 3D model in TLS came out just perfect.