Masthead GoPro project

[RaymondR]

That's a pretty good image with enough resolution to be useful. Good work.

I'd be inclined to retain the metal lid but just make it a full circle, what's going on on the water, and your boat is the thing of major interest unless you're paranoid about spy planes or drones and the shade will prevent the sun from cooking your camera.

Re-orienting the camera and using the shaded part of the image to get a more panoramic view would be good.

[markxengineerin]

Some updates to improve what was very bad reliability 1) I put the battery back in. It was previously removed due to overheating, but with the cooling fans now present it has not been an issue again. 2) I bought a GoPro SmartRemote, because the iphone/ipad were spotty in having the range to reliably connect for turning the camera on/off. The gimbal used to turn the camera on and off, but since "upgrading" the camera to the gopro labs firmware, that seems to have stopped working. It was an unnecessary upgrade, so I might switch back at some point. I am only on day 1 with the remote, so too soon to report if it's a good solution or not, but it claims up to 600 feet range. Here is some video, will try the setup out while sailing soon:
https://youtu.be/-w96kWwfO3I

I also made a new roof from lexan, and painted only the center white. It allows for a slightly better view while still giving some sun protection.

My next plans are to try different setting to see what looks best. Youtube probably degrades the videos, but they display at 2.7k60 on the tv and record to pc at 1080p (limit of the HDMI capture card I have) . I don't like the fishbowl effect much, and could get rid of it with "linear" mode instead of wide, but I think i would be sacrificing some field of view. It's not too bad when the camera is level, which it mostly will be, but gets bad when aimed up or down.

[Pizzazz]

I had some spare time to look into good boat cameras, especially comparing to high performing commercial options, such as the Flir M300C that retails for $6995. Ouch! Here is what I found out and what I am currently building. So, I set out to build a better camera for less, relaxing some of the constraints of the commercial version, improving on others.

The FLIR M300C uses a Sony Exmor R 1/2.8 sensor, most likely the IMX290 or equivalent with a sensitivity value of 5486 (don't worry about the units for now). It uses a 30X zoom that goes from ~4mm to ~120mm, F-stop from 1.8 to 3.9 and field of view from 60 degrees to 2 degrees on the tele-end. It sounds impressive and I am sure that the glass is excellent but it is only 2-axis stabilized (nearly unusable at sea) and the lens is quite closed (F3.9) at the tele-end.

So, I found this mini-wifi camera below that uses the Sony IMX307 sensor that is newer and (some say) better as it has low light sensitivity of 7747 (nearly 40% better than the Flir). The camera costs $100 and it is tiny, approx. 1" by 1".

https://www.longse.com/Products/Cons...0623/2401.html

I put this camera on a Zhiyun Q2 gimbal ($100 for 3-axis stabilization, very smooth). It works very well but the camera is too light for the gimbal, so I had to use a counterweight. My next idea is to get a second camera and mount them side-by-side to balance. The lenses are replaced with longer, more powerful versions. One is currently a 12mm lens (20 deg angle), the other one will be a 50 mm lens (5 deg angle). The good thing is that the 50mm lens for example has F1.6, which means it let's in approx. 5x more light than the Flir (minus the cheap glass quality). And it only costs $40.

The current setup is, one phone gimbal controlled by a phone at the nav station, two cameras with different fixed zooms on the gimbal, a tiny 5V-12V converter to power the wifi cameras from the gimbal power supply. All of this mounted on the arch pointing down. Cost is around $400 including all the parts.

So, how does it compare to the Flir? In my calculations, on the long-tele end I am getting 5x more powerful lens and 40% better sensor or 7x expected better performance. Furthermore, it is much better stabilized. The Flir probably has higher quality optics and better software, although all of this is questionable. The Flir has longer zoom but in all of the Raymarine videos the demonstration boat is in absolutely flat water, at night. I can't imagine a 2 degree field of view being useful in waves with a 2-axis platform. Also, I kind of like having two cameras on two side-by-side displays. The Flir is water resistant, in my case the cameras can be sealed nicely but the gimbal is not. In any case, should this work well, there are more expensive weather-proof gimbals. All of this is powered by a battery or 5V usb cable attached to the gimbal.

Once I get all the parts I will provide some videos but based on this, what am I missing? Why did Flir go for 2-axis only stabilization? Further, technology progresses quickly. The Exmor -290 sensor was good five years ago. Today the 307 (40% better) is widely available, and the next version (327) adds another 40% sensitivity gain. Some telescope guys are already using the IMX662 sensor that is 3.5x better than the Flir. Hopefully, it will become available in more cameras that we can purchase inexpensively.

If you guys come across some cool camera (have to be light to mount on the gimbal, around 300 grams), feel free to share in the thread.

[markxengineerin]

Pizzazz, have you tried yours on a mast yet? The numbers all sound good but I want to see how it looks!

Here is an update on my project

https://www.youtube.com/watch?v=trrHXlNa59s

Notes
-It had not rained in about a month, the enclosure is dirty but usually rain cleans it very well
-Stabilization is excellent, no problem at all there w/ the Hohem gimbal
-The audio source I captured was accidentally the laptop mic, not the camera, ignore
-The view is much better when not pointing into the sun. I would like to implement some kind of polarizer to get rid of all the reflections off the water but don't think it's possible when looking through the plexiglass enclosure.
-I can see freighters at about 3 miles, from my memory, need to double check that number. It was really no better than the view from the human eye on deck because of the lack of zoom. But I do find it useful for spending more time in the cabin, which was the goal- more time lying on the settee (when I want it) and less pop outs to look around.
-At night (not shown) it is pretty useless. A FLIR camera attempt would be interesting. I was hoping this one would pick up a distant ship's lighting on a dark horizon, but it does not.
-I wonder if mounting to the bowsprit would be more useful and much easier to implement, this seems too high up, at least without zoom. I don't know what zoom would do to the stability.