Scott Partridge was kind to post Canadian UAV info.

I missed that post. These rules seem to be good at first read. The ones I have would fall in the least restrictive.

Here is a link to the official document from Transport Canada with complete list of operating parameters, requirements, and exemptions for UAVs in the Great White North.

http://www.tc.gc.ca/media/documents/ca-opssvs/ac-600-004.pdf

FINE PRINT

Did anyone read the last bit?

DO NOT carry dangerous good or lasers.

Does that mean that LiDAR is out?

You must stay 5 miles away from any group of buildings bigger than a farmstead. That right there kills most of the uses surveyors would have.

(26) The pilot operating under this exemption shall only operate a UAV at least
five (5) nautical miles away from a built-up area.

Note: UAVs operated under these exemptions are not required to meet any technical
airworthiness standards meaning that there is no assurance of the airworthiness or
capabilities of the UAV system. This increases the risks to persons and property on the
ground, therefore, UAVs must not be operated near populated areas. Built-up areas are
considered areas with groups of buildings or dwellings including anything from small
hamlets to major cities. Anything larger than a farmstead should be considered a built
up area.

The key to this is that the user can still apply for a Special Flight Operations Certificate for work that does not meet the exemptions. For example,flights over built up areas. The onus is on the applicant to satisfy the agency that more stringent operating conditions will be met, particularly wrt safety, emergency planning, project management and pilot experience, adherence to applicable aviation regulations, including filing NOTAMS, communication with air traffic control as required, etc.

For companies serious about this work, the process is reasonable.

The Canadian rules are very similar to the UK CAA rules. Operating ceiling is 400 feet.

The proximity limit is "150 metres of any congested area of a city, town or settlement, or within 50 metres of any person, vessel, vehicle or structure not under the control of the Remote Pilot during take off or landing". This takes into consideration the significantly greater congestion of UK property.

Again, all things may be possible IF a detailed flying plan and risk assessment are agreed with the CAA and Air Traffic Control in advance.

UK requires not just a theoretical exam, but also a flying exam which includes emergency response. As with the Canadian rules the pilot licence is specific to a particular UAV; a further flight exam is needed for any new aircraft added.

For business purposes, I wonder if anyone has tried this option:

Somehow Install a DSLR camera in the side basket of an helicopter, set the camera auto-record and have the pilot fly the planned lines. Something like a Nikon D7100 can record for quite a while.

The goal is to get the camera in the air and get good overlap pictures. The other goal is keep the camera from being damaged. Within the confinement of the side basket, the gear is safe and FAA requirements are met I would say.

This method could by-pass paper work and investment costs? On many remote projects, helicopters are at site anyway, so the flying cost can be budgeted as operation cost and provided to the client(s).

Sounds OK in theory - it's always good when the client supplies the vehicle.

Two points to consider - firstly vibration which might be transmitted to the camera. Secondly, systems such as the UX5 log the location of each photo which makes the initial assembly of the images much easier, especially if they are fairly featureless.

I'm not sure if the downdraught from the copter will create misleading patterns in grassed areas, which might make image matching difficult. I've had this on power lines where the draught can generate conductor swing and create multiple LIDAR images.

Depending on what accuracies you are after from the mapping, a low flying height and closely spaced lines may be necessary, which will eat into the copter's flying hours. Typically for image matching software you are looking for overlap cover to give at least 5 images, preferably 9 or more on every point.

> Sounds OK in theory - it's always good when the client supplies the vehicle.
>
> Two points to consider - firstly vibration which might be transmitted to the camera. Secondly, systems such as the UX5 log the location of each photo which makes the initial assembly of the images much easier, especially if they are fairly featureless.
>
> I'm not sure if the downdraught from the copter will create misleading patterns in grassed areas, which might make image matching difficult. I've had this on power lines where the draught can generate conductor swing and create multiple LIDAR images.
>
> Depending on what accuracies you are after from the mapping, a low flying height and closely spaced lines may be necessary, which will eat into the copter's flying hours. Typically for image matching software you are looking for overlap cover to give at least 5 images, preferably 9 or more on every point.

Thank you for this information. Out of curiosity, why does the new technology require so much overlap? The traditionnal photogrammetry method is something like forward overlap 60% and side overlap 30%.

Compare camera costs : a photogrammetric camera costs several hundred times more than a conventional camera. Having said that then a good "consumer" camera can be calibrated to a relatively good accuracy.

UAV photography is flown much lower (normally - excluding specific tasks such as power lines and road condition survey) than conventional photography, so there is much less likelihood of well defined features to be used as ties appearing in suitable places on the photography. The camera platform is also much less stable, hence the yaw and roll may vary substantially from frame to frame, giving rise to a lack of coverage unless there is a high percentage of overlap (typically 80% overlap).

Many years ago (1968 if I'm honest) I had a series of overlaps to control, where the only obvious features were gorse bushes scattered all over the hillside. Spent two days counting and ticking off hundreds of bushes to identify specific ones. It worked though; very small residuals - very scratched legs!

Thank you for the information, good stuff.