Mapping a ball field for drainage correction

I looked over a couple of mapping jobs today for a client who’s been commissioned to correct a badly built (poor drainage) softball field and an adjacent baseball field at a new high school. I’m concerned that the nice turf inside the ball park won’t provide enough uniquely identifiable points in the images to allow a good mapping result. Also the elevation changes of interest are about 24-inches of variation from high point to low.

I’m going to do a trial run tomorrow just to see if the photogrammetry process can work on this application. I’m planning to run a cross-hatch pattern, and am thinking about running two different altitudes, one for each direction. Maybe 250’ AGL in one direction and 200 feet in the other. Would like to run best practice conditions, and the size is not large, so time/batteries/image count is not limiting.

Would appreciate hearing whether anyone has done this successfully and what you would recommend as best-practice conditions to get the best shot at useful results. Flying a P4P at solar noon on a cloudy day with about 6mph wind. The deliverables don’t have to coordinate with anyone’s CAD drawings (yet), but I’ll be using 10 GCP’s and registering them myself with my own rover on State system corrections.

This mapping trial has already caught the attention of the school district procurement officer who is thinking about other applications if this works. So… I’d like to make it work! Suggestions appreciated. I’ll be running this flight tomorrow!

Thanks… R

2 Likes

I’m new at this, so just my $0.02, but since it’s a small site and time is not an issue I would run more patterns than you need. Maybe run one cross-hatch at a lower altitude, 100-150’, and another at the altitudes you suggest. Also try to include both nadir and oblique. You’ll end up with more photos than you need, but you’ll also have more options for producing a good model.

1 Like

Since it’s a small area, you could do a double cross hatch with separation of 20 degrees between the sets of cross hatches. It’s way over-kill but would ensure absolutely perfect lens calibration (which isn’t much of an issue as you have a good sensor).

2 Likes

Good ideas guys. Thanks… R

2 Likes

When you’ve had a chance to run everything, let us know what you decided to fly and how the processing went.

2 Likes

Will do.

Ran both a baseball field and a softball field yesterday (4/19). For each field I ran a 150-ft AGL cross-hatch grid and a 200-ft AGL cross-hatch rotated ~ 45 degrees using Map Pilot from MME for capture control (really like that app). Photo conditions were ideal with overcast skys (no shadows), and enough light to have shutter speed 1/200 for all exposures. Map Pilot slowed the flight speed to match available light, and all legs were 5-9 mph. Ran Manual Focus after setting focus initially, manual white balance, and auto exposure. Photos look terrific.

Put down several GCP’s at selected locations (obvious high/lo areas, concrete spots that won’t move, storm drains, etc. Used precision GPS rover to register the GCP’s. At suggestion of a member of another forum, i dropped about 40 paper plates with numbers on them (not registered with rover), to give the processing engine something to work off of when selecting common points. I think that will make a difference, but haven’t gotten far enough into the processing yet to know results. Have started smaller field (softball) processing and am at the point of entering GCP data.

Thanks all for the input. Will follow up with results, but probably not until 4/22.

Regards… Bob R.

3 Likes

Well,I missed the 4/22 progress update, but here’s a brief attempt. The smaller field (softball) that I described above was running for 19 hours in ODM when I cancelled the run because I needed a deliverable for the client and had no indication how much further it had to go. Restarted in Pix4D and had a finished product in 12 hours.

The results were excellent. The ball field was a lot less uniform from nadir than it appeared from ground level oblique. Got a great orthomosaic as you’d expect, but also had excellent contours with the GCPs. I needed the contours to describe the shape of the field with a max elevation change of 50 inches over a length of 325 feet. The elevations were in meters, so I got contour intervals of 0.1 meter. Worked out great with both files georectified and aligned . Overlayed the contours vector file on the otho-mosiac raster file in ArcGIS and exported both in a pdf containing both layers and delivered to client. Also made a AutoCAD file using trial version of new Pix4d Survey app, but don’t have a way to view that file.

With regard to ODM, I plan to review the settings and restart it this week on this same file to see how long it takes to finish and how the mosaic and contours compare. My PC is a dual quad-core Xeon with SSD and 60GB RAM, so it should handle the work load OK. Don’t understand why it was taking so long. Any suggestions are welcome.

Thanks … Bob R.

3 Likes

Fantastic. Thanks for the followup details.

1 Like