Wednesday, May 27, 2015

Notes on trail running

40 SI sticks: 1400 USD.
Measuring wheel:  110 USD.
Banners: 40 USD
Inverter: 50 USD
Deep cycle battery: 100 USD
Monitor for results: 100 USD
1000 Marking flags: 80 USD

Monday, May 25, 2015

Ski to sea

HBOC was represented in Skitosea as part of Jupiter Mining Corporation, here is the report.

Logistics

Skitosea is a complicated relay: road closures for racers, human transportation class (all equipment to race and from race has to be delivered by human power), safety boats, early waking up, very busy finish line. 

I took part in competitive class team Jupiter Mining Corporation for the first running leg: had to wake up at 4:20, then was driven to road bike start place, then drove to downhill runner start, then drove back to alpine run (which was a trail run) start for me.

Usually the first two legs are XC ski and then downhill ski, but for this year they had to replace these with an "alpine run" - a short 5k trail run with some 600 ft of elevation gain. This leg has the least impact on the race - shortest winning time and, most likely, smallest biggest loss. 

Race

Some 35 minutes before the start, I get in the toilet queue. Assuming a person takes 1 min, I can barely make it - some 200 folks in front of me, 7 toilets. And I do! 3 minutes before the gun I run out of there and pose at the starting line to the drones flying over heads. Some jokes with bystanders and then BANG. A bit unexpected, but the race is on. 

I had tactically chosen the position at the rightmost side of the start line: it worked well since the exit out of the parking lot is in straight and, I believe, shortest line. At the end of the first 300 meters team Jupiter Mining Corporation is 3rd overall. Not bad. 

Since then people kept me passing - I had started too fast, but what the hell, I'm still moving.

The rest was miserable: too much of breakfast so I was considering going behind a tree, but it never materialized. Legs are a bit tired and I have trouble breathing - want to take more air, but that breakfast is there.

On the longest climb a girl passes me - but you should hear how organized was her breezing - I let her go. No chance to stay around. Apparently she is some mountain running champion.

I get some time back at the long downhill: catch a guy or 2. Run like you don't care, let the feet find something under. Just before the spectators, another girl - all in pink - is catching up with me. I pass the spectators and a second later they all cheer for her. Great! This time I am up for a fight, but on the pavement I am not finding enough enthusiasm and she passes me with one more guy.

Just before finish I see a guy who clearly has some trouble - he's being walked out and then the finish and I get a chair from Ryota, our downhill runner and I'm done.

I wait a bit around to learn that the second girl was an Olympic XC skier. Then start hiking down the hill with the chair, occasionally offering it to the downhill runners for a small-but-honest payment, but they never take the offer.

Place: 19th, time 25:48.

Next year?

Jupiter Mining Corporation can shave an hour off of our result very easy, so it would be a satisfying experience to do it again - much of the time lost was due to freak accidents that don't happen often. We ended up being around 180th place out of some 350 or so even with quite a lot of bad luck. See results.

However, it is unlikely I will do it again: it is an expensive event, especially from alpine run point of view - 80 dollars means 16 dollars/km + complicated and tiring logistics if Gunta does it as well. On the other hand - I think I could be somewhat of use - maybe even more than this time - in the downhill run, but do I want to hurt my knees? I would not be useful in the bicycle or boat parts, especially not so in the skiing. I also doubt alpine run will be there next year. I will likely stick to a support role if Gunta kayaks in this race again.

Mt Baker area was very nice, I would love to return there - there was too much of fog, but apart from that weather was exceptional - a bit too cold for tshirt, but no wind.



Thursday, April 23, 2015

Notes on UAVs, LIDAR, etc.

Lidar measurement

  • Velodyne's: Lidar Puck
    • 8k price
  • Velodyne's: HDL-32E
    • includes Garmin 
  • Routescene http://www.routescene.com/products/product/routescene-lidarpod-technical-specifications/
    • contains Velodyne's 35k usd sensor inside. but also spits out reasonable data. 
    • if they used the puck it would be a good e2e solution.
    • not sure if this is correct since hdl-32E already has much of that.
  • Autonomoustuff
    • box with 
      • LIDAR sensor
      • Inertial Measurement Unit
      • Processing Unit
      • Battery
      • Post processing software
    • Product good, price too large.

UAVs

Fixed wing

Rotary

  • DJI, S1000+, http://store.dji.com/spreading-wings/s1000-plus
    • weighs around 4kg, liftoff weight 11kg. 
    • price - 2.5 - 4.5k USD.
  • http://www.xactsense.com/products/
    • this is a running system as it should be. 
    • could be expensive; the puck is 8000USD, comparable DJI is around 3-4k, then that all is put together - unclear what is producing, but it could be .las files that we need. so 2k more. Total - around 14k of price. Then add some margin.

Data

  • National lidar dataset
    • "Production of a conterminous US 1-meter dataset will begin in early 2015 and will increase in geographic coverage as new high-resolution data are acquired."
      • this should satisfy all O' interest.

Thursday, April 2, 2015

Google maps/bing maps

It is hard to improve from simple areal photos. So what do you do? Well, you try to make some 3d picture of various indecency:
google here and bing here.


Not sure if they are using the LIDAR available in Redmond or maybe it is some form of 3d extrapolation from 2d images, but the crippling of images in both cases is quite annoying.

Monday, March 23, 2015

What's that?

What's that? 

Well, this is automatic vegetation (includes housing) from LIDAR data. Streets/fields are easy to be seen.
Formula used is the following: divide fields in 5x5m grid. Count the points that are < 0.5 meters, < 4 meters and rest. Then look at ratio of points between 0.5 and 4 meters divided by all points < 4 meters.
If ratio is > 0.5, all green, if > 0.3 - very green, if > 0.1 - somewhat green. If there are high points (> 4 meters), then make it a forest.

Sunday, January 25, 2015

HBOC Winter Champs

Results are here:
Day 1
Day 2

Route gadgets from Jan:

and also here (use the interface you like better):

Thanks to everyone who attended and volunteers:
Jan, who designed the course for second day,
All control pickers, including but not limited to : John, Roman, Ing, Eric. 

Especially large thanks goes to Dave Bishop who let us use his campground!

Here are maps:


Some race feedbacks:

Wednesday, December 17, 2014

Trying to find cliffs from point grid

The ask is simple: you are given vertices on a regular grid and need to return cliffs: lines where steepness gets larger than some given constant.
I tried algorithm like this:

  • mark all cells for which the borders are steep, 
  • connect 2 steepest sides with cliff segment
  • join segments where possible
End result is best described by this picture:
The small white lines are my calculated cliffs. The background is OCAD 11 generated steepness where black is 45degree steepness. My lines are calculated in a smaller subregion than the whole picture. 

This algorithm does not work without some additional improvements. Something better is needed.