Watch the video

Laura bought some Randy Hare boxes and we pulled BB out to see what he’s think of this method of training. BB is already a trained and certified cadaver dog. It was neat to see that Laura could introduce a distraction into the game almost immediately. We didn’t have any tennis balls on a string handy, so we used a rubber ball on a whip. The whip part was torn off fairly quickly. But that made the ball on a string rather perfect for this purpose.

Since Google Earth lets you import GPX files by drag-and-drop, you can also quickly scan your track visually to see your stats:

Start the clock when you harness the dog and stop when the dog finds the subject. Press the lap button between training problems. I find it’s easiest to save all tracks for the same day in the same file. But you could save your your tracks weekly or even monthly, if you so desired.

Zooey 40 acre 11 March 2008

look more like this:

Here is a very nice close-up on the image above:

To import a track into Google Earth, drag a GPX file onto the Google Earth map. I actually use TopoFusion to download tracks and manipulate them. Then I archive the stored tracks on disk. Later, it’s easy to open them with any mapping application.

I also use overlay data. Google Earth now has street map data available from “layers”. You’ll also see how easy it is to show current weather. Unfortunately, it’s not as easy to find historic weather. Weather Underground often has it. But you can’t show it on your map yet.

If you notice in the top image above, there is very little geographic detail available in the region shown. However, there are plenty of data overlay network links that you can add to display more detailed geographic data. What I use most frequently are maps generated from gpsvisualizer.com. To use these, download this KML network link file to your desktop and open it up in Google Earth.

Make sure you’ve zoomed in on the area you wish to look at in your map.

Then right click on the network link and “refresh”. Once you do that, you can expand the menu and see different sorts of imagery available for your region of interest. In the movie below, I download both USGC topographic maps and also USGC 1 meter imagery.

Once you’ve clicked on the imagery you want and then downloaded it from your browser, you can open it in Google Earth. Give the network links a bit of time to load. Once you do, you’ll be able to easily adjust transparency.

Without further ado, here’s the movie demonstration:

You can also download the corresponding imagery from the same webpage on GPSVisualizer.com that you downloaded the network links from. These you can save to disk and align if you expect not to have network access.

What we all want from Google is the ability to load file-based network links. Also, a much better way to cache map data. But… we’ll just have to wait. In the meantime, they continue to release more and more useful updates.

It’s not yet so easy to create tracks and upload them from Google Earth… but GPSVisualizer comes through yet again! Here is a great freehand drawing utility that let’s you save out GPX tracks that you can then upload on to your GPS (again, I use TopoFusion to handle most of my GPS data exchange). You can also open these up in Google Earth. Here’s a screenshot.

So now… you can give your subject a track to follow and put it on his GPS. Later, you can view your subject’s track and your dog’s track on a very detailed Google Earth map using your imagery of choice!

Notice there are only 4 buttons! As you can imagine, the big red button is for ‘record’. I picked it up and was able to record a simple video, edit and upload in a matter of minutes. The best thing about this camera is that there are no accessories. Even the software resides on the camera. When you plug it into your PC, it launches an application that lets you do simple edits and even walks you through uploading your video to YouTube or Grouper. For the video below, I used the Macintosh software on the Flip Video camera. Though you have to actually install it to your machine, the experience is only slightly more clunky than the very smooth PC experience. And for use in the field? No pieces to lose and two double AA batteries to make it work. I’m not yet sure how it will perform in low light conditions, but I have a black and white IR helmet cam on the way for that kind of use.

In this video, I demonstrate that Zooey and I still battle over control about where the ball should go after a retrieve. She insists that a proper retrieve places the ball at about five feet from the human. I presume because it’s a good soccer distance for a solid kick. Regardless, I threaten to stop playing when she doesn’t bring the ball all the way back, and she continues to insist on the five foot placement. Titan is not sure what he should be doing in all of this, but it’s interesting enough to be involved. He has a svelte summer buzz cut.

I extracted an image from the video but had to adjust lighting, contrast and saturation in order to see Titan’s face clearly. Not bad, but it’s a shame it needed so much adjustment. (But yea for iPhoto for making this so easy.) Conveniently, the Flip seems to save a good quality image frame every second. This means it’s pretty simple for the user to “snapshot” save an image from the video. However, if you want a frame that falls between two seconds, you’re out of luck. You need to go to the original AVI on the Flip disk rather than through the provided software. That’s actually what I did to produce the image below. Consider that an expert feature that they allow for, but don’t expose in software.

The YouTube video quality is pretty poor. That’s probably because the video is being transcoded twice. For a 1.31 minute video, flip video stored a 32 MB AVI. The quality is actually very good. To upload to YouTube, the flip transcoded the video to QuickTime (about 10 MB). It uses a Sorenson Video 3 codec at a data rate of over 800 kb per second. Not so bad, but there are better codecs. Then YouTube transcoded a second time to Flash Video (H.263 and down to 3.7 MB). And, again, they don’t use the best codec or setting possible. So the video you see above — after being processed twice — is very poor.

Here is the same video uploaded directly to Blip.tv. Because I uploaded the original AVI without transcoding and then blip transcoded using On VP6 and dropped the frame rate to 15. The quality is quite good. Plus, Blip gives you access to the original (or add additional formats), if you prefer. The file size of the embedded FLV below is about 6 MB:

Click To Play

And here is a link to the original AVI!

The good thing, from my point-of-view, is that you don’t have to go through this transcoding process if you don’t want. I can grab the original AVI off the camera and transcode and post it the way I like. But that would marr the very “easy” nature of the Flip, wouldn’t it?

Looking at the metadata in the avi, I can see that the creation date of the video is the same day as I recorded it but approximately 8 hours off. Since there is no way to set the camera clock, I suppose this makes some sense. However, I’m not sure what would happen if I let the batteries (which came pre-installed) die and then replaced them.

Finally, a most interesting and exploitable idea for this camera is that one could easily replace the software on the camera with custom software. That means, someone who wanted to hand these cameras out to a particular group of people could provide a means for automatically processing and uploading the video to some custom website. Of course, this potential may exist with other cameras that act as disk drives, but this is the first that really demonstrates the enormous potential for seamless capture, processing and upload at such an accessible price.

What’s the point? Well, the video of the dog work is great for learning to read your dog and also reviewing a training session. I had imagined that having the track layer capture video might be interesting if you wanted to carefully review a known track before running it. But until I’ve got time synchronization working right between GPS track and video on a map, it’s not really worth the effort. What you really want to see are critical intersections and landmarks. And this is much easier to do if reviewing the track on a map. Hopefully, later this summer we’ll be able to navigate the video on the map and see how well it works for this purpose!

Let’s start with the Kestrel 4500. Why is this such an exciting product? For three reasons:

• Because the Kestrel gives wind direction along with wind speed — there is a built-in digital compass
• It also calculates crosswind, tailwind and headwind with reference to a user-set target heading.
• And, very importantly, it is a data logging device that stores time-stamped measurements!

Imagine this as an expensive puff bottle for checking wind direction and windspeed working air scent searches or even trailing problems. But it not only gives you the information you need while working a strategy, but also stores that information so that it can be later plotted on the map. Or, at least, that’s what I intend to do with it. Stay tuned for a demo this summer.

Next up is the new Samsung SC-X210WL. I’ve used the two previous Samsung helmet cameras in the past. The first was horrible, and I probably blogged on its performance somewhere. The second was quite good. The two problems that I had with it were: 1) battery life and 2) wire between the cam and the camera body. Samsung has eliminated the wire and so this camera should be even easier to operate and use in the field. The battery life should not be appreciably different (e.g., orders of magnitude better). So I carry lots of spares into the field. Or you can carry a battery power pack that can charge the camera for some hours.

Next is perhaps the most exciting product. The Garmin Astro comes out this month. It is a GPS-enabled dog tracking system designed for hunters. The handler carries a Garmin handset (includes the high-sensitive SiRF Star III chip) while the dog wears a collar containing a GPS and transmitter. The dog’s location is transmitted every 5 seconds to the handler and both tracks are saved on the handset. You can mark the location where the dog makes the find (making distance calculations a breeze). You can also track multiple dogs… so working joint operations with multiple dogs is a good possibility.

Finally, happily, Garmin is releasing the new Rino 530HCx. The Rino, of which I make much use, finally has a new high sensitive chipset and a also a microSD card. It’s not clear if it has a 3-axis compass that the Astro boasts on having.

Happily, Garmin has also set up a developer website. They’ve made it easy for developers to interact with Garmin devices and data via web services and plugins. I’m sure I’ll be testing these capabilities and writing about them soon!

One thing I noticed switching back and forth between hybrid aerial map view and birdseye view is that there is some distortion of the track in birdseye view. That is, along some segments, the track is displaced by up to 10 meters or so. You can test this for yourself. I presume this is caused by the technique used to create the effect. Most likely, what we are seeing is the track moving from a planar surface to one that crosses planes. Microsoft should be able to correct for how they handle polylines when users switch from one map view to another… if they so desired. The aerial view is lots more accurate, though you can’t see nearly the same detail as in the birdseye view.

I’d still like to do some fancy things like include speed information as in the image below. If you notice, you can see the latter part of the track has two lines. This is because the tracklayer did not turn the GPS on until most of the track had been set. The ability to compare what the dog did with what the tracklayer did is very useful.

Although there are all sorts of great mapping applications online such as Google Maps/Earth, Microsoft Virtual Earth, Yahoo Maps, and ESRI Arcweb Explorer, it takes a bit of searching around to find a service with updated and high resolution images for any particular area. Fortunately, most of them have nice programming interfaces that let us easily create custom maps and applications. It would be easy to create an application mashup that would let you switch from map service to map service to find the best imagery… but no one seems to have done this such that it’s also easy to upload your GPS tracks and do interesting things with them.