automatic F3B trigger system - Page 20

jjmouris · Nov 07, 2009, 05:42 AM

Now your talking!

Is this the cheap fisheye type lens grafted onto an unsuspecting webcam?!

Joe, who still likes the double webcam stack idea.

moreil · Nov 07, 2009, 07:21 AM

Quote:

Originally Posted by Mutchy

I tested the FxxTrack Ver 1.06 today with my el cheapo wide angle lens at Picton with excellent results. Chimes are satisfactorily audible. One missed chime in 80 laps.

Wow! Very, very good to hear.
Any useful video out of it?

Now I deeply regret my procrastination in re-writing the video saving

Mutchy · Nov 07, 2009, 06:48 PM

Picton weather was overcast with one very light shower of rain. The clouds were low and mottled white and grey (see later comment). The wind was light and variable with some mild thermal turbulence.
I had the usual set up of the MacBook running Windows XP sitting on a chair with the camera mounted on a tripod next to us at Base A. The wide angle (x=.67 ) was stuck on the front of the Logitec Vision Pro. Pictures below. The field of view is slightly less than 90 degrees probably about 85 degrees but there is some spherical aberration and the corners of the software window are shrouded by the lens body. Also the bottom of the filed of view is circular and the horizon is distorted somewhat due to the extra lens resulting in a short horizon and odd appearance of objects on the horizon.

A set of 10 climbs and laps with my "Kompressor" (3kW F5B model) yielded good results with the camera being able to pick up and signal the low lap 4 return leg at 2 metres or so above the ground and the subsequent entry to the course at about 120 metres altitude. The model has yellow wings and red fuselage. Chimes were quite loud enough for me even though I am 66 years of age and have some industrial deafness.

I left the whole setup running for about an hour and a half. During this time there were maybe a dozen times when spurious chimes sounded. None of the four of us had any idea what caused the chimes but it could have been insects or birds. Cars travelling along the road opposite (which contrary to my previous post is actually about 250 metres away) were observed by the camera but did not generate chimes.

A set of ten climbs with 4 laps each with my Karia (2kW LP class F5B model) which is a bit slower than before yielded similar results with one missed chime at the exit of the 4th lap (may have been a bit low

or the distortion in the extreme bottom of the lens) and two or three missed entries to the course. The model is white with one red wingtip and one yellow tip and against the low clouds was sometimes difficult to see on the climb and turn into the course due to the silhouette of the model being grey lost in the grey bottoms of the clouds. For Base B use this could be important at the end of the first leg and will need to be observed when I next use the system which will be at Base B. (don't fly on cloudy days eh?)

In all the excitement I remembered that I forgot to start the video recordings

Richard was flying a small white trainer with a student so I took a video of that flight and will upload it via ftp so you can have a look at the software performance. The trainer disappeared into the background when at high altitude and you can see that happening a lot. When Richard found out I was capturing video he did some aerobatics down lower where the tracking was very reliable. I will upload the video later this morning.

Steve

PS Lens from here: http://usbfever.com/index_eproduct_v...roducts_id=789

jjmouris · Nov 08, 2009, 04:03 AM

And what is the field of view of a Microsoft Cinema on it's side?

About to pull the trigger on buying one.

Joe

moreil · Nov 08, 2009, 06:43 AM

Lovely writeup, Mutchy.

The spurious chimes are almost always insects these days. I'm amazed at how well it picks them up. I had a camera pointed out of the window the other day, it was constantly finding little flying insects that I couldn't see with the naked eye.

Now if I could just think of a way to reliably distinguish between a plane that's 100m away, and an insect that's 50cm away

A progress report of sorts: I've spent some of the past week amending the detection to be a little more sensitive with some degree of success. I've also spent the last few days re-writing the video file saving. In a fit of frustration I ended up writing my own video compression codec, because I just couldn't find anything that was fast enough and flexible enough. And it turned out to have (much) better compression. Apparently it's possible to get lucky some times

It's not quite ready for release; probably ~ tuesday it'll be up to some testing.

jjmouris · Nov 09, 2009, 05:50 AM

Seriously, how long would it take to put in a 90degree angle change function or perhaps just a horizontal line on the screen vs a vertical line?! It seems like the most sensible option to inscrease the field of view. Well, how about it?

I just ran my 44 leg flight video through the latest 1.6 version and it seemed pritty good to me. Yes there are a fair few insect related buzzes but more importantly the buzzer goes off for the plane every time it is within the view of the camera. The tracking also seems to start a little more early then just at passing the base like it did before. So obviously the plane is getting marked as a plane a little more early on.

I also note that the real action is only going on in the mid 50% of the screen wipth. The outer 25% parts left and right do not really seem needed. Ashame we can just select a different recording format on the camera like widescreen and then tilt the camera.

Personally i want to do plenty of mid winter flying getting my planes dialed in and keeping my practice up. Allowing for cold finger syndrome ofcourse. If we can get this auto base B thing going on my old laptop then that would be totally ace. Worth the purchase of a webcam.

Now i notice still only the Logitech camera is getting used right?! Nobody having any luck with the Microsoft Cinema one?!

moreil · Nov 09, 2009, 03:49 PM

jjmouris, the Microsoft Cinema seems to work just fine.

Note that while the displayed trace is mostly in the mid 50% of the screen, the other parts aren't useless

The program is actually tracking right from the edge of the screen (well, from 15 pixels in) but doesn't display the trace until it builds enough confidence that it's actually a plane and not just noise.

Having said that, I'll add '90 degree rotation' up the to-do list a bit. It should_be fairly easy. At the very least I'd add a 'rotation baseline 90 degrees' which is trivial.

jjmouris · Nov 10, 2009, 11:35 AM

Thank you!

The Cinema webcam is on it's way. Time will tell if my old laptop will handle it.

Note that the Japanese have found your website also;

http://blog.airplanemaniac.com/?page=1
http://blog.airplanemaniac.com/images/auto-base-B.jpg

Anyone care to translate what they are saying?

Darron · Nov 10, 2009, 02:13 PM

Quote:

Originally Posted by jjmouris

I just ran my 44 leg flight video through the latest 1.6 version and it seemed pritty good to me.

How do you get it to play your video?

jjmouris · Nov 10, 2009, 03:16 PM

I just open the program, click away the no camera attached warning, file > open then "video position" and run. I have had everything work in different folders and also closed the program using the (X) and also Exit. Seems fine to me.

The thing is, you have to be carefull that the file is COMPLETELY downloaded. Check the file size. Otherwise you will only see the first frame and the video does not run at all.

Note my old laptop goes to 100% CPU usage running the video and the frame rate is only just short of 10fps when playing the video! This is why i am worried the whole thing might not work on my old laptop in real time. Even without recording the session.

Joe

Mutchy · Nov 10, 2009, 09:58 PM

Testing completed at Base B today using FxxTrack version 1.06.

Sunny day warmish at 29 deg C high strato cirrus (thin ice crystal type) cloud at 2/10ths coverage. Sun from the left quarter of the course.

Computer/tripod/webcam set at Base B. iPhone wide angle lens installed. Walkie talkie with jumper across PTT switch for communication.

First flight with Karia was very successful with only one missed turn signal going away and two coming back. (I've been flying the course way too long!)
At the second missed return chime I realised I had (fortunately -see later) not ticked the video record box. Thumbs up - big smile!

Checked computer and found it had frozen with the "Windows screen saver showing. Restarted and ticked video record box.

2nd Flight with Kompressor, a faster and more colourful model. Only one or two chimes heard during the whole flight. Checked the computer and it had stopped recording because it had exhausted the Hard Disk capacity. A file of about 600Mb had been recorded. I had to discard some old and temporary files to get Windows to reboot and hence recover the FxxTrack video for ftp transmission. Hopefully the video may show you what was happening.

Steve

Alan Flockhart · Nov 11, 2009, 04:11 AM

Quote:

Originally Posted by Mutchy

First flight with Karia was very successful with only one missed turn signal going away and two coming back. (I've been flying the course way too long!)
At the second missed return chime I realised I had (fortunately -see later) not ticked the video record box. Thumbs up - big smile!

With the missed turn signals are you sure you are flying past the center line of the system? In the UK we had a couple times where people turned right on the line and the system at that point should only buzz once.

You could have also been too high/too low. We have seen that all season in the UK. First leg high going past the line is missed as out of camera field of vision. Diving turn carried out means plane is visible by FXX on the rtn and you get a buzz.
Same for the bottom leg. Going out of course you get a buzz then on rtn you are below the camera angle and FXX can't see you.

I don't know if it is possible but having someone watching the screen whilst you are flying can catch the high/low items.

Quote:

Originally Posted by Mutchy

Checked computer and found it had frozen with the "Windows screen saver showing. Restarted and ticked video record box.

2nd Flight with Kompressor, a faster and more colourful model. Only one or two chimes heard during the whole flight. Checked the computer and it had stopped recording because it had exhausted the Hard Disk capacity. A file of about 600Mb had been recorded. I had to discard some old and temporary files to get Windows to reboot and hence recover the FxxTrack video for ftp transmission. Hopefully the video may show you what was happening.

Steve

I'd think that if the computer was running out of disk space then the OS wouldn't be giving FXX much chance to do stuff. The OS would have been "worrying" about what to do with its own data never mind trying to write the GB's which FXX would have been throwing at it.

jjmouris · Nov 11, 2009, 04:26 AM

What he said.

I also wonder if in general recording the video interferes with the operation of the program. Seems like any processing capabilaty and also RAM memory taken away from the program will result in hickups?!

Alan Flockhart · Nov 11, 2009, 04:30 AM

Very possibly. My lappie when running FXX and saving to disk doesn't put much above 40% use on the CPU. memory is within the bounds on the machine as well (Though I do shutdown everything like IM clients, phone sync etc etc to be sure it isn't trying to do something in the background)

Mutchy · Nov 11, 2009, 04:44 AM

You could have also been too high/too low. We have seen that all season in the UK. First leg high going past the line is missed as out of camera field of vision. Diving turn carried out means plane is visible by FXX on the rtn and you get a buzz.
Same for the bottom leg. Going out of course you get a buzz then on rtn you are below the camera angle and FXX can't see you.

Alan,
As far as I am aware I flew exactly the same with the second model as the first. The two flights were within 20 minutes of each other. I believe it is unlikely that during the first flight the system could chime 37 times out of 40 and only 3 out of 40 the second flight with my flying being too high/low being the cause.

I'd think that if the computer was running out of disk space then the OS wouldn't be giving FXX much chance to do stuff. The OS would have been "worrying" about what to do with its own data never mind trying to write the GB's which FXX would have been throwing at it.[/quote]

Yes. The OS would be trying to find space to store all that uncompressed video data.
I have cleaned up the partition and found enough space to store a couple of videos so tests tomorrow morning will tell me if my CPU can walk and chew gum simultaneously.

Steve

Nov 07, 2009, 06:48 PM	#288
Mutchy Registered User Join Date: Dec 2003 Location: Wollongong Australia Posts: 76	Picton weather was overcast with one very light shower of rain. The clouds were low and mottled white and grey (see later comment). The wind was light and variable with some mild thermal turbulence. I had the usual set up of the MacBook running Windows XP sitting on a chair with the camera mounted on a tripod next to us at Base A. The wide angle (x=.67 ) was stuck on the front of the Logitec Vision Pro. Pictures below. The field of view is slightly less than 90 degrees probably about 85 degrees but there is some spherical aberration and the corners of the software window are shrouded by the lens body. Also the bottom of the filed of view is circular and the horizon is distorted somewhat due to the extra lens resulting in a short horizon and odd appearance of objects on the horizon. A set of 10 climbs and laps with my "Kompressor" (3kW F5B model) yielded good results with the camera being able to pick up and signal the low lap 4 return leg at 2 metres or so above the ground and the subsequent entry to the course at about 120 metres altitude. The model has yellow wings and red fuselage. Chimes were quite loud enough for me even though I am 66 years of age and have some industrial deafness. I left the whole setup running for about an hour and a half. During this time there were maybe a dozen times when spurious chimes sounded. None of the four of us had any idea what caused the chimes but it could have been insects or birds. Cars travelling along the road opposite (which contrary to my previous post is actually about 250 metres away) were observed by the camera but did not generate chimes. A set of ten climbs with 4 laps each with my Karia (2kW LP class F5B model) which is a bit slower than before yielded similar results with one missed chime at the exit of the 4th lap (may have been a bit low or the distortion in the extreme bottom of the lens) and two or three missed entries to the course. The model is white with one red wingtip and one yellow tip and against the low clouds was sometimes difficult to see on the climb and turn into the course due to the silhouette of the model being grey lost in the grey bottoms of the clouds. For Base B use this could be important at the end of the first leg and will need to be observed when I next use the system which will be at Base B. (don't fly on cloudy days eh?) In all the excitement I remembered that I forgot to start the video recordings Richard was flying a small white trainer with a student so I took a video of that flight and will upload it via ftp so you can have a look at the software performance. The trainer disappeared into the background when at high altitude and you can see that happening a lot. When Richard found out I was capturing video he did some aerobatics down lower where the tracking was very reliable. I will upload the video later this morning. Steve PS Lens from here: http://usbfever.com/index_eproduct_v...roducts_id=789 Attached Thumbnails All the pieces of the wide angle lens. The small rings are strongly magnetised and have double sided tape on one surface. 41.5 KB · Views: 31 The lens fitted to the webcam. 43.4 KB · Views: 34 The view with lens fitted. The webcam is about 1.5 metres from wall. 58.7 KB · Views: 46 The view without the wide angle lens 54.4 KB · Views: 33 Lens assembly with lens cap and magnetic base and carrying lanyard. 41.3 KB · Views: 28 Last edited by Mutchy; Nov 08, 2009 at 04:43 AM. Reason: URL added

Nov 10, 2009, 11:35 AM	#293
jjmouris High Performance Join Date: Aug 2008 Location: Southampton, UK Posts: 2,519	Thank you! The Cinema webcam is on it's way. Time will tell if my old laptop will handle it. Note that the Japanese have found your website also; http://blog.airplanemaniac.com/?page=1 http://blog.airplanemaniac.com/images/auto-base-B.jpg Anyone care to translate what they are saying? Last edited by jjmouris; Nov 10, 2009 at 01:21 PM.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Discussion Does anyone know where I could get an Electronic Automatic Boiler Management System	Deserteagle	Scale Boats	1	Jul 21, 2008 06:45 PM

Nov 07, 2009, 05:42 AM	#286
jjmouris High Performance Join Date: Aug 2008 Location: Southampton, UK Posts: 2,519	Now your talking! Is this the cheap fisheye type lens grafted onto an unsuspecting webcam?! Joe, who still likes the double webcam stack idea.

Nov 08, 2009, 04:03 AM	#289
jjmouris High Performance Join Date: Aug 2008 Location: Southampton, UK Posts: 2,519	And what is the field of view of a Microsoft Cinema on it's side? About to pull the trigger on buying one. Joe

Nov 08, 2009, 06:43 AM	#290
moreil Registered User Join Date: Jun 2005 Location: Sydney, Australia Posts: 293	Lovely writeup, Mutchy. The spurious chimes are almost always insects these days. I'm amazed at how well it picks them up. I had a camera pointed out of the window the other day, it was constantly finding little flying insects that I couldn't see with the naked eye. Now if I could just think of a way to reliably distinguish between a plane that's 100m away, and an insect that's 50cm away A progress report of sorts: I've spent some of the past week amending the detection to be a little more sensitive with some degree of success. I've also spent the last few days re-writing the video file saving. In a fit of frustration I ended up writing my own video compression codec, because I just couldn't find anything that was fast enough and flexible enough. And it turned out to have (much) better compression. Apparently it's possible to get lucky some times It's not quite ready for release; probably ~ tuesday it'll be up to some testing.

Nov 09, 2009, 05:50 AM	#291
jjmouris High Performance Join Date: Aug 2008 Location: Southampton, UK Posts: 2,519	Seriously, how long would it take to put in a 90degree angle change function or perhaps just a horizontal line on the screen vs a vertical line?! It seems like the most sensible option to inscrease the field of view. Well, how about it? I just ran my 44 leg flight video through the latest 1.6 version and it seemed pritty good to me. Yes there are a fair few insect related buzzes but more importantly the buzzer goes off for the plane every time it is within the view of the camera. The tracking also seems to start a little more early then just at passing the base like it did before. So obviously the plane is getting marked as a plane a little more early on. I also note that the real action is only going on in the mid 50% of the screen wipth. The outer 25% parts left and right do not really seem needed. Ashame we can just select a different recording format on the camera like widescreen and then tilt the camera. Personally i want to do plenty of mid winter flying getting my planes dialed in and keeping my practice up. Allowing for cold finger syndrome ofcourse. If we can get this auto base B thing going on my old laptop then that would be totally ace. Worth the purchase of a webcam. Now i notice still only the Logitech camera is getting used right?! Nobody having any luck with the Microsoft Cinema one?!

Nov 09, 2009, 03:49 PM	#292
moreil Registered User Join Date: Jun 2005 Location: Sydney, Australia Posts: 293	jjmouris, the Microsoft Cinema seems to work just fine. Note that while the displayed trace is mostly in the mid 50% of the screen, the other parts aren't useless The program is actually tracking right from the edge of the screen (well, from 15 pixels in) but doesn't display the trace until it builds enough confidence that it's actually a plane and not just noise. Having said that, I'll add '90 degree rotation' up the to-do list a bit. It should_be fairly easy. At the very least I'd add a 'rotation baseline 90 degrees' which is trivial.

Nov 10, 2009, 03:16 PM	#295
jjmouris High Performance Join Date: Aug 2008 Location: Southampton, UK Posts: 2,519	I just open the program, click away the no camera attached warning, file > open then "video position" and run. I have had everything work in different folders and also closed the program using the (X) and also Exit. Seems fine to me. The thing is, you have to be carefull that the file is COMPLETELY downloaded. Check the file size. Otherwise you will only see the first frame and the video does not run at all. Note my old laptop goes to 100% CPU usage running the video and the frame rate is only just short of 10fps when playing the video! This is why i am worried the whole thing might not work on my old laptop in real time. Even without recording the session. Joe

Nov 10, 2009, 09:58 PM	#296
Mutchy Registered User Join Date: Dec 2003 Location: Wollongong Australia Posts: 76	Testing completed at Base B today using FxxTrack version 1.06. Sunny day warmish at 29 deg C high strato cirrus (thin ice crystal type) cloud at 2/10ths coverage. Sun from the left quarter of the course. Computer/tripod/webcam set at Base B. iPhone wide angle lens installed. Walkie talkie with jumper across PTT switch for communication. First flight with Karia was very successful with only one missed turn signal going away and two coming back. (I've been flying the course way too long!) At the second missed return chime I realised I had (fortunately -see later) not ticked the video record box. Thumbs up - big smile! Checked computer and found it had frozen with the "Windows screen saver showing. Restarted and ticked video record box. 2nd Flight with Kompressor, a faster and more colourful model. Only one or two chimes heard during the whole flight. Checked the computer and it had stopped recording because it had exhausted the Hard Disk capacity. A file of about 600Mb had been recorded. I had to discard some old and temporary files to get Windows to reboot and hence recover the FxxTrack video for ftp transmission. Hopefully the video may show you what was happening. Steve

Nov 11, 2009, 04:26 AM	#298
jjmouris High Performance Join Date: Aug 2008 Location: Southampton, UK Posts: 2,519	What he said. I also wonder if in general recording the video interferes with the operation of the program. Seems like any processing capabilaty and also RAM memory taken away from the program will result in hickups?!

Nov 11, 2009, 04:30 AM	#299
Alan Flockhart Alan Join Date: Sep 2005 Location: Cambridge Posts: 196	Very possibly. My lappie when running FXX and saving to disk doesn't put much above 40% use on the CPU. memory is within the bounds on the machine as well (Though I do shutdown everything like IM clients, phone sync etc etc to be sure it isn't trying to do something in the background)

Nov 11, 2009, 04:44 AM	#300
Mutchy Registered User Join Date: Dec 2003 Location: Wollongong Australia Posts: 76	You could have also been too high/too low. We have seen that all season in the UK. First leg high going past the line is missed as out of camera field of vision. Diving turn carried out means plane is visible by FXX on the rtn and you get a buzz. Same for the bottom leg. Going out of course you get a buzz then on rtn you are below the camera angle and FXX can't see you. Alan, As far as I am aware I flew exactly the same with the second model as the first. The two flights were within 20 minutes of each other. I believe it is unlikely that during the first flight the system could chime 37 times out of 40 and only 3 out of 40 the second flight with my flying being too high/low being the cause. I'd think that if the computer was running out of disk space then the OS wouldn't be giving FXX much chance to do stuff. The OS would have been "worrying" about what to do with its own data never mind trying to write the GB's which FXX would have been throwing at it.[/quote] Yes. The OS would be trying to find space to store all that uncompressed video data. I have cleaned up the partition and found enough space to store a couple of videos so tests tomorrow morning will tell me if my CPU can walk and chew gum simultaneously. Steve