Today the webserver with the Caedmon webcam pointing to the Acre and the James Graham Building was restarted again. It had been off for an unknown time, probably due to a temporary shut-down a short while ago. Since I no longer have an office in which this webcam is located, I was not able to immediately restart it after I had noticed the shut-down. Another issue hid this shutdown from me for a while: due to a wrong configuration of the 2nd webcam server (which points to the parking lot between Priestley and Caedmin), the picture of that server replaced the one from the main webcam, with the result that the supervisory process (which checks if there is no new image within 24 hours) has not detected a lack of images, and therefore has not sent a notification about the webcam outage.
This has now been fixed, and both webcams run properly as intended. The Headingley Caedmon Webcam has indeed been running for several months without me ever visiting or maintaining the server.
Friday, November 19, 2010
Wednesday, August 4, 2010
One week camera outage
The webcam (with its PC) had been running continuously since March 2010 - quite robustly without any problems. On Friday 23.July 2010 in the afternoon a brief power outage occurred at the Headingley campus. And after that the webcam image was no longer updated. I had no opportunity to fix this problem before now, so the cam image remained steady for more than a week - apologies!
When I checked today, the computer was completely shut-down. It might have also been that the PC was shutdown by someone, unrelated to the power outage. Restarting it had a slight problem, due to two different webcam software version that had been installed. Version 1.0.5 was the one that ran now properly. In any case the camera is up and running now fine.
When I checked today, the computer was completely shut-down. It might have also been that the PC was shutdown by someone, unrelated to the power outage. Restarting it had a slight problem, due to two different webcam software version that had been installed. Version 1.0.5 was the one that ran now properly. In any case the camera is up and running now fine.
Monday, March 15, 2010
Work-around for camera capture problem
I was able to find a work-around for the problem of the camera capture software: After the porting of the software to the "new" computer it sometimes went into a state where the DirectX camera preview just stopped. This did not manifest itself as an error, but instead the software kept running as if nothing had happened - and kept uploading the image that had last been captured ok. Only a restart of the software helped - so I programmed into the software an automatic restart at midnight.
This seems to work - I noticed yesterday that the image had again been frozen; now, however, the image is being updated ok.
Is not an ideal fix - I would like to explore what the true reason for this error is... but at least for now the camera capture works as intended.
This seems to work - I noticed yesterday that the image had again been frozen; now, however, the image is being updated ok.
Is not an ideal fix - I would like to explore what the true reason for this error is... but at least for now the camera capture works as intended.
Wednesday, January 20, 2010
Some Problems...
Since the porting of the webcam onto the new PC at the beginning of this year, there have been some instabilities: the image appears suddenly not to be updated, while the software still sends new images in a regular time interval to the server. But these images are constant - so on the webcam viewers (embedded images in websites such as this blog) the image appears not to change, while the timestamp is constantly being updated.
On the capture PC, this problem shows up as the software running properly, but the image just being black. No error message or exception was reported. The same version of the software runs fine in the Back-Caedmon camera; so I assume that the problem lies with the PC in 101 Caedmon which may be too slow. Today I have reduced the image capture size from 1280x1024 to 640x480, in order to reduce the requirements of bandwidth, image storage and processing for each captured frame. I will have to monitor if this solves the problem.
On the capture PC, this problem shows up as the software running properly, but the image just being black. No error message or exception was reported. The same version of the software runs fine in the Back-Caedmon camera; so I assume that the problem lies with the PC in 101 Caedmon which may be too slow. Today I have reduced the image capture size from 1280x1024 to 640x480, in order to reduce the requirements of bandwidth, image storage and processing for each captured frame. I will have to monitor if this solves the problem.
Friday, January 15, 2010
Back-Caedmon Camera Off
The camera in my new office facing into the back parking lot between Caedmon Hall and Priestley Hall has been switched off over the weekend - the USB cam is needed for another project. So for the time being, there will be the only the most recent image shown in the top right corner of this blog.
I assume that I will be able to switch this camera on again on Monday morning.
I assume that I will be able to switch this camera on again on Monday morning.
Thursday, January 14, 2010
New Software Testing
The latest version of the camera capture software is now under beta test with the Back-Caedmon camera in my office (see top right small image on this blog).
The main feature of this version is that it measures the change within an image with respect to a reference image, and when this change exceeds a threshold, it automatically captures a new image. This will change the picture automatically more frequently when there are activities in the scene.
There is a baseline capture frequency, set at 10 minutes. When 10 minutes are over, a new image is captured, independent of the change, and it is then used as the new reference image. This reference baseline capture could be extended to longer times.
From first experiments today, it can be seen that the time interval for the automatically captured images is around 3-7 minutes, when the threshold for change is set to a value of 10 (this value is normalised to 255: a max value of 255 would mean that all pixels in an image change from completely black=0 to completely white=255).
I will keep the Back-Caedmon camera running with this software, and if it runs stable I will then update the regular Headingley-Caedmon Webcam with this version.
The main feature of this version is that it measures the change within an image with respect to a reference image, and when this change exceeds a threshold, it automatically captures a new image. This will change the picture automatically more frequently when there are activities in the scene.
There is a baseline capture frequency, set at 10 minutes. When 10 minutes are over, a new image is captured, independent of the change, and it is then used as the new reference image. This reference baseline capture could be extended to longer times.
From first experiments today, it can be seen that the time interval for the automatically captured images is around 3-7 minutes, when the threshold for change is set to a value of 10 (this value is normalised to 255: a max value of 255 would mean that all pixels in an image change from completely black=0 to completely white=255).
I will keep the Back-Caedmon camera running with this software, and if it runs stable I will then update the regular Headingley-Caedmon Webcam with this version.
Tuesday, January 12, 2010
Problem with the Webcam
There appears to be a problem with the webcam system: the communication between the camera and the capture software stops, and no new image is captured. This happened now already for several times in the last few days - the current image, for example, is one from during the day. It appears that the software keeps running and uploading the same file - hence the new image date/time. But the image itself remains the same.
The reason for this could be either the cold which influences the camera itself, or the new PC and its OS environment where the capture software is running. I can now compare this PC with the "old" PC which now runs the Caedmon-Back-Camera. This one also had shown the connection error today during the day, but now appears to run stable. I will check tomorrow.
The reason for this could be either the cold which influences the camera itself, or the new PC and its OS environment where the capture software is running. I can now compare this PC with the "old" PC which now runs the Caedmon-Back-Camera. This one also had shown the connection error today during the day, but now appears to run stable. I will check tomorrow.
New Software Version - and Additional Location
The capture software has been upgraded to allow a slightly easier use.
In addition, a second instance of the webcam has now been set up in my office, showing a view onto the parking lot between Caedmon Hall and Priestley Hall. This webcam will not be advertised much, but will be embedded in a few of my sites for esperimentation. It also will allow me to test new features before releasing them onto the main webcam which points to the Acre.
In addition, a second instance of the webcam has now been set up in my office, showing a view onto the parking lot between Caedmon Hall and Priestley Hall. This webcam will not be advertised much, but will be embedded in a few of my sites for esperimentation. It also will allow me to test new features before releasing them onto the main webcam which points to the Acre.
Wednesday, January 6, 2010
Webcam Running with Acre View Again
The Leeds Met Caedmon Webcam is again at its location in office 101. After I had moved office in the beginning of November 2009, I had to dismantle the camera setup and take it with me to my new office location, which did not have such an attractive view (into the back of the buildings Caedmon Hall and Priestley Hall). However, the plan was to setup a new PC, and the webcam would be operated further.
This did not work initially, as there were some technical issues with that PC, and I did not have the time to take care of these issues during the past 2 months. Yesterday, however, I was able to setup the software properly and to reinstate the web cam setup at its original location. Since this morning, camera images from the Acre and the James Graham Building are sent again to our web server.
As you see, there is now a winter wonderland here in Leeds - a snow blanket of at least 15 cm is covering everything, making road traffic a nightmare, but providing a wonderful sight!
This did not work initially, as there were some technical issues with that PC, and I did not have the time to take care of these issues during the past 2 months. Yesterday, however, I was able to setup the software properly and to reinstate the web cam setup at its original location. Since this morning, camera images from the Acre and the James Graham Building are sent again to our web server.
As you see, there is now a winter wonderland here in Leeds - a snow blanket of at least 15 cm is covering everything, making road traffic a nightmare, but providing a wonderful sight!
Tuesday, January 5, 2010
Webcam at Temporary Location
A happy New Year 2010 to everyone!
In the past two months the webcam has been down. The reason for this was that a new PC had to be used in order to keep the webcam at the same location, and there were some issues regarding the installation of that software which I had no time to address, due to other work that I had to do. I am now addressing these issues, and I hope to have the webcam up and running again at the previous location.
As a temporary fix, I am now using the previous PC where the functioning release of the capture software is running. This does NOT point to the Leeds Met Acre and the James Graham building, but points instead to the parking lot between the Caedmon and Priestley buildings. This view is less attractive than the front view of the James Graham building, but since it is currently snowing, the view is not that bad.
The issue regarding the software is quite strange: despite having installed the latest .NET releases, the installation routine requires some particular files to be installed. It could be that these are a part of MS Visual Studio, which had been installed on both the PCs where I had developed and used the webcam before. I hope to figure this out soon. In the meantime, the camera will be operational from this view behind Caedmon Hall. The building in sight (to the right) is Priestley Hall. In the background on the left you can see the Metceno building. And at the extreme right there is a part of Caedmon Hall.
In the past two months the webcam has been down. The reason for this was that a new PC had to be used in order to keep the webcam at the same location, and there were some issues regarding the installation of that software which I had no time to address, due to other work that I had to do. I am now addressing these issues, and I hope to have the webcam up and running again at the previous location.
As a temporary fix, I am now using the previous PC where the functioning release of the capture software is running. This does NOT point to the Leeds Met Acre and the James Graham building, but points instead to the parking lot between the Caedmon and Priestley buildings. This view is less attractive than the front view of the James Graham building, but since it is currently snowing, the view is not that bad.
The issue regarding the software is quite strange: despite having installed the latest .NET releases, the installation routine requires some particular files to be installed. It could be that these are a part of MS Visual Studio, which had been installed on both the PCs where I had developed and used the webcam before. I hope to figure this out soon. In the meantime, the camera will be operational from this view behind Caedmon Hall. The building in sight (to the right) is Priestley Hall. In the background on the left you can see the Metceno building. And at the extreme right there is a part of Caedmon Hall.
Subscribe to:
Posts (Atom)
