-
Notifications
You must be signed in to change notification settings - Fork 277
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Tracking without active browser session #36
Comments
I see what you mean. I'll figure out the fastest workaround, the best would be to just hold 1 state and no matter how many browser window you open, you see the current state of the process. But this may not be so straightforward to implement. The other workaround would be something like you say, keep the current behaviour (it actually doesn't stop counting when you close the window, it stops counting when you refresh the page or make another request from another browser), but store the counting data in a file that we can access later on.. |
Ok done, now the behaviour is that if you refresh the windows (or close it and open it later), it will display the current state of the process and not restart everything. So this workflow should works now:
Please test it and tell me if this works. I've added a note about updating the code of the app in the README: https://github.com/moovel/lab-open-data-cam#bonus-update-the-open-data-cam-app-when-you-have-already-set-up-everything Also great work on the more detailed install guide, but just one modification to it, you do not need to install next globally, it will install it localy on npm install. I just had forgotten to add in the README the build step... I've seen you figured it out, just for you to know that |
@florianporada and I just tried the data cam after your update. We noticed the following at the counter:
Any idea why this happens? |
Could you try to run the node app manually in the terminal to see what happen in the logs ? It seems that the app is crashing when starting yolo. Do something like (directly on the jetson ubuntu interface)
Interact with the app and see if the logs at telling you something meaninfull |
Here is the log:
|
I tried killing the FFserver because of the
|
There is a problem when starting YOLO... Weird, is the PATH_TO_YOLO well configured in the config.json file ? Can you try to manually go from the terminal to the darknet folder and try to run this command ?
|
...i`m getting the feeling that this could be a permission problem? |
YOLO work when you start it directly... This is weird. First we should make sure this is not caused by my last update, try to revert to the commit before the update: (I don't think it is but we never now)
If it also fails, this means it's not the new update that causes the bug, and it's something else. I'm not sure it is a permission problem because you successfully runned the thing before... I'm thinking what else could cause this to fail |
Also double check the PATH_TO_YOLO_DARKNET in config.json |
I'm kind of thinking it is caused by the PATH_TO_YOLO_DARKNET beeing wrong... Because the error ENOENT means it can't find the command |
It was indeed a problem with the path. I used the relative path ( |
Great ! I'm curious, the absolute path doesn't work in your case ? eg: |
Everything works fine now with the absolute path! 😃 |
Ok, yep maybe it needs to be documented a bit better in the README (I mention it should be the absolute path): https://github.com/moovel/lab-open-data-cam#5-download-and-install-the-open-data-cam-node-app-on-the-jetson |
FYI: The session based tracking is working but there is an issues: |
Ahhh, yes, very true... I'll fix that. Have you been able to test it in a street ? |
Yesterday I set it up in the office and tested it for a couple of hours with an external battery. So this was kind of the first test with the hardware been completely off the gird. We counted some people passing through the office. (...cars aren't allowed in yet. :) ) |
Ok, great ! I just pushed the fix you mentioned, you can update the app code and should be able to see all the counting areas when restoring the page. |
I just tested it - Looks good! 👍 |
After I tested the app on the Jetson board, I noticed, that the tracking is only happening when there is an active browser session on the client.
After closing the browser window the collected data is lost.
It would be cool, to be able to trigger the process on the jetson via the client and retrieving the active tracking-process after accessing the web-interface again.
I think it would be enough to store the tracking data in a temp json/csv file and just expose it to the webinterface as a download link.
@tdurand maybe you have an idea on how to handle the connection between running process and browser session. ? :)
something like:
(notrealcode)
If this would be implemented we also would not have the issue that the app crashes (at least the videostream) when opening multiple web sessions to the jetson board.
The text was updated successfully, but these errors were encountered: