Code to download webcams from URLs
Fetching latest commit…
Cannot retrieve the latest commit at this time
--- README --- This python software package automatically downloads images from online webcams. It was used to gather the Webcam Clip Art Dataset available at: http://graphics.cs.cmu.edu/projects/webcamdataset The code has the following features, see below for additional details: - multi-threaded (one main thread, one or many downloading threads); - computes the sun position based on gps and time to download only daytime images; - checks for image corruption (this happens often with webcams); - emails periodically to let you know that it's still running; - emails if there is a problem; - avoids downloading duplicates; - tags imags with current date and time (in UTC coordinates); - supports different sampling intervals and total time (can be set to infinite, or continuous download); - logs output to file with different levels of verbosity (adjustable); - uses an xml file for listing the webcams to download, so the list can be changed on the fly without having to restart; Most of these options can be turned on/off manually. -- Command line example -- python DownloadImagesThreads.py /path/to/webcams.xml /path/to/images/ 10 /path/to/logdirectory/ -- Checking for image corruption -- Checking for image corruption (filterJpg=True) requires the third-party C software package 'Jpeginfo' by Timo Kokkonen. Version 1.6.0 is included herein for convenience. Make sure the path to the compiled executable is given to DownloadImagesThread.py. -- Sun position -- Computing the sun position (filterSunAltitude=True) requires the third-party python package 'pysolar' by Brandon Stafford. I have included a version herein for convenience. Make sure the PYTHONPATH points to that package as well. -- Duplicates -- We can avoid downloading duplicates by setting filterExistingFile=True. --- Emails --- The program supports email notifications when emailNotification=True. It should send email: - at a certain frequency (e.g. every day), to keep you informed that it is still running (so you don't have to check); - set emailRunningNotificationInterval to the interval at which you want to receive emails - if a thread is still trying to download after the time interval has elapsed; - if launching all the threads take longer than the time interval. -- Logging -- The caller thread and each downloading threads have their own log files, which get created automatically from the caller log file passed in as input. You can change the verbosity of the loggers via the 'setLevel' function. By default, the caller thread has level INFO, and each downloading thread DEBUG. -- Python version -- This has been tested on Mac OSX 10.6 with python 2.6.2.