This document is still a work in progress, and this is an old version of this code (that may end up being developed in parallel with the new version). Consider everything broken unless proven otherwise.
Python 2.7 is required - if you already have a copy of Python 2.7 (such as with ArcGIS), you'll need to make sure you have "pip" - otherwise, this installer comes with it. Download and install this file (Python C++ compiler support) Download and install this file (database support) Download and install this file (image support) Open a command prompt and run the following commands pip install arrow pip install six pip install numpy pip install usgs-api pip install matplotlib pip install docopt Download CWS Code Library - Unzip, the files. The files will be installed in the same location you extracted them, so extract to somewhere permanent. Then run the file "install_code_library.py" inside the first subfolder. If you installed Python using the above installer, double clicking should be enough to run it. Otherwise, you may need to log out, then log back in, open a command prompt, navigate to that folder the code library files are in, and run "python install_code_library.py". When installing, if it pops up a black box and asks you to "Press enter to close" then your installation should have been successful. Download and install this file if you want to use calculated sun angles to sort your images from your game camera - accepting the defaults should be sufficient. Download the image day/night sorting script - extract somewhere you can find it later. Install Strawberry Perl Download and extract VirtualDub to a location you'll remember - if you already have and know how to use FFMPEG to merge frames into a video, you can use that instead
See the video training on using this software on YouTube
- Start with gage/logger data and photos from each river
- Copy code from moving_hydrograph folder=code for each video. Get new copy of this every time and then edit. Nick has blank copies and can give to you, or you can retrieve from this github repository for this need to add moving_hydrograph folder
- First, deal with image folder:
Go to image folder and thumbnail view. Find script, sort_by_sun_angle.py, copy and paste it into image folder (open command prompt first within the picture folder—by holding down shift, this opens the folder and then you can run the script within the folder). Run script. If this works, you’ll get two folders—one for day and one for night. Need to provide two parameters—latitude and longitude. Ex: --lat=37.7 --long=120.7 (need to look this up for your location each time you do this)
Or, you can just sort by size and delete night images. Not as good as the script, but is suitable if it doesn’t work This method has issues with photos on the fringe (low light/dusk time)—this method is more tedious but it works.
Once script runs, review folders to look for dark and/or fringe pictures and cull as necessary
- Then, decide which river will be your master video (aka video on the left of screen, everything else is based off of this one and its frame rate)
- Move pictures from master video into “pics” folder, secondary river into “pair_pics” folder. You can do these streamlapses with just one river, just don’t do anything to the pair folders and set the correct parameters in
composite.pl
- Now, need to generate graphs
- Need to generate a graph for each time/photo in this streamlapse
- Cut down your stream data to only the time you need for your video, do .csv edits in Notepad++ (not Excel, usually tries to format dates and messes things up)
- Move two data files into appropriate folders (
graphsandpair_graphs) - Need two python files (from moving hydro folder) into
graphsfolders - Decide what frame size to use
- To help decide: open an image in an image editor, determine your pixel width and height. Height is number you see, then multiply that by 4/3 to get the width. This makes it look nicer and consistent.
- Change
frame_size(what you just determined) inmoving_hydrographs_figures.py - What you need to configure goes down to ~line 51 (probably shouldn’t mess with any of the other stuff unless you know what you’re doing)
Output_dpisets pixel density of images on outputOutput_folderwhere it’s going to put the graphs- Setting colors: Set colors for data in past, present, future
- Colors are in web format
- Keep # sign in front of color names in the code
- Rename csv data (ex. TUO_data.csv—this has to match one of the prefixes in the code, needs to end in _data)
- Check date and flow fields
- Run
moving_hydrographs_figures.py - Don’t leave folder open while it’s running, might slow process down
- Check the output to make sure the points are moving along the graphs logically
- If data is noisy and hard to look at, set moving average to True
- If you want to set moving average to true for flow as well, you’d
have to add another line (is false by default):
Plot_item[1].use_moving_average = True - You can’t really see the axis labels so don’t worry about the labels, we can fix this, but not super important. These graphs are potentially something we could plot in R and make look much nicer.
- Repeat this for
pair_graphs
- Next script—indexes graphs and pair rivers images, assigns photo a time range, and matches it to the graph. Takes graph and puts it on an image, does this for pair river too, after this, should have all frames of video (but not a video yet)
- Open Perl script
- If you want only one image, set:
paired_images = 0, if not = 1 - Open command window, type:
perl composite.pl
- Run VirtualDub—turns image frames into a video
- Open first image from folder with composite images (small version), program will load remaining files.
- Click on Video, then Frame Rate, then set frame rate (usually 10fps because we have hourly images, or for 15min camera = 30fps)
- Save as .avi in output folder (or wherever you like).
- Once have AVI, bring into Windows Movie Maker (or similar) and add in annotation (such as "UC Davis Center for Watershed Sciences - http://watershed.ucdavis.edu" - then, export as mpg or wmv to get compressed