mtl

mtl (moving time-lapse) is a python tool to create time lapse animation from photos taken not from a fixed camera (hence 'moving') with identifiable markers.

mtl align time series photos with markers (3 or 4 markers) provided as .TPS file (digitized with TPSDig software), and output the aligned photos and time-lapse movie.

Left: unaligned photo sequence; Right: aligned by mtl

requires

mtl is based on OpenCV's [1] implementation of affine transformation (with 3 markers provided) and perspective transformation (with 4 markers provided) [2].

Output of time-lapse video is based on ffmpeg [3]. To use mtl, both OpenCV and ffmpeg are required.

how to use?

1. Use as a python package. mtl is on PyPI and can be installed with pip (in command line, i.e. cmd in Windows [4] / terminal in Unix systems):

pip install mtl

2. Directly use the mtl.py python module, if you prefer. Download the file.

mtl can be directly used as command line script, with the following arguments:

-h, --help	show this help message and exit
-t, --tps	path to tps file containing landmarks for alignments
-i, --img	path to the directory containing images to be aligned
-s, --sep	separator between individual and time in image name. NOTE: use single quote (') for special character in Unix systems

Alternatively, mtl can be imported into python (in python):

from mtl import align

The main function of mtl is align, which provides more options. For further details run (in python):

help(align)

preparing images and markers file

mtl supports batch processing of multiple time series photos. Different time series (such as 'individuals') and time points should be indicated in the file name of the images. For examples, 1-1.tif, 1-2.tif, ..., 1-100.tif and a-1.tif, a-2.tif, ..., a-100.tif will be processed as two different time series of '1' and 'a' with time points of 1, 2, ..., 100. These images should be placed in a single directory. A dash '-' is used to separate the time series and time points here so this should be instructed to the program. Only a single .TPS file is required for processing multiple time series photos, and it should contains markers for all images in the directory to be processed.

application

This tool was written for a friend's thesis, for monitoring organism's motility:

Lin, J-C. (2018). Effects of water-flow and distance from neighbours on the mobility of the turtle barnacle Chelonibia testudinaria (Crustacea: Cirripedia). (MSc Dissertation). National Taiwan Ocean University. Available at: https://hdl.handle.net/11296/8ue54q

Which resulted in this publication:

Five hundred million years to mobility: directed locomotion and its ecological function in a turtle barnacle

notes

[1]	OpenCV's installation guide for Windows; Installation guide for Debian systems.

[2]	A nice explanation on the transformation methods can be found here.

[3]	FFmpeg's installation guide for Windows; In Ubuntu it can be installed with apt (apt-get install ffmpeg)

[4]	Use pip in Windows