About

KM-MOD provides a standalone image classifier for use with date-time stamped (YYYYMMDD-HHMMSS) images which most security cameras produce. Repetitive images are filtered out leaving only images of interest. Selected images are converted into a time-lapse video.

Requirements

• RPI4 2GB or higher
• Python 3.4 or higher

Setup

python -m pip install -U pip
python -m pip install -U scikit-image 
pip install opencv-python
pip install shutils
pip install -U scikit-learn (for kmeans)
pip install matplotlib

sudo apt update
sudo apt upgrade
sudo apt install ffmpeg

Getting Started

git clone https://github.com/SensorAnalyticsAus/KM-MOD.git

cd KM-MOD

Config

Paths and the DV identifier string need to be set to actual paths/value at the beginning of these files. The DV value must be the same among daily-driver-mp, moviefrm-list, and moviefrm-list-ni

config.py
daily-driver-mp
moviefrm-list
moviefrm-list-ni (non-interactive version used by `daily-driver-mp`)

Example 1

Step 1 train

/path/to/.venv/bin/python train-km-mp.py on 1 10 20240506000000 20240506235959

Step 2 predict (output frames from selected clusters from step 1)

/path/to/.venv/bin/python predict-km.py 80

Step 3 create a time-lapse video

./moviefrm-list 20

Example 2

./daily-driver-mp on

or

./daily-driver-mp off

off prevents the progress-bar from showing, useful for cron This is best run at say 7am and 7pm daily as a cron job