Use JavaCV on a raspberry pi

Original goal was to check the difference of speed between standard compiled OpenCV (so origami) and JavaCV. I also originally thought JavaCV was compiling OpenCV with Gpu bindings, it is not.

more relevant javacv samples

http://bytedeco.org/news/2015/04/04/javacv-frame-converters/ https://github.com/bytedeco/javacv/blob/master/samples/ImageSegmentation.java

Installation.

extra libraries missing on modern raspbian ...

Things that were missing on the raspberry pi

apt install libgfortran3

Jupyter notebook integgration for both clojure and java

pip install jupyter

extra clojure jupyter setup

## install leiningen
## https://leiningen.org/#install

lein jupyter install-kernel

extra java jupyter setup

# download: https://github.com/SpencerPark/IJava/releases
wget https://github.com/SpencerPark/IJava/releases/download/v1.3.0/ijava-1.3.0.zip
unzip  ijava-1.3.0.zip
# with pyenv change the prefix for the kernel
python3 install.py --prefix=/home/pi/.local

# make sure you are using java 8 for java (broken with later sdks ..)
sudo update-alternatives --set java /usr/lib/jvm/java-8-openjdk-armhf/bin/java

# Although this seems to work ?
sudo update-alternatives --set java /usr/lib/jvm/java-11-openjdk-armhf/bin/java

now check the java and clojure kernels are installed:

pi@raspberrypi:~/Downloads $ jupyter kernelspec list 
Available kernels:
  java            /home/pi/.local/share/jupyter/kernels/java
  lein-clojure    /home/pi/.local/share/jupyter/kernels/lein-clojure
  python3         /home/pi/.local/share/jupyter/kernels/python3

# start the notebook on a given ip
export LEIN_WORKING_DIRECTORY=$PWD
jupyter notebook --ip=0.0.0.0