-install opencv 4.4.0 sesuai jumlah core CPU
-install darknet yolov4
-download/git clone : https://github.com/AlexeyAB/darknet
-edit dulu darknet/scr/detector.c agar bisa verbose report ke file.txt
-build darknet

Ektrak plastik-pic.zip dan zip file dalam folder plastic-pic-1024 yang ada di ciliwung-images
https://drive.google.com/drive/folders/1nnvJ41VV45ncbqa6nMQY3ImjVByOHnjd?usp=sharing

Rename plastik-pic jadi plastic-pic.

direktori :
|
|---plastic-pic
|
|---plastic-pic-1024
|
|---darknet

copy isi folder darknet(cfg dan *.sh) dalam ciliwung-tools

https://drive.google.com/drive/folders/1DALM9oE-qm7ZTEfTFvrzyIc4Sgpk5vx2?usp=sharing

masuk folder darknet,
run :

~/darknet$ ./yolov4tinyplastik.sh

Untuk labelling/annotation, gunakan labelimg. Contoh :

python3 ../labelImg/labelImg.py to_label/ classes.txt

Direktori to_label adalah direktori images yang dilabelin. Jangan lupa copy classes.txt dalam folder to_label

labelimg ada di ciliwung-tools/labelimg.zip

plastic-pic-1024
├── botol-1024
│   ├── test
│   └── train
├── botol-pestisida -> ../plastic-pic/botol-pestisida/
├── botol-plastik -> ../plastic-pic/botol-plastik/
├── ciliwung -> ../plastic-pic/ciliwung
├── ciliwung-1024
│   ├── test
│   └── train
├── ciliwung-ori -> ../plastic-pic/ciliwung-ori/
├── kresek-1024
│   ├── test
│   └── train
├── plastic -> ../plastic-pic/plastic/
├── plastic-1024
│   ├── test
│   └── train
├── transect
│   ├── test
│   └── train
└── uav -> ../plastic-pic/uav/