Auto_Weed_IP

Automatic Weed Image Processing

Automatic weed classification & segmentation powered by YOLO.

Requirements

python 3.13 (maybe), go 1.24.4

Installation & Server Deploy

• Build from source
  1. Clone the repo : git clone https://github.com/haochend413/Auto_Weed_IP.git
  2. System permission: chmod -x setup.sh localserver.sh
  3. Environment setup: ./setup.sh. *Note: This might take a while;
  4. To run server locally on 8000 port : ./localserver.sh (Right now the only option).
  5. build & start terminal CLI:
     • cd frontend
     • go build -o awd

CLI

Built with Golang & Cobra.

• detect : use detect with -i flag ./awd r -d path_to_the_images_folder

• segmentation : use seg with -i flag ./awd r -s path_to_the_images_folder

• classification : use cls with -i flag ./awd r -c path_to_the_images_folder

ID	Common Name	Role	Species	EPPO Taxon Code	Name
0	chinee apple	weed	ziziphus mauritiana	ZIPMA	weed: ziziphus mauritiana
1	lantana	weed	lantana camara	LANCA	weed: lantana camara
2	parkinsonia	weed	parkinsonia aculeata	PAKAC	weed: parkinsonia aculeata
3	parthenium	weed	parthenium hysterophorus	PTNHY	weed: parthenium hysterophorus
4	prickly acacia	weed	vachellia nilotica	ACANL	weed: vachellia nilotica
5	rubber vine	weed	cryptostegia grandiflora	CVRGR	weed: cryptostegia grandiflora
6	siam weed	weed	chromolaena odorata	EUPOD	weed: chromolaena odorata
7	snake weed	weed	gutierrezia sarothrae	GUESA	weed: gutierrezia sarothrae
8	negative	na			none

The results will be stored in ~/Awd_Results folder as a zip file named after time stored.

GUI

Built with TypeScript + Next.js.

• automatic detection / segmentation / classification
• local upload
• manual labeling / fine-tuning results
• downloadable image / data

Server & Database

Built with Fastapi + SQLModel + PostgreDB. Powerful Component functionalities provided by shadcn-ui