Data

Tools

Indexes

Applications

Experiments

Open Source Science

weed_detection_csn

Weed detection method with training datasets

https://github.com/fcc-d23/weed_detection_csn

Science Score: 26.0%

This score indicates how likely this project is to be science-related based on various indicators:

  • ○
    CITATION.cff file
  • ✓
    codemeta.json file
    Found codemeta.json file
  • ✓
    .zenodo.json file
    Found .zenodo.json file
  • ○
    DOI references
  • ○
    Academic publication links
  • ○
    Academic email domains
  • ○
    Institutional organization owner
  • ○
    JOSS paper metadata
  • ○
    Scientific vocabulary similarity
    Low similarity (8.2%) to scientific vocabulary

Keywords

dataset deep-learning weed-detection
Last synced: 9 months ago · JSON representation

Repository

Weed detection method with training datasets

Basic Info
  • Host: GitHub
  • Owner: fcc-d23
  • License: mit
  • Language: Python
  • Default Branch: master
  • Homepage:
  • Size: 1.36 GB
Statistics
  • Stars: 2
  • Watchers: 1
  • Forks: 1
  • Open Issues: 0
  • Releases: 0
Topics
dataset deep-learning weed-detection
Created about 3 years ago · Last pushed over 2 years ago
Metadata Files
Readme License Citation

README.md

Weed detection method

The main code used by the method described in the paper is uploaded in a separate folder, which contains both the libraries used to train the models and perform segmentation over full images. This folder contains also contains a simple example of use of the method over a folder of images (in this case the one in the database dubbed "Full Images") which can be used to assess the method in question. The code was tested using Numpy 1.19.5, TensorFlow-GPU 1.14.0 , Keras 2.2.3 and OpenCV 3.4.7.28.

Weed detection database

This repo contains the database that was used for training both the region detection model (the CSN architecture) and the sub-frame detection model (the CNN architecture). The database is structured as follows: 1. First, images are divided between full pictures (from which examples come from) and the examples themselves 2. In the examples folders the pictures are then arranged into Region and Sub-Frame folders 3. In these folders examples are classified into Structured and Unstructured weeds, as well as into a Grass Subfolder 4. Finally, for each of the latter class-folders, sets of train and test images (which were generated randomly via a python code) are presented as were used during training, validation and testing stages

Additionally, outputs of the proposed method over selected testing images are available in the Segmented images folder.

Owner

  • Login: fcc-d23
  • Kind: user

GitHub Events

Total
  • Watch event: 1
Last Year
  • Watch event: 1