               ######        ###       ##     ##    ########    ########     ######   
              ##    ##      ## ##      ##     ##       ##       ##          ##    ##  
              ##           ##   ##     ##     ##       ##       ##          ##        
               ######     ##     ##    ##     ##       ##       ######      ##   #### 
                    ##    #########    ##     ##       ##       ##          ##    ##  
              ##    ##    ##     ##    ##     ##       ##       ##          ##    ##  
               ######     ##     ##     #######        ##       ########     ######

SAUTAG English Document

Dataset picture display

Project Introduction

SAUTEG is an open agricultural pest and disease image dataset, and the project aims to provide an open agricultural pest and disease image dataset collection platform.

Operating Instruction

You should understand the dataset file structure before using the project

Directory Structure

//The names of crop species, pests and diseases involved should be given priority to Latin scientific names
SATEG
├─Oryza sativa L        //Latin scientific name for a crop species
│  ├─Diseases           //This folder is used to hold the diseases dataset of the crop
│  │  ├─!Phytophthora fragariae Hickm.var.oryzo-bladis Wang et Lu    //"!"The sign indicates that there are not enough samples
│  │  ├─Entyloma oryzae Syd
│  │  └─Fusarium moniliforme Sheld
│  └─Pests             //This folder is used to hold the Pets dataset of the crop
│      ├─!Ancyllomia japonica zeller
│      ├─Aleurocybotus indicus David et Subramaniam
│      ├─Ampullaria gigas Spix
└─Sample               //This folder is used to store the optimized pest and disease training samples
   └─Oryza sativa L.Pests     //One of the training samples,semi-colon separated crop species and sample type
        ├─test                //This is the test dataset folder
        │  ├─Aleurocybotus indicus David et Subramaniam
        │  └─Ampullaria gigas Spix
        |─train               //This is the train dataset folder
        |   ├─Aleurocybotus indicus David et Subramaniam
        |   └─Ampullaria gigas Spix
        └─README             //The documentation for the training sample should be described here when the dataset has been specially processed.
                             // For example by changing the default directory structure of the training sample document

The names of crop species, pests and diseases involved should be given priority to Latin scientific names.
Folder name before ! represents too few samples.
Such as:
!Phytophthora fragariae Hickm.var.oryzo-bladis Wang et Lu //The sample is insufficient
The folder of the Latin name of the crop in the root directory contains the original data of the crop, with Diseases holding the disease data and Pets holding the pest data.
Data in the Diseases and Pets folders should be classified as folders under Latin names.
The folder name & connects different kinds of data, which access similar diseases or pests.
Such as:
Cletus punctiger&Clketus trigonus //This folder holds two types of familiar pests

In the use of this method, species have great similarity, there is no need to separate in the prevention view.

When the Latin name of some folders exceeds the maximum length of GitHub, the use of the acronym format, in the corresponding root directory has README file to explain.
Such as:
H.S.2F.M
The name of the folder that is too long after the merging of 5 kinds of acquaintance.Using the abbreviation form, create the README file in the root directory to explain.
The Sample folder stores samples that can be used for training directly. The Sample folder name adopts the naming method of cropName. Category
The sample folder root may have a README file that describes the dataset sample

Manual

Quick Start

If you want to quickly build your recognition model, just find the pre-designed dataset in the Sample folder as needed, and the dataset has been named according to the cropName. Category.

There may be a README file in the folder that explains the sample details

Original Data

The original data is stored in the root directory of the dataset with the crop Latin scientific name as the folder, and the Diseases folder holds the disease data and the Pets folder holds the pest data.

Version History

version_1.0

updated rice pests and diseases.
Oryza sativa L.Pests training samples were added to the Sample folder.

Upload Rules

You can upload and modify the project data, perfect the dataset, upload your dataset, or upload your carefully designed training sample.

Upload must according to the directory structure requirements and comply with the following rules:

Priority should be given to Latin scientific name for the names of crop species, pests and diseases involved.
If the sample quantity is too small, !should be added before the sample folder.
For example:
!Phytophthora fragariae Hickm.var.oryzo-bladis Wang et Lu //The sample is insufficient
The folder with the Latin scientific name of the crop in the root directory should contain the original data of the crop, with Diseases for disease data and Pets for pest data.Data in the Diseases and Pets folders should be classified as folders under Latin names.
You can use a folder to access several similar diseases and pests, different kinds of Latin scientific name application & division.
For example:
Cletus punctiger&Clketus trigonus //This folder holds two types of familiar pests