30_plant_types_image_detection Open-source Plant Image Recognition Model

30 Plant Types Image Detection

Developed by dima806

This is an image classification model based on the Vision Transformer (ViT) architecture, capable of accurately identifying 30 different types of plants with an average accuracy of 93%.

Image Classification

Transformers

Open Source License:Apache-2.0 #High-precision plant identification #Agricultural image classification #ViT visual model

Downloads 27

Release Time : 10/29/2023

Model Overview

The model uses google/vit-base-patch16-224-in21k as its base model and is specifically designed for plant type recognition tasks. It can predict the specific species of plants from input images and has broad application value in fields such as agriculture and botanical research.

Model Features

High Accuracy

Achieves an overall accuracy of 93% on the 30-class plant classification task, with certain categories such as coconut and sweet potato reaching over 99% accuracy.

Multi-category Recognition

Capable of identifying 30 different plant types, covering common crops, fruits, and medicinal plants.

ViT-based Architecture

Utilizes the advanced Vision Transformer architecture, which outperforms traditional CNN models in image classification tasks.

Model Capabilities

Plant species identification

Image classification

Agricultural image analysis

Use Cases

Agriculture

Automatic Crop Identification

Used in farmland monitoring systems to automatically identify different crop types

Can accurately identify major crops such as rice, corn, and cassava

Fruit Quality Inspection

Identifies different fruit types on sorting lines

Achieves over 95% accuracy in recognizing fruits like bananas, oranges, and mangoes

Botanical Research

Plant Specimen Classification

Assists botanists in quickly classifying plant specimens

Can identify various plants, including medicinal plants like aloe vera and turmeric

🚀 Plant Type Prediction Model

This project provides a model that predicts the plant type from an image with approximately 93% accuracy. It's a valuable tool for plant identification tasks.

🚀 Quick Start

The model predicts plant types from images. It uses the base model google/vit-base-patch16-224-in21k and can achieve about 93% accuracy.

For more detailed information, please refer to this Kaggle notebook.

📦 Installation

No specific installation steps are provided in the original document.

💻 Usage Examples

The following is the classification report of the model:

Classification report:

               precision    recall  f1-score   support

        guava     0.9846    0.9600    0.9722       200
     galangal     0.9418    0.8900    0.9152       200
      bilimbi     0.9949    0.9750    0.9848       200
        paddy     0.9731    0.9050    0.9378       200
     eggplant     0.9848    0.9700    0.9773       200
     cucumber     0.9561    0.9800    0.9679       200
      cassava     0.9899    0.9800    0.9849       200
       papaya     0.9851    0.9950    0.9900       200
       banana     0.9950    0.9900    0.9925       200
       orange     0.9534    0.9200    0.9364       200
   cantaloupe     0.5271    0.3400    0.4134       200
      coconut     0.9950    1.0000    0.9975       200
     soybeans     0.9754    0.9900    0.9826       200
       pomelo     0.9563    0.9850    0.9704       200
    pineapple     0.9703    0.9800    0.9751       200
        melon     0.5000    0.6150    0.5516       200
      shallot     0.9949    0.9750    0.9848       200
   peperchili     0.9755    0.9950    0.9851       200
      spinach     0.9231    0.9600    0.9412       200
      tobacco     0.9151    0.9700    0.9417       200
     aloevera     0.9949    0.9800    0.9874       200
      curcuma     0.9005    0.8600    0.8798       200
         corn     0.9610    0.9850    0.9728       200
       ginger     0.8551    0.8850    0.8698       200
sweetpotatoes     1.0000    0.9950    0.9975       200
         kale     0.9268    0.9500    0.9383       200
    longbeans     0.9850    0.9850    0.9850       200
   watermelon     0.9252    0.9900    0.9565       200
        mango     0.9239    0.9100    0.9169       200
   waterapple     0.8807    0.9600    0.9187       200

     accuracy                         0.9292      6000
    macro avg     0.9282    0.9292    0.9275      6000
 weighted avg     0.9282    0.9292    0.9275      6000

📚 Documentation

Model Information

Property	Details
Model Type	Based on `google/vit-base-patch16-224-in21k`
Metrics	Accuracy, F1

📄 License

This project is licensed under the Apache-2.0 license.

Featured Recommended AI Models

Empowering the Future, Your AI Solution Knowledge Base

English 简体中文繁體中文にほんご