DeepForest Object Detection Model for Immature Oil Palm Tree

About

This application allows for the multi-class detection of immature oil palm trees in images using a custom-trained model built with the DeepForest library. It provides a user-friendly interface to upload images and visualize the detection results.

The model is trained to identify the following classes:

• tbm-1 (0-12 months)
• tbm-2 (13-24 months)

Run Demo

Features

• Detect objects (Immature Oil Palms) in user-uploaded images.
• Supports various image formats (JPG, JPEG, PNG, TIF, TIFF).
• Displays detection bounding boxes, class labels, and confidence scores.
• Allows download of the visualized image and a CSV summary of detections.
• Uses a demo image if no image is uploaded.

Tech Stack

• Python
• Streamlit
• DeepForest
• PyTorch
• OpenCV
• Pillow
• RasterIO
• Pandas
• NumPy

Run Locally

Clone the project

  git clone https://github.com/ramalpha/ImmatureOilPalmTree-ObjectDetection-App

Go to the project directory

  cd ImmatureOilPalmTree-ObjectDetection-App

Install dependencies

  pip install -r requirements.txt

Start the server

  streamlit run tbm_od_app.py

Author

• Ramadya Alif Satya