Image-Based Layout Modeling and Pathfinding

Overview

This project aims to convert real-life maps, layouts, or floor plans into computer-understandable models for various applications, such as shortest path finding. A key application is to use scaled maps of warehouse interiors to model warehouse shelves and optimize navigation for store personnel.

Features

• Image Processing: Enhance contrast, apply Gaussian blur, adaptive thresholding, and upscaling.
• Rectangle Detection: Detect and filter rectangles representing shelves or obstacles.
• Clustering: Cluster detected rectangles by size or distance using KMeans.
• Edge Connection: Create connection lines between non-intersecting rectangles.

Installation

Prerequisites

• Python 3.7 or higher

Steps

1. Clone the repository:

   git clone git@github.com:akgunfu/layroad.git
   cd layroad

2. Run the installation script:

   ./install.sh

Usage

Command-Line Interface

The main script, main.py, can be run from the command line to process images:

python main.py -i <input_directory> -o <output_directory> -m <max_images> -f <file_path>

Arguments

Argument	Description	Default
-i, --input_dir	Directory containing input images	assets
-o, --output_dir	Directory to save output images and shapes	outputs
-m, --max_images	Maximum number of images to process	3
-f, --file_path	Path to the input image file	None

Example

Bare usage

python main.py

To process images from the assets directory and save the results to the outputs directory:

python main.py -i assets -o outputs -m 3

Server Interface

The server script, server.py, allows for uploading and processing files via a web interface.

1. Run the Server:

   python server.py

2. Upload a File: Use curl or Postman to upload a file to the server.

   Using curl:

   curl -F "file=@/path/to/your/file.pdf" http://127.0.0.1:5000/process

Server Endpoints

Endpoint	Method	Description
/process	POST	Upload and process a file

Troubleshooting

Common Issues

1. Missing Dependencies:

   • Ensure all required Python packages are installed by running ./install.sh.

2. File Not Found:

   • Ensure the input directory or file path specified exists and contains the correct images.

3. Permission Errors:

   • Ensure you have the necessary permissions to read from the input directory and write to the output directory.

4. Invalid Image Formats:

   • Ensure the images are in supported formats (PNG, JPG, JPEG, PDF).

Sample Errors and Solutions

1. Error: OSError: poppler not found

   • Solution: Install poppler on your system:

     brew install poppler

2. Error: FileNotFoundError: Image not found

   • Solution: Ensure the image file paths are correct and the files exist in the specified directory.

Documentation (In Progress)

Rectangle Detection

Line Detection