Implementing an Open-Source SaaS Web Tool with OpenCV on my Website

1

Hello OpenCV community,

I am developing an open-source Software-as-a-Service (SaaS) web tool and plan to use OpenCV’s powerful computer vision capabilities as its core. My goal is to integrate this tool directly into my website.

My Project Idea:
The website is a resource for all things related to Michigan counties. I want to build a tool where users can upload an image (e.g., a map, a historical document, a photo of a landmark) and the tool will use OpenCV to analyze it. Potential functionalities could include:

  • Object Detection: Identifying and outlining counties, cities, or specific landmarks on a map.

  • Color Detection/Filtering: Analyzing images to find areas with specific colors (e.g., highlighting all forested areas in green on a satellite image).

  • Feature Matching: Comparing an uploaded historical map to a modern one to find correspondences.

  • Document Processing: Automatically cropping, deskewing, or enhancing scanned historical documents related to Michigan counties.

My Technical Questions & Roadblocks:

I have a basic understanding of OpenCV in Python/C++, but I’m unsure how to bridge the gap to a web environment. My main questions are:

  1. Architecture & Backend: What is the recommended backend architecture for this?

    • Should I run OpenCV in a Python backend (using a framework like Flask or Django) and have my website send image data via an API?

    • Is OpenCV.js a viable alternative for running the CV algorithms directly in the user’s browser? What are the performance and compatibility trade-offs?

  2. OpenCV.js vs. Server-Side Processing:

    • For tasks that are computationally intensive (like processing high-resolution satellite imagery), is it better to offload the work to a server to avoid freezing the user’s browser?

    • Are there specific OpenCV classes or modules (like cv.imread, cv.Mat, cv.CascadeClassifier, cv.findContours) that are fully supported and performant in OpenCV.js?

  3. Practical Implementation:

    • Could you point me to any tutorials or example projects that demonstrate a similar workflow? Specifically, uploading an image from a web form, processing it with OpenCV (either on the server or client-side), and then displaying the results (e.g., an image with bounding boxes) back on the webpage.

    • What is the best way to handle the image data transfer? Should I convert images to base64 strings for the API, or use multipart/form-data for file uploads to the server?

  4. Performance & Scalability:

    • As this is intended to be a SaaS tool, are there any best practices for managing server load if I go the backend route? For example, using a task queue (like Celery with Redis) for heavy processing.

I am very excited about this project and its potential to add unique value to my website. Any code snippets, architectural advice, or links to relevant resources would be immensely helpful!

Thank you in advance for your time and expertise.