🎉Celebrating 25 Years of Tech Excellence and Trust - Learn More

Video Stream Processing

Image for OPENCV FFMPEG

Radixweb was associated with an Italy-based client, who was akin to find a partner to help them with video and image processing solution for their existing system.

The Problem

The client had two input video streams running on RTMP/HTTP video streaming server.

First, live video streaming had text code (ex. 000321FG4) masked on the video at any time and for any time duration.

Second stream is a simple black stream with same code masked on it at same position, same time and for same duration as live video stream.

The main objective of this project was to remove the masked code along with the code text detection using OCR in the live streaming video, re-encoding and re-streaming the resultant video on an RTMP server.

The Solution

We created a Linux based C++ application that took various inputs in command line.

  • Iterate each frame in black input stream using FFmpeg library.
  • We work on LAMP configuration for development and the Linux Live CD provided us with the much need SDK that serve all its needs well.
  • For each frame, we processed them using OpenCV to detect text code in the frame image. We applied OCR on detected code to verify that the correct code is processed. We applied algorithms and configurations to the OpenCV provided OCR to improve the results.
  • When the code is detected, we created a blurred mask with the help of surrounding background for code and overlaid the mask on actual live streaming input wide. The team also applied different processing using OpenCV to blur the code area so that the code cannot be read in the video.
  • After the code is processed and removed from live video, we re-encoded frames into H.264 and streamed them live on RTMP server as final output video using FFmpeg. Syncing audio stream from input live stream with processed video stream was the crucial requirement to be fulfilled.
  • We re-encoded the output stream to different video resolution based on an input provided to the application.

The whole application along with final package, source code and documentation was delivered.

Highlights

  • Radix developed effective code finding an OCR algorithm which runs easily on different kind of video inputs.
  • The processing and algorithm for removal of the code is efficient so that user cannot notice the code in video.
  • The solution was so configurable that it can be used with various different processing parameters like bits per second, output video quality, resolution, etc.
  • The application is processing and memory efficient so that it can used on a single stream for longer duration.
Get in Touch

Let’s discuss your project and build something extraordinary

Great place to work
Top Rated Software Development Companies - Radixweb
Goodfirms - Top Software development company - Radixweb
IAOP 2024 The Global Outsourcing 100
Techviewer.co Top Software Developers
ITFirms Top Web Development Company
Top 10 Custom Software Development
Clutch - Top Software Development Company
Great place to work
Top Rated Software Development Companies - Radixweb
Goodfirms - Top Software development company - Radixweb
IAOP 2024 The Global Outsourcing 100
Techviewer.co Top Software Developers
ITFirms Top Web Development Company
Top 10 Custom Software Development
Clutch - Top Software Development Company
Great place to work
Top Rated Software Development Companies - Radixweb
Goodfirms - Top Software development company - Radixweb
IAOP 2024 The Global Outsourcing 100
Techviewer.co Top Software Developers
ITFirms Top Web Development Company
Top 10 Custom Software Development
Clutch - Top Software Development Company
Great place to work
Top Rated Software Development Companies - Radixweb
Goodfirms - Top Software development company - Radixweb
IAOP 2024 The Global Outsourcing 100
Techviewer.co Top Software Developers
ITFirms Top Web Development Company
Top 10 Custom Software Development
Clutch - Top Software Development Company