The client was developing a drone-based application capable of capturing 4K video at 60fps for inspecting indoor or hostile environments with minimal ambient light. However, they encountered visual artifacts in the video output and experienced delays when capturing snapshots—issues that impacted performance and reliability in real-time operations.
To address these challenges, we delivered a GStreamer-based, zero-copy video processing solution optimized for the NVIDIA® Jetson™ platform. This approach significantly improved image quality and reduced latency, enabling smoother and more responsive video capture in demanding environments.
SERVICES
INDUSTRY
Unmanned Vehicles
TECHNOLOGIES
GStreamer, codecs (audio and video),embedded devices
OUTCOME
High-quality, low-latency 4K drone video at 60fps
SERVICES
INDUSTRY
Unmanned Vehicles
TECHNOLOGIES
GStreamer, codecs (audio and video),embedded devices
OUTCOME
High-quality, low-latency 4K drone video at 60fps
Proposed solutions
High-quality drone video with optimized pipelines
Stabilizing real-time video capture in demanding environments
We worked with the client to resolve intermittent video freezes and visual artifacts that occurred during H.264 recording to the drone’s onboard memory. Our solution focused on optimizing the video encoding and data handling processes to ensure stable, high-quality recording performance under real-time operating conditions.
For further insights on how we tackle memory handling and advanced capture pipelines in GStreamer, explore our articles Don’t touch my memory and Capture the best shots and metadata with our GStreamer PoC. Both highlight techniques that complement the optimizations applied in this project.
Proposed solutions
High-quality drone video with optimized pipelines
Stabilizing real-time video capture in demanding environments
We worked with the client to resolve intermittent video freezes and visual artifacts that occurred during H.264 recording to the drone’s onboard memory. Our solution focused on optimizing the video encoding and data handling processes to ensure stable, high-quality recording performance under real-time operating conditions.
For further insights on how we tackle memory handling and advanced capture pipelines in GStreamer, explore our articles Don’t touch my memory and Capture the best shots and metadata with our GStreamer PoC. Both highlight techniques that complement the optimizations applied in this project.