DJI Launches GPU-based High Performance Embedded Computer for Drones

Leading camera drone manufacturer DJI unveiled their new NVIDIA Tegra TK1-powered “Manifold” embedded computer this week. Now, developers can transform aerial…

Brad Nemire
1 min readintermediate
--
View Original

Overview

DJI has introduced the Manifold, a high-performance embedded computer powered by the NVIDIA Tegra K1, designed for developers to enhance drone capabilities with advanced computing and image processing. The device supports Ubuntu, CUDA, OpenCV, and ROS, making it suitable for AI applications such as computer vision and deep learning.

What You'll Learn

1

How to utilize the Manifold for advanced image processing in drones

2

Why the NVIDIA Tegra K1 is suitable for AI applications in drones

3

When to implement the DJI Onboard SDK for intelligent control

Prerequisites & Requirements

  • Familiarity with drone technology and embedded systems
  • Basic understanding of CUDA, OpenCV, and ROS(optional)

Key Questions Answered

What are the main features of the DJI Manifold embedded computer?
The DJI Manifold features an NVIDIA Tegra K1 quad-core ARM Cortex-A15 processor with 192 GPU cores, capable of achieving 326 gigaflops throughput. It runs on a built-in Ubuntu operating system and supports CUDA, OpenCV, and ROS, making it ideal for advanced image processing and AI applications.
How does the Manifold enhance drone capabilities?
The Manifold enhances drone capabilities by enabling complex computing tasks and advanced image processing in real-time, allowing drones to sense their environment, identify objects, and respond instantly, which is crucial for applications in AI and machine learning.
What is the processing power of the NVIDIA Tegra K1 used in the Manifold?
The NVIDIA Tegra K1 processor in the Manifold features a quad-core ARM Cortex-A15 architecture with clock speeds up to 2.2GHz and 192 GPU cores, achieving a throughput of 326 gigaflops, making it highly capable for demanding computational tasks.

Key Statistics & Figures

Throughput
326 gigaflops
Achieved by the NVIDIA Tegra K1 GPU cores in the Manifold.
Clock Speed
2.2GHz
Maximum clock speed of the NVIDIA Tegra K1 processor.

Technologies & Tools

Hardware
Nvidia Tegra K1
Serves as the processing unit for the Manifold embedded computer.
Software
Cuda
Used for parallel computing tasks on the Manifold.
Software
Opencv
Facilitates advanced image processing capabilities.
Software
Ros
Supports robotic applications and development on the Manifold.

Key Actionable Insights

1
Developers can leverage the Manifold's capabilities to create intelligent drone applications that perform real-time data analysis and object recognition.
This is particularly useful in fields such as agriculture, surveillance, and search and rescue, where drones need to process and respond to environmental data quickly.
2
Utilizing the built-in Ubuntu operating system allows developers to easily integrate existing software tools and libraries, enhancing development efficiency.
This flexibility is beneficial for rapid prototyping and testing of new drone functionalities, enabling faster innovation cycles.

Common Pitfalls

1
Overlooking the importance of understanding the underlying hardware capabilities when developing applications for the Manifold.
This can lead to suboptimal performance or failure to utilize the full potential of the Tegra K1 processor, which is crucial for demanding AI tasks.