Jetson Xavier Developer Documentation

RidgeRun Knowledge Sharing

The Jetson AGX Xavier is a newly released SOM by NVIDIA. Introduced in 2018, the Xavier can achieve 20 times the performance and 10 times the energy efficiency of its predecessor, the Jetson TX2.

This SoM includes:

• Volta GPU with 512 cores
• 8-core ARM v8.2 CPU
• 16 GB LDDR4
• 2 NVDLA Deep Learning Accelerators
• 2 4Kp60 video codecs

RidgeRun knows how important documentation is for your project and how a single source of information may decrease the time-to-market of your product. That's why we've decided to share our expertise with the developer community, creating a centralized wiki with technical documentation, tutorials, examples and much more!

Enter one of the biggest sources of knowledge available for the new Jetson Xavier RidgeRun's: developer's wiki

i.MX8 Developer Documentation

RidgeRun Knowledge Sharing

The i.MX8 processor series is the latest addition to the i.MX SoC family from NXP Semiconductors. Taking advantage of a multi-core architecture, great perf/watt and heterogeneous processing, these highly integrated devices are well suited for the next generation of IoT devices, bringing up a stable, widely supported and scalable platform for your embedded products focused in advanced media processing, healthcare, automotive industry, domotics / home automation, and much more. The high end SoC in the family includes:

• 2 ARM Cortex A72
• 4 ARM Cortex A53
• 2 ARM Cortex M4F
• 1 DSP
• 2 GPU
• 4Kp60 decoding, 1080p60 encoding

RidgeRun knows how important documentation is for your project and how a single source of information may decrease the time-to-market of your product. That's why we've decided to share our expertise with the developer community, creating a centralized wiki with technical documentation, tutorials, examples and much more!

Get the most out of the i.MX8 processors by exploring our developer's wiki

LibGUVC v1.4.0

RidgeRun Product

We are glad to announce the release of libGUVC v1.4.0, now supporting NVIDIA Jetson platform family, thanks to the new bulk transfer support. You can see it in action on the video below.

It has never been easier to implement a UVC application on your hardware, libGUVC make it easy to interact with the UVC driver and expose a variety of useful features such as:

• USB 2.0 and USB 3.0 support
• Isochronous and bulk endpoint support
• YUV2, MJPEG and H264 video streaming support
• Extension Unit support
• MMAP and UserPtr support.

Since the driver is agnostic to the platform, you can run it on almost any platform with quality USB and UVC drivers, making it really simple to convert it into a UVC capable device.

See our developer's wiki for more information

Xavier IMX219 Driver

RidgeRun Product

We are glad to announce the support for the Sony IMX219 V4L2 Driver for the Jetson Xavier board. Xavier is one of the most powerful embedded platforms used for Computer Vision and Deep Learning and, on the other hand, the IMX219 sensor is one the most used cameras due to its size, simplicity and interesting features.

Resolutions:

• 3280x2464@15
• 1920x1080@30
• 1640x1232@30
• 1280x720@30
• 820x616@30

**Format: Bayer - RGGB10

The driver works with the following setup:

• Jetson Xavier board
• Carrier board (sold separately):
• J20  Auvidea board
• JetPack versions available: 4.1
• Capture elements: V4l2Src, NVarguscamerasrc

See our developer's wiki for more information

GstInference

RidgeRun Project

Deep Learning has revolutionized classic computer vision techniques to enable even more intelligent and autonomous systems. Multimedia frameworks, such as GStreamer, are a basic complement of automatic recognition and classification systems. GstInference is an ongoing open-source project from RidgeRun Engineering that allows easy integration of deep learning networks into your existing pipeline.

Checkout RidgeRun's presentation at the GStreamer's conference 2018:

Leverage GStreamer's flexibility and scalability with your existing models and achieve high performance inference. Use your pre-trained model from the most popular machine learning frameworks (Keras, Tensorflow, Caffe) to infer and execute them in a variety of platforms (x86, iMX8, TX1/TX2, Xavier). Link in your tracking, recognition and classification networks into your existing pipeline and achieve real-time deep learning inference.

Monitor the project progress in the project's GitHub repository:

See our developer's wiki for more information

GstColorTransfer v0.6.1  

RidgeRun Product

Are you developing a multi-camera system? You'd probably noticed that two images of the same scene will hardly ever have the same color distribution. For example, two cameras capturing a soccer or baseball field will have a different green tint.

GstColorTransfer is a GStreamer plug-in that performs high-level image post-processing to transfer the color distribution information from one image to another. Its CUDA accelerated algorithms allow for real-time processing in a GStreamer element that may be linked in your existing pipeline.

The latest version 0.6.1 features:

• Real time performance at 1080p30
• CUDA acceleration
• Support for Ubuntu 18.04
• User assisted overlap specification
• Supported platforms: Tegra X1/X2, Xavier and x86 (with NVidia GPU)

See our developer's wiki for more information

DispTEC

RidgeRun University Partnership

Modern advances on embedded systems architecture have bring complex computer vision solutions to the embedded world. Nowadays, it is possible to run amazing artificial intelligence algorithms as well as classic computer vision algorithms on such small systems providing a new spectrum of applications.

RidgeRun knows how important is to understand the system capabilities for computer vision in order to provide our customers an efficient solution according to their requirements, is for this reason that we have started working together with universities in order to explore the computational limits of NVIDIA based systems for non-traditional computer vision problems.

Check out some of our work with the academia at:

https://developer.ridgerun.com/wiki/index.php?title=RidgeRun/Academy_Research_Partnership_Program

Let us know if you have an amazing idea, we can make it happen!

GstCUDA v0.5.1

RidgeRun Product

New features and fixes were added to our latest v0.5.1 release.

GstCUDA is a framework and GStreamer plug-in enabling CUDA algorithm easy integration into GStreamer pipelines on the Jetson Tegra X1/X2 and Xavier boards. Check out GstCUDA in action in our promotional video below.

Version 0.5.1 includes:

• GStreamer single input / single output base class
• GStreamer multiple inputs / single output base class
• Support for multiple instances
• Unified memory GStreamer allocator
• Support for Jetpack 3.3

Make the most out of the Jetson TX1/TX2 and Xavier GPU architecture and GStreamer with performances up to dual 4K 60pfs data throughput.

See our developer's wiki for more information
Go to our online store.