How to make 10GigE machine vision reliable and cost-effective?

2024-01-15

Multiple visual system application scenarios, such as rapid detection lines, semiconductor factories, intelligent transportation systems, motion analysis, and volume capture, require high resolution, high FPS, and high data transmission rates to achieve better results. For vision system engineers who want to use machine vision cameras with faster frame rates and higher resolutions to improve output, upgrading from 1GigE to 10GigE is an obvious choice.

However, according to research by the Association for Automated Imaging (AIA), its adoption is quite slow. Considering the three technical challenges brought about by this upgrade: reliability (packet loss), high CPU usage, and high latency, this can be understood. This article introduces how the Teledyne FLIR Oryx+Myricom bundled solution addresses these challenges.

Update 1: Reliable Performance

Although the bandwidth of 10GigE Vision is 10 times higher than the GigE Vision protocol, the performance of the 10GigE host adapter has not been correspondingly improved. The data transfer from the camera to the host usually leads to CPU overload, causing application buffer overflow, and unacceptable packet loss levels for demanding applications.

By utilizing a host adapter to directly process packet reception and image reconstruction on the card, the CPU no longer needs to manage these tasks. The Teledyne FLIR Oryx+Myricom bundled solution is designed specifically to address such situations. As shown in our test results below, the system reliability can be significantly improved, thereby significantly reducing packet loss and thus reducing frame loss.

This bundled solution can be seamlessly integrated with our new custom SDK driver, specifically designed to handle data provided by Myricom cards. Through this combination, image data can be reliably transmitted from the camera to the host PC. The test results can be found in the appendix below: Reliability and CPU Usage Test.

The high cost-effectiveness of Teledyne FLIR Oryx+Myricom bundled solution makes it an obvious choice; Compared to separately purchasing hardware and then integrating, thisan economical and highly reliable setup.

Update 2: CPU usage can be managed

In theory, the CPU can process input data from 10GigE with up to 100% of one kernel, and can use multiple kernels when running multiple applications/cameras. By using Myricom cards to manage packet reception and image reconstruction, the CPU usage of each application can be as low as 1%, allowing more CPU cycles to be used for image processing. The test results can be found in the appendix below: Reliability and CPU Usage Test

Update 3: Delay reduction

The frame delay of 10GigE Vision is not fixed; This means that the arrival of frames may be accompanied by significant time base jitter. In some cases, especially for switches, there is not only packet loss, but also reverse order in frame reception. The Teledyne FLIR Oryx+Myricom bundled solution solves this problem by timely notifying frame completion to shorten latency and reduce time base jitter.

Appendix: Reliability and CPU Usage Testing

Test 1: High bandwidth 7-day streaming

Using a custom console application created through the Teledyne FLIR Spinnaker API, set up an 8.9 megapixel Teledyne FLIR Oryx camera to continuously capture images and track any incomplete images without additional processing or third-party resource intensive programs running simultaneously.

Test results: Approximately 40 million frames of images were collected; Detected 0 frames of incomplete/missing images.

Note: During the entire 7-day testing period, CPU usage was checked and found to remain at 1%. By disabling the new Myricom driver and relying solely on the FLIR standard filter driver, the CPU usage of the CPU core specifically designed for applications remains at approximately 100%.

Test 2: Dual camera streaming

This test includes two Oryx cameras (ORX-10G-123S6M and ORX-10G-89S6C) running in the same custom console application, each with a bandwidth of 6.7 Gb/s and running continuously for 24 hours.

Test results: Each camera captures approximately 6 million frames of images; Detected 0 frames of incomplete/missing images

Test 3: 24-hour CPU stress test

This test includes an Oryx camera (ORX-10G-123S6M) with the same settings as Test 1.

Used the same console application as Test 1, but this time used another application at the same time; This custom application aims to simulate high workloads with a total CPU usage of approximately 90% (for all eight cores).

Test results: Approximately 6 million frames of images were collected; Detected 0 frames of incomplete/missing images

Test system hardware and software specifications:

I7-9700k @ 3.6GHz | 16GB | Windows 10 1809

Teledyne FLIR Spinnaker 2.1.0.82 and PGRLwf 2.7.3.397 with customized version 2.3.0. x with Myricom support

Oryx ORX-10G-123S6M

Oryx ORX-10G-89S6C

The United States has developed technology that is millions of times faster than current PCs