Matrox debuts CoaXPress frame grabber, MIL enhancements at Vision 2011
November 11, 2011. At Vision 2011, Matrox Imaging announced the Matrox Radient eV-CXP, a CoaXPress (CXP) frame grabber derived from the original Matrox Radient family. The company also highlighted enhancements to the Matrox Imaging Library (MIL) 9 related to code reading and verification, GPU-based processing, and GenICam™ support.
The Matrox Radient eV-CXP provides four independent CXP links through the required BNC connectors. This allows for simultaneous capture from up to four cameras each running at different CXP speeds. For high-bandwidth applications, the frame grabber can also capture from a single camera transmitting image data at up to 25 Gbps using link aggregation.
To handle high data rates, the Matrox Radient eV-CXP has a PCIe 2.0 x8 host interface—with a peak transfer rate of up to 4 GB/s—combined with up to 4 GB SDRAM for on-board buffering. In cases with extremely high video frame rates (for example, thousands of frames per second), users have the option of relying on the Matrox Radient eV-CXP’s on-board MicroBlaze™ soft processor to offload image acquisition management from the host CPU. The frame grabber can also offload the host CPU from having to perform image preprocessing (for instance, Bayer interpolation, color space conversion, and LUT mapping) tasks.
Beyond host offloading, the Matrox Radient eV-CXP simplifies overall system integration by providing camera power, trigger, and control over each CXP link, as well as four independent sets of auxiliary I/O for interfacing with rotary encoders, photoelectric sensors and strobe controllers. By having the primary set of auxiliary I/Os on the same bracket as the BNC connections, the Matrox Radient eV-CXP offers a true single PCIe® slot solution for single camera applications.
The Matrox Radient eV-CXP is fully supported by the Matrox Imaging Library (MIL) toolkit on 32-bit and 64-bit Windows 7 and Linux.
Matrox Imaging Library Enhancements
Matrox Imaging also announced at Vision 2011 enhancements to the Matrox Imaging Library (MIL) 9, a software development toolkit for image analysis, machine vision, medical imaging, and video analytics. These enhancements include improvements to code reading and verification, GPU-based processing, and GenICam™ support.
“The ongoing implementation of product serialization to meet track and trace standards and regulations has brought about the continued development of the 1-D/2-D code reading and verification technology,” said Pierantonio Boriero, product line manager. The latest result is an update to MIL 9 that makes it easier to read distorted DataMatrix, QR, and 4-State/IMB codes, and allows Micro QR codes to be read. The update also introduces the verification of DataMatrix and QR codes based on the relevant ISO-IEC and AIM-DPM standards.
MIL’s GPU-based processing capability has been updated using the Microsoft®DirectX® 11, or more specifically, the DirectCompute application programming interface (API). “The latest version of this proven API has allowed Matrox Imaging to implement and optimize operations like histogram,” said Arnaud Lina, processing software development manager. Also illustrated in this update is the interoperability of MIL code with custom DirectCompute and OpenCL™ code for GPU-based processing. In addition, the update enables the use of a GPU that is not connected to a monitor, which is particularly useful for demanding applications requiring multiple GPUs.
MIL’s support for the GenICam™ CLProtocol will be introduced first for the Matrox Solios Camera Link® frame grabber family. “The GenICam™ CLProctocol is a new standardized method to access and control features of Camera Link® cameras and its use is being gradually phased in by leading camera vendors,” said Stephane Maurice, software development director.