An embedded vision standard can be built on earlier standardization experiences.
In many technologies, standardization is the technological gateway to a barrier-free world. Standards simplify the adoption and integration of a technology; they reduce learning costs for users as well as development costs for suppliers. For machine vision as a still relatively young technology, the global, joint standardization efforts of the past 15 years have been key to adoption. Since 2009, international machine vision associations have been working under the umbrella of the so-called G3 Initiative to advance existing standardization initiatives and explore new fields of standardization both within machine vision and in cooperation with adjacent vertical industries. The European Machine Vision Association (EMVA) is one of the five G3 partners.
Standardization cannot be taken for granted
It is important to understand the mechanisms for how standardization comes about. The associations, as the hosting organizations of a standard, primarily provide the framework. The actual work is done within the standard working groups, whose members are usually seconded by companies that have an interest in the development of the respective standard, such as to integrate it into their products. As a result, companies that would otherwise be competing in the market cooperate within the working group in the development of the standard. Further members of the working groups represent the scientific and research communities. All working groups are basically open for cooperation. For the G3 effort, therefore, anyone who would like to contribute is invited to contact the relevant working group or the EMVA as one of the hosting associations.
EMVA 1288 and GenICam for image processing
The EMVA currently hosts four machine vision standards. One is EMVA 1288, which has become a globally established standard for describing and comparing the image quality of cameras. Selecting the appropriate camera for a particular image processing application often proves to be a challenging task. The data sheets provided by manufacturers can be difficult or even impossible to compare. In addition, important information is not always available, leaving users little option other than to perform a costly comparison test. This is where the EMVA 1288 standard comes into play. It defines a uniform method for measuring, calculating, and presenting specification parameters for cameras and image sensors, thereby creating transparency for comparing them.
Equally fundamental to the machine vision industry is the Generic Interface for Cameras, better known as GenICam. Today’s digital cameras are packed with a multitude of functionalities. As a result, the complexity of the programming interface for cameras has continually increased. In this context, the GenICam standard has become the worldwide backbone for plug-and-play operation of cameras and devices in machine vision, with all other established machine vision hardware interface standards — such as CameraLink, CameraLink HS, CoaXPress, GigE Vision, and USB3 Vision — building on its discovery, control, and stream services. GenICam provides a generic programming interface for all types of devices and applications. Regardless of the interface technology used or the functions implemented, the application programming interface (API) is always identical.
These two EMVA-hosted standards have been adopted internationally since they ensure that buyers of machine vision cameras or software get the image quality of their choice with comparable components selected on the basis of EMVA 1288 criteria, and that the image capture for the application is easily and reliably working by means of GenICam.
New standardization initiatives and cooperation
New standardization projects are reviewed for their necessity within the G3 by all participating associations. One of the recent EMVA initiatives, recognized by the G3, builds on the existing EMVA 1288 and GenICam guidelines and attempts to control the interaction of optics/shutter/mechanics around the camera so that all lens conditions can also be set and controlled in practice. The EMVA Standard Group Open Optics Camera Interface (OOCI) therefore aims to create a common standard — closely tied to GenICam —for lens control, even if the optical elements have different mechanical and electrical connections such as RS-232/I2C/SPI.
The Embedded Vision Interface Standard, or emVision, is another EMVA standardization initiative. The combination of a processing board with a small camera makes it possible to design a very compact vision system. These embedded vision systems are of great interest to the machine vision industry and many others as well. Most embedded vision applications run on Linux with MIPI cameras and APIs such as V4L2, GStreamer, and further developments like libCamera.
These user interfaces, however, do not provide sufficient support to address problems holistically in the machine vision environment. Even if custom extensions are provided in V4L2, for example, from a user perspective these lead to incompatible solutions across manufacturers as implementations.
Because of the diversity in the camera hardware interface, camera API, and possible extensions, one individual driver must be programmed for each camera module that is connected to a system-on-chip. This effort is added to the number of MIPI modules offered and multiplied by the number of SoC suppliers on the market — an almost incomprehensible expenditure of programming resources!
A major part of these programming costs could be avoided through joint efforts for a generic user interface, which would lead to lower development costs for manufacturers and customers in the embedded arena. The emVision effort is intended to meet those needs.
In response to industry requests, the EMVA and The Khronos Group joined forced in February 2021 to form an Embedded Camera API Exploratory Group, open to all at no cost, to explore industry interest in the creation of open API standards for controlling embedded cameras and sensors. All participants — currently, 60 companies, with 145 representatives — are able to discuss use cases and requirements for new interoperability standards to accelerate market growth and reduce development costs in embedded markets using vision and sensor processing and associated acceleration. All sensor and camera manufacturers, silicon vendors, and software developers working on vision and sensor processing can participate in this initiative. If the exploratory group reaches significant consensus, then EMVA and Khronos will work together to initiate the proposed standardization initiatives at the appropriate organization.
Standardization should benefit everyone
The emVision standard and the Embedded Camera API Exploratory Group particularly exemplify how the need for a new standard is formed by the industry and further developed in a joint effort. Clearly, standardization will remain a highly dynamic process. All working groups of the machine vision standards are therefore open to collaboration with other companies. By sending employees to a working group, companies can contribute their own preferences and ideas to a common specification of a common challenge of the industry — an opportunity called standards development.
For the EMVA as a hosting association, together with the other G3 partners, the task is to counteract possible market foreclosures in order to continue to live up to the idea of an open and fair market that is inherent in standardization. However, that does not preclude the establishment of royalty or licensing models for the use of an established standard.
This article was originally published on EE Times Europe.