Vision sensors & intelligent cameras
In industrial automation technology there has been a trend towards increased use of camera inspection systems for years. In the 1990s, almost only PC-based inspection systems fulfilled these tasks. The enhanced computing power of CPUs and DSP makes it possible to do without a PC system for many mass applications.
Intelligent cameras, vision sensors and code readers conquer more and more fields of application which could only be fulfilled by a PC some years ago. Resolution and evaluation speed are continuously increased.
Particularly compact versions are referred to as vision sensors, with optics and lighting usually already being integral parts of the system. The software can quickly be set up even by non-professionals without special knowledge of machine vision: the fields of application are mainly restricted to simple identification tasks, presence checks or simple measuring applications. The configuration is partly done via small operation and control units (in a control cabinet or as a compact pocket interface) or using simple PC software. Simple models can be acquired for clearly less than 1000 euros.
The passage from vision sensors to intelligent cameras is fluent and a distinction is not always easy. Intelligent cameras, like vision sensors, have the complete image acquisition and evaluation electronics integrated in the housing. Normally, however, optics and lighting are not built in and must be selected by the users themselves. This results in the same possibilities like when using a conventional PC-based camera system. Except for few completely freely programmable smart cameras, the devices normally have a supplied software environment. The user mainly accesses the device via the Ethernet interface (also during operation) and creates an application by means of a graphic user interface. Everything is available on the market: from very simple software packages to complex image processing programmes with a wide range of functions (often comparable to PC systems) and scripting possibilities. Today the camera resolution ranges up to 5 megapixels, even intelligent line scan camera systems are available on the market.
Advantages & disadvantages
An advantage is the smaller overall dimensions of all components as a whole. By melting together image capture, digitalisation and evaluation into a small module, designing the system has become much easier for the integrator and user. Most of the time a quite user-friendly software interface is supplied which essentially simplifies the use for many standard applications. Due to the compact design, many vision sensors and smart cameras have high IP protection classes (IP67, etc.) and can also be used in rough environments.
A possible disadvantage is the clearly restricted selection of available camera versions with different resolutions. The computing power is limited compared to PC systems due to the compact format. If optics and lighting are integrated into the system and cannot be replaced by standard components, further functional restrictions must be taken into consideration. Typically the system is bound to a special software package provided by the manufacturer.
Intelligent embedded PC cameras try to compensate this disadvantage by offering a complete Windows environment for the classic (proprietary) software. Yet, compared to DSP-based intelligent cameras, they are clearly larger.
A particularly important argument is (as usual) the price. Individual, compact sensors and smart cameras are usually clearly cheaper than a PC system. Only in case of higher resolutions, very fast evaluations and multi-camera solutions, PC systems are advantageous. The supplied software programmes are sufficient for many test tasks and are easy to learn.