New device generation: designed for AI and Deep Learning using user-generated apps
The limitations of using traditional sensors becomes evident when specialisation is needed and the sensor's capacity for flexibility is reached. Offering maximum flexibility for applications that are beyond the scope of standardised camera functions, the iDS NXT series is a new generation of industrial cameras based on the smart phone's app principle.

The introduction of a new generation of vision-app devices extends the usefulness of the old standard sensors by adding the capacity for the user to create and install their own vision app that works in tandem with the manufacturer's functions. Pre-processed data or results are easily transferred to a PC or machine control unit.

With the capacity of combining user-designed apps with the manufacturer's standard functions, and billed as the latest in customised application imaging, the iDS NXT models are fully-fledged industrial cameras with the added advantage of being multifunctional devices equipped with user-created vision apps and designed for AI and Deep Learning.

Manufacturers' strictly-defined functions are not very customisable. The NXT platform, however, is not tied to any predefined task but allows the user to develop and install vision apps for themselves, setting up and changing them at will. Reassuringly, app development does not need additional specialist knowledge or learning time.

Ideally equipped for numerous tasks in the industrial arena, using the standard manufacturer's functional scope in combination with the user- created-and-integrated app capability, the camera can be used in exactly the same way as if both platforms were part of the original product. Applications can be run classically on a PC workstation, independently on the IDS NXT camera, or collaboratively on both devices

Extending the limits of traditional sensors
Having a facility where users can integrate vision-compliant camera features they have created for themselves offers a number of benefits:

• The application processes can be scaled between camera and host PC, offering the option of partially or completely shifting the decision process from the host PC to the camera, providing:
• Lower network load, hardware-accelerated processing
• Lower energy consumption
• Lower system costs.
• Frequently-occuring tasks to be established and changed easily
• No longer just image suppliers but can also work completely independently, trigger subsequent processes directly or support a connected PC through individual pre-processing

How does the self-programmed app work?
If users want to deploy custom-written functions on smart cameras, they require a high-speed data interface and the ability to process data in a third-party application easily; conditions not usually met by conventional cameras/sensors.

The solution provided by the iDS NXT rio and rome models solves these problems by supplying Smart GeniCam functionality which allows the user to upload their vision app. Firmware then integrates the app code into the camera's selection of functions and all are made available in standardised form. Third party applications, such as HALCON, then use these features as if they were part of the standard functionality of the industrial camera.

AI-ready (rio and rome models)
Designed for artificial intelligence and deep learning, the NXT models have the capacity to learn more about what they are doing, with time, and therefore, make better decisions. The NXT models carry an especially-developed AI vision app which makes them into universal, high-performance inference cameras with an integrated artificial neural network accelerator having inference times of a few milliseconds.

The NXT rio and rome models have the capacity to learn more about what they are doing, with time, and therefore, to make better decisions.

Applications
These cameras expand application possibilities in numerous fields, notably optical quality assurance; the cameras are suitable as an analytical device in medical technology, in face-recognition tasks or in vehicle/people-count applications.

Designed for artificial intelligence and deep learning,

Currently, three models are offered:
• NXT Vegas - an embedded vision system with a light-sensitive 1.3 MP CMOS sensor, a liquid lens with autofocus and an integrated Time-of-flight sensor
• NXT rio and • NXT rome - intelligent cameras