Embedded vision and 3D sensors: what it means to be smart

Fixed Sensor Design and the Machine Vision Status Quo

Today’s simple sensors (including ultrasonic, capacitive, inductive, temperature, accelerometer and more) are built with a single purpose to measure and deliver a specific output. In the world of machine vision, the basic sensor is the camera chip for which manufacturers add sufficient hardware to control and deliver raw images from the camera to a PC. Developers then add other machine vision elements such as optics, lighting, a PC, and 3rd party image processing to solve real world problems.

How Embedded Processing Makes Hardware “Smart”

By adding an embedded processor, memory, and flash to a fixed hardware design, it is possible to add new functionality and extend the use of a low level, single purpose sensor into a broader market— making the hardware “smart” in the process. This idea is not new. Embedded processing drove the development of 2D smart cameras, however, the implementation of “smart” has fallen short of many user needs. This paper presents ideas that will drive the smart hardware industry and evolve user experience to the next level.