Industrial vision

Guidance

Firstly, machine vision systems can determine the position and orientation of a part. They compare them with a given tolerance and make sure they’re at the right angle to check for correct assembly. Secondly, guidance can be used to indicate the location and orientation of a part in 2D or 3D space to the control system of an industrial robot or special machine. This indication will enable the robot to locate the part, or the machine to align the part.

Identification

A machine vision system for part identification and recognition reads a lot of digital elements. These include barcodes (1D), Data Matrix codes (2D), DPM codes (direct marking) and characters printed on parts, labels and packaging. An optical character recognition (OCR) system reads alphanumeric characters without any prior knowledge. In addition, an Optical Character Verification (OCV) system confirms the presence of a character string.

Measurement

For measurement purposes, a machine vision system calculates the distance between two or more points, or between geometric locations on an object. It determines whether these measurements comply with specifications. If this is not the case, the machine vision system sends a failure signal to the control system. This triggers a rejection mechanism that ejects the object from the line.

A machine vision system consists of lighting, optics and machine vision processing.

Lighting is one of the keys to achieving satisfactory results in machine vision. Machine vision systems create images by analyzing the light reflected by an object, not by analyzing the object itself. The principle of lighting is to generate a light source and position it judiciously in relation to the room and the camera.

The optics acquire the image and transmit it to the camera’s image sensor. Optics vary in optical quality and price. Consequently, the optics selected determine the quality and resolution of the image captured. Most machine vision cameras feature two main types of optics: interchangeable and fixed. Interchangeable optics are generally fitted with C or CS mounts.

Processing is the process of extracting information from a digital image. It can be carried out externally in a PC-based system or internally in a stand-alone machine vision system. Processing is carried out by software and consists of several stages.

1/ An image is acquired using the sensor. In some cases, pre-processing may be required to optimize the image and ensure that all the necessary features stand out.

2/ The software locates the specific characteristics of the part, takes measurements and compares them with the specifications.

3/ A decision is made and the results communicated.

There are three main categories of machine vision systems.

1D machine vision analyzes a digital signal, one line at a time, rather than the entire image. For example, it evaluates the differences between a group of ten recently acquired lines and an older group. This technique generally detects and classifies defects in materials manufactured in a continuous process. These materials include paper, metal, plastic, foil and non-woven rolls.

Linear systems offer other advantages over matrix systems. For example, the inspection of round or cylindrical parts may require several matrix cameras to cover the entire surface of the part. However, by rotating the part in front of a single linear camera, the entire surface can be acquired and the image projected. Linear machine vision systems adapt more easily to smaller spaces. For example, when the camera has to pass between cylinders on a conveyor to see the lower part of a part.

3D machine vision systems generally comprise several cameras, or one or more laser displacement sensors. Multi-camera 3D vision in robot guidance applications shows the robot the orientation of the part. These machine vision systems consist of several cameras mounted at different locations and triangulation on an objective position in 3D space.