Line scan cameras

A line scan camera captures individual sensor lines continuously. The relative movement of the material assembles those lines into a two-dimensional image.

Particularly with: continuous web material very wide objects continuous production very high resolution rotating cylinders web inspection

These indicate roughly the number of sensor pixels along one line. An 8K line scan camera has approximately 8,000 pixels across the scan direction, for example. Basler offers current line scan systems up to 16K.

A color line scan camera produces color information from several color-sensitive pixel lines or other multispectral sensor concepts. Because the color channels can be captured with a spatial or temporal offset, motion synchronization is particularly important.

A bilinear line sensor uses two sensor lines. Depending on the sensor design, these can capture different color or brightness information, for example.

A trilinear color sensor typically has separate sensor lines for red, green and blue. Because these lines are physically offset, the color information has to be spatially corrected to match the object movement.

A multispectral camera captures several spectral bands. Visible color information and NIR can be combined, for example.

object resolution across the direction of travel = field of view / number of pixels Example: FOV = 500 mm sensor = 4,096 pixels 500 / 4,096 = 0.1221 mm/pixel

For roughly square object pixels, the distance traveled between two captured lines should correspond approximately to the horizontal object resolution. line rate = object speed / desired line spacing Example: speed = 1,000 mm/s desired line spacing = 0.122 mm 1,000 / 0.122 ≈ 8,197 lines/s so roughly: 8.2 kHz

If the line rate is too low or too high in relation to the movement, the image is stretched or compressed in the direction of travel. The geometry is then rendered incorrectly.

If the conveyor speed varies, a constant time-based line rate is not enough. An encoder measures the actual movement. Line capture can then be triggered in proportion to the distance actually traveled.

A quadrature encoder typically provides two phase-shifted signals, A and B. This allows both movement and direction of movement to be evaluated.

A divider reduces the number of encoder pulses used for image capture. A multiplier or rate converter computes a finer or adjusted trigger frequency. Frame grabbers such as the Coaxlink series offer programmable rate converter functions for this.

Pixel aspect ratio describes the ratio of the effective pixel size in the X and Y directions in the object image produced. With line scan, the X resolution is determined optically by the sensor and FOV, the Y resolution by speed and line rate. For geometrically correct imaging, a ratio of roughly 1:1 is usually the aim.

TDI means time delay integration. Several sensor stages integrate the signal of a moving object in sync with the movement. This collects more light information and increases sensitivity.

Not automatically. TDI primarily increases the usable light sensitivity and signal level at high speed. Basler explicitly points out that TDI does not automatically increase the line rate.

High-intensity line lighting is normally used. As the material speed and line rate increase, the available exposure time decreases - which normally increases the luminous flux required.

The line rate does not match the material speed. Remedy: calculate the correct line rate use an encoder set the rate converter check the pixel aspect ratio

mm/pixel = scan width / number of line pixels

line rate [Hz] = speed [mm/s] / line spacing [mm]

data rate = pixels per line × lines/s × bits/pixel / 8

Speed, line rate and encoder have to be matched to one another.