- Accuracy
- How close a measurement is to the true value; verified with certified references.
- Anomaly detection
- Deep-learning method trained mainly on good parts that flags anything that looks different.
- Area-scan camera
- Camera that captures a full 2D image in one exposure.
- Backlight
- Light placed behind the part so it appears as a dark silhouette.
- Bandpass filter
- Lens filter that passes only a narrow wavelength band, blocking ambient light.
- Bayer filter
- Color filter mosaic on color sensors; colors are interpolated, reducing effective resolution.
- Bit depth
- Number of gray levels per pixel (8 bit = 256, 12 bit = 4096).
- Blob analysis
- Tool that finds and measures connected regions of similar pixels.
- Bright field
- Lighting where specular reflection from flat areas enters the lens, making them bright.
- Calibration
- Relating pixels to real-world units and coordinates, and correcting lens distortion.
- Caliper / edge tool
- Tool that finds edge positions along a line, often with sub-pixel precision.
- CoaXPress
- High-bandwidth camera interface over coaxial cable, used with frame grabbers.
- Coaxial illumination
- Light sent along the camera axis via a beam splitter; flat reflective surfaces appear bright.
- Collimated light
- Light with parallel rays, used for precise silhouettes and with telecentric lenses.
- Dark field
- Low-angle lighting where flat areas stay dark and edges or defects scatter light into the lens.
- Deep learning
- Neural-network methods that learn to classify, detect or segment from labeled images.
- Depth of field (DOF)
- Distance range in front of and behind the focus plane that appears acceptably sharp.
- Diffuse lighting
- Light arriving from many directions (dome, flat dome) to suppress glare and texture.
- Distortion
- Geometric deviation of the image from the object, usually given in percent.
- DPM (direct part mark)
- Code marked directly on a part by dot-peen, laser or ink-jet.
- Dynamic range
- Ratio between the brightest and darkest signal a camera can capture.
- Encoder
- Sensor that reports motion; used to trigger cameras at constant spacing.
- Exposure time
- Time the sensor collects light for one image.
- F-number (f/N)
- Ratio of focal length to aperture diameter; higher numbers mean smaller apertures.
- Field of view (FOV)
- Area on the object captured by the camera.
- Focal length
- Optical property of a lens that, with the sensor size, sets the field of view.
- Frame grabber
- PCIe card that acquires images from high-bandwidth cameras.
- Gain
- Electronic amplification of the image signal; increases noise.
- Gauge R&R
- Study of how much of the tolerance is consumed by measurement variation.
- GenICam
- Generic standard for camera control, enabling software to work with many cameras.
- GigE Vision
- Camera interface standard over Gigabit Ethernet, up to 100 m cables.
- Global shutter
- Sensor that exposes all pixels at the same time — no distortion of moving parts.
- Image circle
- Diameter of the sharp image a lens projects; must cover the sensor diagonal.
- Jumbo frames
- Large Ethernet packets that reduce CPU load for GigE cameras.
- Laser triangulation
- 3D principle measuring height from the shift of a laser line seen at an angle.
- Line-scan camera
- Camera with one or a few pixel lines that builds images as the object moves.
- Lens mount
- Mechanical interface between lens and camera (C, CS, M42, F, …).
- lp/mm
- Line pairs per millimetre — a measure of lens resolution.
- Magnification
- Ratio between image size on the sensor and object size.
- Minimum object distance (MOD)
- Closest distance at which a lens can focus.
- Module (code)
- Smallest bar or square of a barcode or 2D code.
- Nyquist frequency
- Highest spatial frequency a sensor can sample: 1 line pair per 2 pixels.
- OCR / OCV
- Optical character recognition (reading text) / verification (checking expected text).
- Overdrive
- Driving LEDs above continuous current for very short, bright pulses.
- Pattern matching
- Tool that finds a trained pattern's position and orientation in an image.
- Pixel size
- Physical size of one sensor pixel, e.g. 3.45 µm.
- Point cloud
- Set of 3D points representing a surface.
- PoE
- Power over Ethernet — power and data on one cable.
- Polarizer
- Filter that passes light of one polarization; reduces glare.
- Quantum efficiency
- Share of photons converted into signal at a given wavelength.
- Quiet zone
- Empty margin required around a code for reliable decoding.
- Repeatability
- Spread of repeated measurements of the same object.
- Resolution
- Number of pixels of a sensor, or the smallest detail a system can resolve.
- Rolling shutter
- Sensor that exposes rows one after another; skews fast motion.
- ROI (region of interest)
- Part of the image that is read out or processed, increasing speed.
- Strobe
- Light switched on only during exposure.
- Structured light
- 3D method projecting patterns and measuring their deformation.
- Sub-pixel
- Edge or position calculated finer than one pixel from gray-level transitions.
- SWIR
- Short-wave infrared (about 900–1700 nm); sees moisture and through silicon.
- TDI
- Time delay integration — line-scan technique adding signal over several lines for sensitivity.
- Telecentric lens
- Lens with constant magnification over a depth range; used for measurement.
- Time-of-flight (ToF)
- 3D method measuring the travel time of light to each pixel.
- USB3 Vision
- Camera interface standard over USB 3, for short cable distances.
- Working distance (WD)
- Distance from the front of the lens to the object.
No matching terms.