Home / Technologies
The vision software and processing tools we deploy — selected according to inspection complexity, production speed, accuracy requirement, environmental conditions, and automation needs, since the smartest camera is only as good as the algorithm behind it.
For software specifically, this is where a captured image actually becomes a pass/fail decision, a measurement, or a decoded value. It’s also where the biggest either/or decision sits: rule-based tools are fast, deterministic, and easy to validate when the defect or feature is well-defined; deep learning earns its cost when the defect is too variable or subtle to describe with a fixed rule.
How well-defined the feature is — a fixed rule vs. a trained model.
Algorithm runtime has to fit inside the cycle time your line demands.
Sub-pixel measurement tools versus simpler presence or read checks.
How much part or lighting variation the software needs to tolerate.
How results need to reach the controller, PLC, or plant database.
Some of these tie directly into a product or application page — follow through for the full picture.
Full vision software suite for image analysis, measurement, and pattern matching on compatible hardware, configured per application.
Filtering, thresholding, and contrast enhancement that most other inspection tools depend on as a first step.
Locate, verify, or align a part against a stored template, independent of position or rotation.
Sub-pixel edge detection and dimensional gauging tools for length, diameter, gap, and position checks.
Read and verify printed or embossed text for date codes, lot numbers, and label content.
Decode 1D and 2D barcodes for traceability, sorting, and verification at line speed.
Trained models for defects too variable or subtle to describe with a fixed rule — used selectively, not as a default.
Full vision software suite
Runs on compatible smart cameras
Enhancement & pre-processing
Foundation for other tools
Presence, orientation, alignment
Rule-based, deterministic
Dimensional gauging
Sub-pixel edge detection
Text and print verification
Font and print-quality dependent
1D / 2D code decoding
Fast, deterministic
Variable or cosmetic defects
Needs training data, more tolerant of variation
Tell us the defect or feature you need to check and how variable it is. Our application engineers will recommend rule-based tools, deep learning, or a mix of both.