Knowledge · DPM
With Direct Part Marking, the code is applied to the part. Readability depends on the mark, surface, presentation and light reaching the camera. Decoding comes after image formation.

Definition
A DPM code may be laser marked, dot-peened, etched or inked onto metal, plastic or a coating. Each process changes the surface differently. Laser marking may alter colour or gloss; dot-peen creates indentations; ink adds material. Wear, cleaning and later processing affect each mark differently, so required durability must be tested for the actual part and process.
Why it is hard
A decoder can tolerate some missing modules or distortion, but it cannot recover information that is not visible. The optical setup must preserve separation between code cells and background across the allowed part variation.
Directional light can reveal height changes and scratch direction. Diffuse light suppresses local glare. The useful geometry depends on the combination of mark and surface; increasing intensity alone does not solve it.
On cylinders and cast parts, viewing angle, focus and apparent cell shape vary within one image. Test the full permitted range of position and orientation, not only one centred sample.
For moving parts, exposure time, light pulse and trigger position need enough margin against motion blur. Speed and position variation belong in the test conditions.
Method
Reliable reading under agreed production conditions is different from formal code verification or grading to a standard. For a read test, document the sample set, speed range, positions, contamination and settings. Count successful reads, no-reads and incorrect results separately. Repeat boundary samples and test the PLC response, reject action and manual recovery route. Formal grading requires a suitable verifier, calibration and a defined method.
Next steps
Compare a dedicated reader, PC-based vision and line-scan reading by task, integration and ownership.
Discuss parts, production variation, speed, available space and the required response to a no-read.
Review reader families after the optical constraints and communication needs are defined.
Use an application example only when the mark, test conditions and outcome are documented.