1. Convert the perception task into an optical requirement
Lane markings, vehicles, pedestrians and signs occupy different numbers of pixels at different distances. Begin with the smallest feature that must remain usable, the required detection range and the algorithm’s contrast needs.
This connects optical performance with what the ADAS stack must perceive instead of using a generic “8MP” target.
- Target object and detection distance
- Minimum useful feature size
- Required field coverage
- Algorithm sensitivity to contrast and artifacts
2. Set MTF conditions that match the sensor
Pixel size determines sampling, but the lens, sensor and image-processing chain jointly determine system contrast. Specify review frequencies below and around relevant sensor sampling limits, and include center, mid-field and edge positions.
State whether curves are polychromatic or monochromatic, the wavelength weighting, f-number, focus condition and object distance. Without those conditions, two MTF values are not directly comparable.
- Spatial frequency and contrast target
- Center, mid-field and edge positions
- Wavelength weighting and f-number
- Nominal focus and object distance
3. Design for HDR scenes, not only laboratory charts
ADAS cameras encounter low sun, headlamps, tunnel exits, wet roads and high-contrast edges. Lens requirements should include representative ghost and flare source positions, intensity ratios and acceptance criteria.
Coating, element shape, barrel geometry, aperture surfaces and sensor cover stack can all influence stray light. Evaluation should include the complete optical path where possible.
- Low-sun and headlamp source positions
- Ghost location and intensity criteria
- Veiling glare / contrast loss
- Complete sensor-filter-lens stack
4. Protect focus and field performance across temperature
Refractive index, material dimensions, barrel length, adhesive and sensor position all change with temperature. A lens can remain visually in focus at the center while edge MTF or chromatic focus becomes unacceptable.
Define focus strategy, temperature points, stabilization conditions and evaluation fields before optimization. Camera-level thermal behavior should be reviewed with the mechanical and module teams.
- Operating and storage temperature
- Focus shift and field MTF
- Material CTE and refractive-index change
- Camera-module thermal stack
5. Freeze production controls with the optical specification
The best nominal design is not enough if element decenter, tilt, spacing, adhesive cure, barrel variation or sensor alignment can move production units outside the target. Use tolerance analysis to identify the sensitive contributors.
Agree on incoming inspection, assembly feedback, MTF or image-quality testing, traceability and change control before SOP. Sampling plans and limits must be aligned with the camera program’s risk level.
- Tolerance sensitivity and compensation
- Assembly and active-alignment method
- Image-quality end-of-line controls
- Traceability and engineering change control