Image capture: simulation of sensor responses from hyperspectral images.

Area of Science:

• Computer Vision
• Image Processing
• Sensor Technology

Background:

• Digital cameras utilize sensors with varying spectral sensitivities.
• Understanding sensor response to irradiance is crucial for image quality.
• Existing models may not fully capture complex sensor interactions.

Purpose of the Study:

• To design and validate an image capture simulator.
• To model image capture devices with multiple sensor classes.
• To provide a tool for analyzing sensor design parameter effects.

Main Methods:

• Developed a simulator based on a general model of image capture devices.
• Input: narrow-band images from a hyperspectral camera.
• Parameters: sensor class number, spectral sensitivities, noise, quantization, spatial arrangement, exposure.

Main Results:

• The simulator outputs raw image data mimicking a target device.
• Validated against a calibrated Kodak DCS-200 digital camera.
• High agreement between simulated and acquired images confirmed model accuracy.

Conclusions:

• The image capture simulator accurately models sensor performance.
• The simulator is a valuable tool for digital camera design and analysis.
• Further research can explore advanced sensor characteristics within the simulator.