Digital Domain TDI-CMOS Imaging Based on Minimum Search Domain Alignment.

Area of Science:

• Digital imaging and signal processing
• CMOS sensor technology
• Dynamic imaging systems

Background:

• Accurate image alignment is crucial for dynamic imaging, especially with Time Delay and Integration (TDI) CMOS sensors.
• Low-contrast and dark imaging conditions present significant challenges for feature matching and motion estimation.
• Image motion and platform jitter degrade the quality of captured dynamic imagery.

Purpose of the Study:

• To develop a robust digital domain TDI-CMOS dynamic imaging method.
• To enhance the alignment speed and accuracy of images acquired under challenging low-contrast and dark conditions.
• To mitigate the impact of image motion and spatial camera jitter on imaging quality.

Main Methods:

• A five-step digital domain TDI-CMOS dynamic imaging method based on minimum search domain alignment.
• Optimization of search box size and position using image motion compensation and satellite platform jitter models.
• Feature pair matching within optimized search boxes, followed by kernel density estimation for displacement probability.
• Digital domain alignment and superposition of frames for TDI accumulation.

Main Results:

• Greatly improved alignment speed and accuracy for dark and low-contrast images.
• Effective mitigation of image motion and spatial camera jitter effects.
• Fitted global image motion error maintained below 0.01 pixels.
• Improved Modulation Transfer Function (MTF) coefficient to 0.68.

Conclusions:

• The proposed digital domain TDI-CMOS method significantly enhances dynamic imaging performance.
• The method successfully addresses challenges in aligning low-contrast and dark images.
• The technique effectively reduces motion and jitter artifacts, leading to superior imaging quality.