FPGA-Based On-Board Geometric Calibration for Linear CCD Array Sensors.

Area of Science:

• Remote Sensing
• Computer Engineering

Background:

• Real-time remotely sensed imagery is crucial for military, disaster response, and security applications.
• On-board geometric calibration of linear CCD sensor arrays is a significant challenge.

Purpose of the Study:

• To develop and validate an efficient on-board geometric calibration method for linear CCD sensor arrays using FPGA chips.
• To optimize FPGA resource utilization and calibration accuracy.

Main Methods:

• Implementation of parallel computations for observation equations and least squares adjustment using FPGA chips.
• Utilizing a decomposed matrix inversion method on a Xilinx Virtex-7 FPGA.
• Validation using MOMS-2P data from DLR.

Main Results:

• A 50-bit data width is recommended for FPGA-based geometric calibration to balance resource consumption and performance.
• Six ground control points (GCPs) are sufficient and recommended for calibration without significant resource increase.
• The FPGA-based method achieves approximately 24 times the speed of PC-based methods.
• Calibration accuracy is comparable to inflight calibration when using the same model and GCPs.

Conclusions:

• FPGA-based on-board geometric calibration offers significant speed advantages for real-time applications.
• The proposed method provides a practical and efficient solution for geometric calibration of linear CCD sensors.
• Optimized data width and GCP selection are key to efficient FPGA implementation.