Electromagnetic Imaging of Dielectric Objects Using a Multi-Directional Search-Based Simulated Annealing.

Area of Science:

• Electromagnetics
• Computational Physics
• Signal Processing

Background:

• Microwave imaging requires accurate methods to determine object electrical properties.
• Existing optimization techniques like simulated annealing can be slow and computationally intensive.
• Global optimization is crucial for reconstructing complex target characteristics.

Purpose of the Study:

• To introduce a novel global optimization method for microwave imaging.
• To enhance the performance and speed of simulated annealing using a multi-directional search.
• To reconstruct object shape, location, and material properties without a priori information.

Main Methods:

• A hybrid global optimization algorithm combining simulated annealing and multi-directional search.
• Utilizing a nonlinear simplex search for effective step-taking towards the global minimum.
• Application to a microwave-imaging system for electrical property retrieval.

Main Results:

• Significant improvements in the performance and speed of the simulated annealing method.
• Successful reconstruction of 2D images of multiple objects, including shape, location, and material properties.
• Demonstrated accuracy and applicability through numerical simulations.

Conclusions:

• The proposed global optimization method offers an efficient and accurate solution for microwave imaging.
• The hybrid approach overcomes limitations of traditional simulated annealing.
• This technique enables non-invasive characterization of objects using microwave imaging.