Improving the Efficiency of Backprojection and Vehicle Detection in Synthetic Aperture Radar images

School Name

South Carolina Governor's School for Science and Mathematics

Grade Level

12th Grade

Presentation Topic

Computer Science

Presentation Type

Mentored

Abstract

In the modern world, militaries use radars to form images instead of optical images to compensate for the many cons of optical imaging through the use of Synthetic Aperture Radar. SAR works by collecting data over a region of interest and using this data to form an image through an algorithm called backprojection. My research at MIT Beaver Works Summer Institute was to work on methods to make communication with radars and backprojection more efficient. We did this by using interpolation to form more points on a graph of signal amplitudes gathered by the radar. We also used multithreading to utilize a heavier load of a computer's CPU to make the code run faster. We implemented many error catches on the command and control code that helped prevent any errors that could occur when collecting data from the radar. My team developed a command and control code that was very close to entirely error-free. Our backprojection code was the fastest out of the other teams and on par with the instructor. We learned that interpolation and multithreading are very important in improving code and making it faster. Our first implementation of the code took over 45 minutes for a high-resolution image of 1200x1200 pixels over 200 scans. In our final implementation, it took less than 9 seconds to do the same.

Location

HSS 209

Start Date

4-2-2022 9:45 AM

Presentation Format

Oral Only

Group Project

No

COinS
 
Apr 2nd, 9:45 AM

Improving the Efficiency of Backprojection and Vehicle Detection in Synthetic Aperture Radar images

HSS 209

In the modern world, militaries use radars to form images instead of optical images to compensate for the many cons of optical imaging through the use of Synthetic Aperture Radar. SAR works by collecting data over a region of interest and using this data to form an image through an algorithm called backprojection. My research at MIT Beaver Works Summer Institute was to work on methods to make communication with radars and backprojection more efficient. We did this by using interpolation to form more points on a graph of signal amplitudes gathered by the radar. We also used multithreading to utilize a heavier load of a computer's CPU to make the code run faster. We implemented many error catches on the command and control code that helped prevent any errors that could occur when collecting data from the radar. My team developed a command and control code that was very close to entirely error-free. Our backprojection code was the fastest out of the other teams and on par with the instructor. We learned that interpolation and multithreading are very important in improving code and making it faster. Our first implementation of the code took over 45 minutes for a high-resolution image of 1200x1200 pixels over 200 scans. In our final implementation, it took less than 9 seconds to do the same.