Embedded Files
Background
Image enhancing techniques such as structured illumination microscopy and subpixel resolution enhancement (SRE) are able to drastically increase the resolution of image sampling devices, allowing cameras to surpass the diffraction limit of light or simply increase sensor resolution. Thus far, these techniques have only been applied to static images with singular photos—this project aimed to apply them to high-speed imaging, or "slow-mo" video.
Image enhancing techniques such as structured illumination microscopy and subpixel resolution enhancement (SRE) are able to drastically increase the resolution of image sampling devices, allowing cameras to surpass the diffraction limit of light or simply increase sensor resolution. Thus far, these techniques have only been applied to static images with singular photos—this project aimed to apply them to high-speed imaging, or "slow-mo" video.
Professor Lisa Poulikakos, one sponsor of this project and professor at University of California, San Diego (UCSD) leads the Poulikakos Lab, researching and innovating on nanophotonics for global health – manipulating light at the nanoscale. She is interested in testing the limits of image enhancing techniques beyond the bounds of a microscopy, particularly in achieving higher resolution slow-mo fluid flow imaging.
Professor Lisa Poulikakos, one sponsor of this project and professor at University of California, San Diego (UCSD) leads the Poulikakos Lab, researching and innovating on nanophotonics for global health – manipulating light at the nanoscale. She is interested in testing the limits of image enhancing techniques beyond the bounds of a microscopy, particularly in achieving higher resolution slow-mo fluid flow imaging.
Professor Abhishek Saha sponsors this project as well and is a professor at UCSD, heading the Flame-Fluids-Energy Lab. He researches combustion, uses high-speed image sampling to resolve certain information about the combustion process. He is interested in using image enhancing techniques to increase the accuracy of the information obtained from image sampling.
Professor Abhishek Saha sponsors this project as well and is a professor at UCSD, heading the Flame-Fluids-Energy Lab. He researches combustion, uses high-speed image sampling to resolve certain information about the combustion process. He is interested in using image enhancing techniques to increase the accuracy of the information obtained from image sampling.
Hy Cao is a Master's student at UCSD jointly in the Poulikakos Lab and Flame-Fluids-Energy Lab, studying the application of super-resolution methods to fluid flow imaging.
Hy Cao is a Master's student at UCSD jointly in the Poulikakos Lab and Flame-Fluids-Energy Lab, studying the application of super-resolution methods to fluid flow imaging.
Project Objectives
High Priority Objectives
Move grating uniaxially
Variable speeds of 5-20 mm/s
Allow for control such that grating moves 5-20 μm between images
Position deviation within ±step size/2 (2.5-10 μm) from setpoint
> 5 seconds of imaging time at each 5-20 μm step size
Sync imaging with grating movement
Ability to send signal to camera to image every discrete 5-20 μm step
Ability to send signal to light to illuminate every discrete 5-20 μm step
Record position of grating with each image taken
Second Priority Objectives
Easily swappable gratings
UI for controlling the grating movement
Third Priority Objectives
Usable with frame-rates as high as 35 kHz
1 mm/s grating speed
Final_Design_Video_For_Website_Front_Page.movVideo of Final Design
Final Design
The final design consists of a ball screw, ball rail, motor, linear encoder system, grating housing, and flywheel. The ball screw, ball rail, and motor function together to move a grating housing uniaxially to be photographed. A 50 nm resolution optical linear encoder was integrated to provide precise position feedback. The encoder allowed for accurate measurements of the distance between each camera image as well as accurate PID control of the grating position. The electrical system needs to be revised to successfully send a rising-edge signal whenever the grating passes an increment of the step size; the system currently cannot send it at the 1000 Hz necessary frequency. From this signal, two methods for image recombination could be used. The camera could be signaled to image, or a light could be turned on at each step size with a light to illuminate the grating with independent camera imaging. Both methods would allow for successful image recombination.
Assembly of Key Components in Final Design
Performance Summary
Overall, the system was able to control the position deviation from setpoint within ±step size/2 boundaries, impressively achieving within ±2.5 µm at a step size of 5 µm. The system provided imaging time over 5 seconds for step sizes up to 15 µm, with step sizes closer and at 20 µm slightly below the desired 5 seconds of imaging time. The system was unable to send a signal every step size increment to trigger camera imaging or illuminate a light source. See the figure below for a graphical representation of the position deviation from setpoint.
Position deviation from position setpoint with 5-20 µm step sizes at 1000 Hz with ± (step size)/2 boundaries. Settling time for each step size shown in dashed vertical lines with their according color. PID constants used: KP=50,000, KI=20,000, KD=1,100.
Each trial in the figure met the required deviation boundaries. All trials besides the 20 µm trial met the required imaging and settling time of 5 seconds, where the 20 µm trial only allowed for 4.316 s of imaging. A summary of the these results is given in the table below.
Aggregated trial data with step sizes from 5-20 μm and 1000 Hz frame rate.
For > 98.95% of the time, the position deviation from the setpoint is within the required boundaries. For step sizes < 15 μm, the imaging time is > 5 seconds, with the trial at 20 μm step size only slightly under 5 seconds. This can still be used for imaging. Overall, the mechanism provides the functionality necessary to use subpixel resolution enhancement techniques with high-speed imaging, only requiring a revision to the electrical system to fully function as necessary.
Project Poster
MAE 156 - Final Poster.pptxPage updated
Report abuse