Image-Free Target Classification With Semiactive Laser Detection System

Wang, Siyuan; Li, Lijing; Yu, Zijian; Sun, Mingjie

doi:10.1109/jsen.2022.3217281

Cited by 8 publications

(1 citation statement)

References 23 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…However, in theory, a complete reconstruction of an image with a resolution of N×N needs a minimum number of modulation patterns of N×N. Some researchers have also proposed algorithms to reduce the number of modulation patterns, such as compressive sensing algorithm [17][18], evolutionary compressive sensing algorithm [19], deep learning algorithm [20][21][22], etc. But these schemes are often accompanied by the sacrifice of imaging quality or great restrictions on application scenarios.…”

Section: Introductionmentioning

confidence: 99%

50k fps computational ghost imaging with an ultrafast LED array

Huang,

Li,

et al. 2023

Third International Computing Imaging Conference (CITA 2023)

View full text Add to dashboard Cite

Computational ghost imaging is a new photoelectric imaging technique mainly characterized by spatial light modulation and single-pixel detection. It provides advantages in multi-wavelength imaging, three-dimensional imaging, and other applications. Improving the image frame rate is always the focus of computational ghost imaging technology, but the image frame rate is mainly limited by the modulation speed of spatial light modulator. In this work, we developed a low-cost LED array module suitable for computational ghost imaging, which can provide high-speed structured illumination at a speed of 25 MHz. With a special scan display scheme, the 32×32 resolution LED array module can display Hadamard sub-patterns at a rate of 12.5 MHz, which is approximately 500 times faster than commercial DMD modulator. Combined with evolutionary compressive sensing algorithm, we obtained the image frame rate of 50k fps in the dynamic imaging experiment. Furthermore, we analyzed the light intensity error caused by the designed LED module and proposed a targeted algorithm to improve the imaging quality in this work. The proposed scheme greatly improves the real-time performance of computational ghost imaging technology and makes it more suitable for dynamic imaging applications. Besides, the LED array presented in this work can also potentially be applied to other technologies requiring high-speed structured illumination, such as LIFI and fringe projection profilometry.

show abstract