以單像素成像系統定位及辨識反射式標記

Localization and Identification of Reflective Markers with Single-Pixel Camera

游子緒(Tzu-Hsu Yu)

指導教授 ： 蔡欣穆

國立臺灣大學/電機資訊學院/資訊工程學研究所/碩士(2020年)

https://doi.org/10.6342/NTU202000915

摘要

基於相機的可見光定位系統有容易受環境光干擾的缺點。本研究提出一基於單像素成像系統的定位系統。此系統可利用頻率選擇性從干擾中分離出調變的光訊號，而一般相機因為其幀率太低無法利用頻率選擇性。另外，此系統亦可利用空間選擇性。基於 retroreflector 的反射式標記可集中反射光至入射光的來向，使得環境中其他物體的反射光相對黯淡。此定位系統中的單像素成像系統由一數位微鏡裝置及一光電探測器構成，並利用壓縮感知重建拍攝的場景。在壓縮感知重建的過程中加入非負的限制可以大幅增強重建影像的品質，但一般的重建演算法需要非常長的重建時間。本研究提出一改進版本的 in-crowd 演算法以加速重建過程，實現實時的定位。

關鍵字

單像素成像系統 ； 壓縮感知 ； 數位微鏡裝置 ； 數位訊號處理 ； 可見光定位

並列摘要

Visible light positioning systems based on cameras suffer from interference from ambient light. A positioning system based on a single-pixel camera is implemented in this work. The system can utilize selectiveness in frequency to extract modulated light signals from interference, which is hard for cameras due to their low frame rate. Furthermore, selectiveness in space can also be utilized. Reflective markers are made of retroreflector, which concentrates reflection from markers, making reflection from the surrounding environment relatively dimmed. The single-pixel camera consists of a Digital Micromirror Device (DMD) and a photodetector, and uses Compressed Sensing (CS) to reconstruct the captured scene. Adding nonnegative constraints in CS improves reconstruction quality significantly, but also slows down the common algorithms. A modified version of the in-crowd algorithm is used to accelerate the nonnegative-constrained CS problem, enabling realtime positioning of markers.

並列關鍵字

Single-Pixel Camera ； Compressed Sensing ； Digital Micromirror Device ； Digital Signal Processing ； Visible Light Positioning

參考文獻

[1] E. J. Candes and M. B. Wakin. An introduction to compressive sampling. IEEE Signal Processing Magazine, 25(2):21–30, 2008.

[2] P. R. Gill, A. Wang, and A. Molnar. The in-crowd algorithm for fast basis pursuit denoising. IEEE Transactions on Signal Processing, 59(10):4595–4605, 2011.

[3] Roshan Ayyalasomayajula, Aditya Arun, Chenfeng Wu, Sanatan Sharma, Abhishek Sethi, Deepak Vasisht, and Dinesh Bharadia. Deep learning based wireless localization for indoor navigation. pages 1–14, 2020.

[4] Jun Qi and Guo-Ping Liu. A robust high-accuracy ultrasound indoor positioning system based on a wireless sensor network. Sensors, 17:2554, 2017.

[5] Yu-Lin Wei, Chang-Jung Huang, Hsin-Mu Tsai, and Kate Ching-Ju Lin. Celli: Indoor positioning using polarized sweeping light beams. pages 136–147, 2017.