Multi-view 3D reconstruction, namely, structure-from-motion followed by multi-view stereo, is a fundamental component of 3D computer vision. In general, multi-view 3D reconstruction suffers from an unknown scale ambiguity unless a reference object of known size is present in the scene. In this article, we show that multi-view images captured using a dual-pixel (DP) sensor can automatically resolve the scale ambiguity, without requiring a reference object or prior calibration. Specifically, the defocus blur observed in DP images provides sufficient information to determine the absolute scale when paired with depth maps (up to scale) recovered from multi-view 3D reconstruction. Based on this observation, we develop a simple yet effective linear method to estimate the absolute scale, followed by the intensity-based optimization stage that aligns the left and right DP images by shifting them back toward each other using cross-view blur kernels. Experiments demonstrate the effectiveness of the proposed approach across diverse scenes captured with different cameras and lenses. Code and data are available at https://github.com/lilika-makabe/dp-sfm-tpami.git

Multi-view 3D reconstruction, namely structure-frommotion and multi-view stereo, is an essential component in 3D computer vision. In general, multi-view 3D reconstruction suffers from unknown scale ambiguity unless a reference object of known size is recorded together with the scene, or the camera poses are pre-calibrated. In this paper, we show that multi-view images recorded by a dual-pixel (DP) sensor allow us to automatically resolve the scale ambiguity without requiring a reference object or pre-calibration. Specifically, the observed defocus blurs in DP images provide sufficient information for determining the scale when paired together with the depth maps (up to scale) recovered from the multi-view 3D reconstruction. Based on this observation, we develop a simple yet effective linear solution method to determine the absolute scale in multi-view 3D reconstruction. Experiments demonstrate the effectiveness of the proposed method with diverse scenes recorded with different cameras/lenses. Code and data are available at https://github.com/kohei-ashida/dp-sfm.

We investigated the influence of surface preparation procedures on the resonance characteristics of silicon nanocavities. We demonstrate that ultrahigh-Q factors larger than one million can be maintained when applying preparation procedures that support attaching nanomaterials.

We employ a super-luminescent diode (SLD)as excitation light source for photonic crystal high-Q nanocavities. We demonstrate that the broad emission of the SLD enables robust and simultaneous excitation of eight high-Q L3-type nanocavities.

Electron-beam lithography allows precise photonic crystal (PC) fabrication, but is not suitable for mass production. Therefore, the development and optimization of CMOS-compatible processes is necessary to implement unique nanocavity technologies in optoelectronic circuits. We investigated the quality factors (Q) and the resonant wavelengths (lambda) of PC heterostructure nanocavities fabricated by the 193-nm argon fluoride immersion lithography on a 300-mm-wide silicon-on-insulator wafer. We measured 30 cavities distributed over nine chips at various positions of the wafer. An average Q of 1.9 million was obtained for the 30 cavities, and the highest value was 2.5 million, which is the highest Q reported so far for a nanocavity fabricated by photolithography. Such high Q were realized by the improvements of the nanocavity design and the fabrication process. All nanocavities exhibited a Q of larger than 1 million and the fluctuation of the chip-averaged Q was independent of the chip location. On the other hand, the measured A tended to shift to shorter wavelengths as the distance between the nanocavity and the substrate center increased. Among the nine chips, the difference of the chip-averaged A was as large as 8.0 nm. We consider that a systematic shift of the average air-hole radius by several nanometers is responsible for the large fluctuation of the chip-averaged A. These statistical studies provide important hints to accelerate the application study of PC high-Q nanocavities.

【 概要 】フォトニック結晶を用いて作製される導波路やナノ共振器は,微小サイズ,スローライ ト,高 Q値などの特徴を持つため,新規光デバイス創生を目指して精力的に研究されてきた.特 に,高 Q 値シリコンフォトニック結晶は,世界最高 Q 値[1],低閾値ラマンレーザ[2],CMOS 互 換プロセスによる大量作製など[3],応用上重要な成果を達成してきている.前回我々は、スーパ ールミネッセントダイオード(SLD)を励起光源に用いたフォトニック結晶デバイス評価を報告 した.波長可変半導体レーザを励起光源に用いる場合と比べて,レーザ波長を所望の動作波長に 合わせる必要がないため,応用において有利になると考えられる.今回我々は,SLD 光源を用い ることで,多数のナノ共振器を配置したデバイスを一括励起できることを報告する.さらに,高 性能エッジフィルタと NIRカメラを利用したシンプルなセンシング手法を示す.

We fabricated photonic crystal high-quality factor (Q) nanocavities on a 300-mm-wide silicon-on-insulator wafer by using argon fluoride immersion photolithography. The heterostructure nanocavities showed an average experimental Q value of 1.5 million for 12 measured samples. The highest Q value was 2.3 million, which represents a record for a nanocavity fabricated by complementary metal-oxide-semiconductor (CMOS)-compatible machinery. We also demonstrated an eight-channel drop filter with 4 nm spacing consisting of arrayed nanocavities with three missing air holes. The standard deviation in the drop wavelength was less than 1 nm. These results will accelerate ultrahigh-Q nanocavity research in various areas.