| Summary: | 博士 === 國立成功大學 === 電機工程學系碩博士班 === 101 === This dissertation discusses the digital consumer electronic devices, and presents the sensor-shift optical image stabilizer (OIS) system and the automatic keystone correction (AKC) system. In general, the precise estimation of the hand-shake vibration and controller performance are the most crucial factors of the stabilization performance. This dissertation proposes a frequency adaptive OIS (FAOIS) system which utilizes a fuzzy sliding-mode controller (FSMC) to compensate the nonlinearities of voice coil motor (VCM) and stabilize the dual-axis OIS module. The FSMC does have good robustness while the system has large gravity fluctuation in vertical axis. An extended Kalman filter (EKF) is designed to estimate the tremor frequency in real-time and tune the hand-shake detector online. A fuzzy EKF (FEKF) is used to improve the settling-time of the EKF-based hand-shake detector. The bandwidth of the OIS system with time-varying hand tremor is expended by improving the phase-shift phenomenon and the settling-time of the hand-shake detector. Furthermore, the AKC system adopts the projector-camera pair to provide a visual feedback for monitoring the keystone distortion, a feature extractor and an inclination classifier to determine the projection inclination between the projector and the projection screen, and a pre-warping process to compensate the keystone distortion, caused by non-normal projection poses. Finally, the stabilization performance and frequency adaption of the FAOIS are illustrated by OIS experiments with time-varying hand tremor. All the mobile projection experiments demonstrate that z-axis rotation does not affect the proposed AKC system.
|
|---|
