Motion Vector Estimation based on RFID Localization and Video Sequence

• Main Authors: Chieh-Ling Huang, 黃詰琳
• Other Authors: Pau-Choo Chung
• Format: Others
• Language: en_US
• Published: 2008
• Online Access: http://ndltd.ncl.edu.tw/handle/20733750207126907745

博士 === 國立成功大學 === 電機工程學系碩博士班 === 96 === Motion plays one significant feature in human behavior analysis. Generally speaking the motion vector estimated by video sequence is a short-term motion that facilitated local human behavior analysis. However, some human behavior analysis need long-term motion, that is a wide range motion, such as route tracking. Based on this consideration, this dissertation used RFID device for estimating long-term motion and used video analysis for estimating short-term motion. This dissertation presents an RFID based tracking algorithm with reliability improvement for long-term motion estimation. In this algorithm, RFID Field Generators and Readers are installed in the environment, near the entrance and exit points, for tracking the moving object. Every moving object carries a tag. The Field Generator constantly transmits trigger signal to any tag within its transmission range. When a tag receives the trigger from a Field Generator, it responds to a Reader with the ID of the Field Generator issuing the trigger. Based on the Field Generator IDs a tag responds, we proposed the dynamic range adjustment localization (DRAL) algorithm that estimates the location of the moving object associated with the tag. However by so doing, it is necessary to have significant dense of Field Generators to obtain significantly accurate location. In this algorithm, reference tags are used as reference basis for increasing the localization accuracy. On the other hand, in order to resolve interferences from different Field Generators, this dissertation also proposed a graph coloring with merging and deletion (GCMD) algorithm in RFID system for solving the interference caused from Field Generators located one near another. The short-term motion in human behavior analysis is required to represent the actual motion displacement, rather than regions of visually significant similarity. In this dissertation, region-based selective optical flow back-projection (RSOFB) which back-projects optical flows in a region to restore the region’s motion vector from gradient-based optical flows, is proposed to obtain genuine motion displacement. The back-projection is performed based on minimizing the projection mean square errors of the motion vector on gradient directions. As optical flows of various magnitudes and directions provide various degrees of reliability in the genuine motion restoration, the optical flows to be used in the RSOFB are optimally selected based on their sensitivity to noises and their tendency in causing motion estimation errors. In this dissertation a deterministic solution is also derived for performing the minimization and obtaining the genuine motion magnitude and motion direction.