2-stage Pipelined GHT Structure for Positioning Simple Label Images

다국어 초록 (Multilingual Abstract)

Almost every products include label image which contain various informations about product. It is necessary to find location and rotated angle of label image to decide whether label image contain defection or not. By using Generalized Hough Transform(GHT) we can find location and rotated angle of label image robustly, But it require huge amount of memory space and long computation time to positioning image. In this paper, a 2-stage pipelined GHT structure which reduce memory requirement and computation time is proposed. Obtaining location and rotation angle of label image is partitioned to reduce memory space. We selectively vote edge points to reduce computation time. An experiment is conducted using many label image samples. The result show that proposed algorithm can reduce enormous memory space and computation time.

목차 (Table of Contents)

• Abstract
• I. INTRODUCTION
• II. PROPOSED 2-STAGE STRUCTURE
• III. EXPERIMENTS
• IV. RESULT
• V. CONCLUSION
• REFERENCES