关键词: 边缘检测, 局部阈值与二值函数, CUDA

Abstract:

Edge detection is an important subject in image processing due to its importance in a wide range of applications.Massive parallelization of algorithms has become very popular in the recent years with the development of the programmable.Graphic Processor Unit,or GPU,has evolved into a highly parallel,multithreaded,many-core process or with tremendous computational power and very high memory bandwidth.A completely different concept is used in the local threshold and Boolean function based edge detection.In that case 3×3 blocks of the image are converted in bit blocks.Some patterns appearing in the bit blocks are considered as indicators of edge existence.This article introduces the principle of local threshold and Boolean function edge detection in detail and its implementation by using CUDA.This algorithm proves more effective and faster than traditional edge detection by comparing the time consumed.

Key words: edge detection;the local threshold and Boolean function;CUDA