几何操作中重采样问题的数据并行实现

Data Parallel Implementation of Resampling Problem in Geometric Operations

通过SIMD PE阵列之间的数据并行传送,数据并行实现重采样,使实现重采样的复杂度由存储器数据流访问实现方法的Ο(MN)降低到Ο(M+N),从而大大提高了处理速度,能更好地满足图像快速实时处理的需要.

Through the data parallel passing between SIMD PE arrays,this paper implements the problem of ressmpling in geometric operations in data parallel.This kind of data parallel implementation reduces implementing complexity of resampling from Ο（MN） implemented by accessing memory to Ο（M＋N）,and greatly enhances processing speed of processor.Thereby,it can meet the requirements of real-time processing of image effectively.