Acoral Operating System Image Processing Function Library Development And Parallel Optimization

As the computational performance of the multicore proce ssors develops very fast,and its cost becomes lower and lower, multicore processors have been widely used inthe embedded systems. Image processing techology is also increasingly used in theembedded systems, with the power of embedded systems has been grown.On the otherhand, as the power of embedded systems has been progressively better, the technologiesabout image procession have been widly used in the embedded systems as well. Imageprocessing technology always deal with a large amount of data information, so thespeed of it working in the singlecore embedded system is always low, which limits itsuse in many applications. To make the image procession progress running in themulticore embedded system can make up for the lack of singlecore embedded system,which can improve its speed of dealing with information, reduce its costs and use it insome new fields.aCoral operating system is an operating system developed by theEmbedded Real-Time Computing Laboratory of UESTC, basing on a multicoreembedded system. It is an opensource operating system, with hign-cutting and highscalability.This thesis is to develop a image processing library for aCoral called as aCoralCV.This library is consisted with three parts, which have different functions, such as gettingimage, producing results, preprocessing and edge detection.This study investigates following aspects:1. To design aCoralCV for aCoral operating system, and achieve three types offunctions, such as image input and output, image data manipulation, ima ge edgedetection.2. To make functions of aCoralCV running parallelly.3. To make some tests for aCoralCV, to make sure its functions works well.Analyzing performance of parallelized functions, to demonstrate the design of parallelprocessing works well.