Parallel algorithm of computing convex hull based on multinuclear framework

Abstract

The invention provides a parallel algorithm of computing a convex hull based on a multinuclear framework specific to the defect of an existing convex hull algorithm. The parallel algorithm of computing the convex hull based on the multinuclear framework comprises the following steps: (1), finding the initial incomplete convex hull of initial point assembly, indicating edges by counter clockwise direction directed edge; (2), classifying a point set based on the initial incomplete convex hull, finding out all of the outer points of directed edges; (3), iteratively and concurrently breeding each of the directed edge of the initial incomplete convex hull; (4), deleting final acquired non-convex-hull peak on the convex hull. The parallel algorithm of computing the convex hull based on the multinuclear framework effectively optimizes an original algorithm and plenarily saves calculation resources, parallelly expands the classification process of the point set and the iterative process to the original algorithm, and fully utilizes the parallel calculation resources of a multi-core processor; so that parallel speed-up ratio is controlled by controlling parallel mission grain size further by self-adaption selection of the parallel and serial, and possibly produced bottlenecks are eliminated.

Description

[0001] technical field [0002] The invention relates to the field of computational geometry, in particular to a parallel algorithm for calculating convex hulls based on a multi-core architecture. Background technique [0003] Convex hull is the most common and basic structure in computational geometry, and occupies an important position in computational geometry. Convex hulls not only have many properties, but are also effective tools for constructing other geometric shapes. The convex hull is also called the smallest convex hull, which is the smallest convex set that contains all objects in the set S. Among them, the convex hull of the plane point set is the most important and basic problem. The convex hull problem of the plane line segment set and the plane polygon set can be converted into the convex hull problem of the plane point set. The convex hull problem of plane point set is widely used in computer graphics, image processing and pattern recognition, geographic ...

AI Technical Summary

Problems solved by technology

In this algorithm, firstly determine which side of the directed line segment the point is on (i.e. positive and negative division), and then use the Euclidean distance to find the point farthest from the directed line segment. The algorithm is cumbersome and takes up a lot of computing resources.

Images