基于改进Hough变换的电梯轿厢内人数统计算法

Counting algorithm for numbers of people in elevator car based on improved Hough transform

考虑电梯轿厢内人数的电梯群控调度算法,可以在有限的时间内,找到更优解,对电梯的调度具有重要意义。针对目前基于视觉的人数统计算法正确率不够稳定,速度较慢的问题,提出了一种基于改进Hough变换检测人头轮廓的电梯轿厢内人数统计算法。该算法将标准Hough变换检测圆时的三维空间转换为一维空间,利用圆的几何特性遍历图像数据,通过一维Hough变换进行半径累积,并和阈值进行比较来确定人头轮廓的半径和圆心。结果表明,和用Hough变换检测人头轮廓的传统方法相比,该算法具有速度快,正确率高,抗噪性能好等优点。

The elevator group control dispatch algorithm, which considers the numbers of people in elevator car, can find better solution in the limited time and is very important to the elevator dispatching. According to the instable accuracy and lower speed produced by the correct people counting algorithm based on vision, this paper proposes a counting algorithm for the numbers of people in elevator car based on improved Hough transform which is used to detect head contour. This algorithm converts three dimensional space based on the circle detection by standard Hough transform to one dimensional space, and adopts the geometrical character to traverse the image data, then accumulates value of radius by the one dimensional Hough transform, finally determines the radius and center of the head contour by comparing the values of radius with the threshold. The results of experiment show that the algorithm greatly reduces runtime and improves accuracy, has good anti-noise performance in contrast with the traditional method of head contour detection by Hough transform.