增强随机蕨的安全帽佩戴检测  

Safety Helmet Wearing Detection Using Enhanced Random Fern

作　　者：张乾 岳诗琴 范玉 白金华 ZHANG Qian;YUE Shi-qin;FAN Yu;BAI Jin-hua(Academic Affairs Office,Guizhou Minzu University,Guiyang Guizhou 550025,China;Key Laboratory of Pattern Recognition and Intelligent Systems of Guizhou,Guiyang Guizhou 550025,Chin;School of Data Science and Information Engineering,Guizhou Minzu University,Guiyang Guizhou 550025,China)

机构地区：[1]贵州民族大学教务处,贵州贵阳550025 [2]贵州省模式识别与智能系统重点实验室,贵州贵阳550025 [3]贵州民族大学数据科学与信息工程学院,贵州贵阳550025

出　　处：《计算机仿真》2021年第7期429-432,共4页Computer Simulation

基　　金：国家自然科学基金(61802082,61263034,61762020);贵州省重点实验室(黔科合Z字[2009]4002号);贵州省教育厅自然科学基金(黔教合KY字[2017]129号);教育部产学合作协同育人基金(201702044007)。

摘　　要：针对随机蕨算法中二元测试的对数和蕨的数量难以确定问题,采用了网格搜索与交叉验证法改进并应用于安全帽检测应用。首先在图像中随机取二元组像素点对比较并形成0/1元素的序列(随机蕨),其次采用密度估计估计其随机蕨分布,然后网格搜索与交叉验证法搜索其中参数进行调整和优化构建增强随机蕨丛。通过在公开数据集上进行实验,实验结果表明,提出的算法较随机蕨算法有较大提升,说明了网格搜索与交叉验证法在随机蕨中的有效性。Aiming at the problem that it is difficult to determine numbers of binary test and the ferns in the random fern algorithm, grid search and cross validation method were used to improve and apply to the application of helmet detection. Firstly, binary pixel pairs were randomly selected from the image and compared to form a sequence of 0/1 elements(random fern). After that, density estimation was used to estimate the random fern distribution. Then, grid search and cross validation were used to search the parameters of random fern for adjustment and optimization to construct it. Experimental results on public data sets show the proposed algorithm is better than the random fern algorithm, which all shows the effectiveness of grid search and cross validation method in random fern.

关 键 词：安全帽检测 随机蕨 二元测试 密度估计 

分 类 号：TP391.9[自动化与计算机技术—计算机应用技术]