[1]崔文飞,边东岩,王会峰,等.基于超像素的感知哈希交通场景图像去重方法[J].机械与电子,2020,(12):9-13.
　CUI Wenfei,BIAN Dongyan,WANG Huifeng,et al.Research on Image Deduplication of Perceptual Hashing Traffic Scene Based on Super Pixel[J].Machinery & Electronics,2020,(12):9-13.
点击复制

基于超像素的感知哈希交通场景图像去重方法()
分享到：

机械与电子[ISSN:1001-2257/CN:52-1052/TH]

卷:
期数:
2020年12期
页码:
9-13
栏目:
设计与研究
出版日期:
2020-12-18

文章信息/Info

Title:
Research on Image Deduplication of Perceptual Hashing Traffic Scene Based on Super Pixel
文章编号:
1001-2257（2020）12-0009-05
作者:
崔文飞 1; 边东岩2; 王会峰1; 武泽键1; 杨文光1
1.长安大学电子与控制工程学院，陕西 西安 710064；
2. 濮阳市高级技工学校，河南 濮阳 457000
Author(s):
CUI Wenfei1; BIAN Dongyan2; WANG Huifeng1; WU Zejian1; YANG Wenguang1
1.School of Electronics and Control Engineering,Chang’an University,Xi’an 710064,China;
 2.School of Puyang Advanced Technical, Puyang 457000,China
关键词:
场景样本库; 图像去重; 超像素; 离散余弦变换(DCT); Jaccard距离
Keywords:
scene sample base;  image de-weighting;  superpixel;  discrete cosine transform(DCT);  Jaccard distance
分类号:
TP391.4
文献标志码:
A
摘要:
针对构建典型交通场景样本库采集数据过程中存在数据重复、相似的问题，提出了一种基于超像素分割下感知哈希的交通场景图像去重算法。首先，对图像超像素分割；然后记录每个超像素区域的像素均值作为图像特征，在分割后的图像中提取像素点组成整幅图像的描述集，并对其进行离散余弦变换（DCT）采用Zigzag模式顺序对变换的系数矩阵编码作为图像DCT特征；最终，采用Jaccard 距离衡量图像相似度，根据权重确定个特征对图像相似度的贡献，确定合理的阈值建立图像去重系统。在KITTI、DeepTesla等数据集上选择部分图像进行实验，实验结果表明，该方法对图像去重的准确率达98.55%，同时具有较高的鲁棒性和稳定性。
Abstract:
Aiming at the problem of data duplication and similarity in the process of building a typical traffic scene sample library to collect data, a perceptual hashing traffic scene image de-duplication algorithm based on superpixel segmentation is proposed. First, segment the image superpixels; then record the pixel average of each superpixel area as an image feature, extract the pixels from the segmented image to form the description set of the entire image, and perform discrete cosine transform (DCT) on it. The Zigzag mode is used to encode the transformed coefficient matrix as the image DCT feature; finally, the Jaccard distance is used to measure the image similarity, the contribution of the two features to the image similarity is determined according to the weight, and a reasonable threshold is determined to establish an image deduplication system. Some images are selected for experiments on KITTI, DeepTesla and other data sets. The experimental results show that the method has an accuracy of 98.55% for image deduplication, and it has high robustness and stability.

参考文献/References:

[1]刘亚芳. 基于感知哈希的图像重复数据删除技术的研究[D].西安：西安电子科技大学,2018.

[2]尚晶.云环境下基于密文图像的块级去重技术研究[D].西安：西安电子科技大学,2019.

[3]SCHNEIDER M,CHANG S F.A robust content based digital signature for image authentication[C]// Proceedings of the IEEE International Conference on Image Processing,1996：227-230.

[4]邓绍江,王方晓,张岱固,等.基于直方图量化和混沌系统的感知图像Hashing算法[J].计算机应用,2008，28(11): 2804-2807.

[5]陈明.图像消冗关键技术研究[D].北京：北京邮电大学,2014.

[6]LI X, LI J, HUANG F L.A secure cloud storage system supporting privacy-preserving fuzzy deduplication[J].Soft computing, 2016,20(4):1437-1448.

[7]LI D P,YANG C, LI C Z,et al.A client-based secure deduplication of multimedia data[C]//IEEE International Conference on Communications,2017:1-6.

[8] 张茹茹.面向图像的云存储重复数据删除方案设计[D].西安：西安电子科技大学,2018

[9]江小平,胡雪晴,孙婧,等.基于分块DCT的图像去重算法[J].中南民族大学学报(自然科学版),2018,37(3):72-75，81.

[10]WANG M R,LIU X B,GAO Y X,et al.Superpixel segmentation: a benchmark[J]. Signal processing: image communication,2017,56:28–39.

[11]ZHOU R G,CHENG Y,LIU D Q.Quantum image scaling based on bilinear interpolation with arbitrary scaling ratio[J]. Quantum information processing,2019,18(9):1332-1573.

[12]HU Z W,ZOU Q,LI Q Q.Watershed superpixel[C]//2015 IEEE International Conference on Image Processing (ICIP), 2015:349-353.

[13]ACHANTA R,SHAJI A,SMITH K,et al.SLIC superpixels compared to state-of-the-art superpixel methods[J].IEEE Transactions on pattern analysis and machine intelligence, 2012,34(11):2274-2282.

[14]WANG H F,WANG Y F, ZHAO X M,et al.Lane detection of curving road for structural highway with straight-curve model on vision[J].IEEE Transactions on vehicular technology,2019,68(6):5321-5330.

[15]FRIDRICH J,GOLJAN M.Robust hash functions for digital watermarking[C]//Proceedings International Conference on Information Technology:Coding and Computing,2000: 178-183.

[16] 丁旭,何建忠.一种由DCT和SURF改进的图像感知哈希算法[J].小型微型计算机系统,2014,35(11):2553-2557.

[17] 寿震宇,钱江波,董一鸿,等.演化森林哈希：一种无监督的在线哈希学习算法[J].电信科学,2020,36(3):71-82.

[18]张晓琳,付英姿,褚培肖.杰卡德相似系数在推荐系统中的应用[J].计算机技术与发展,2015,25(4):158-161,165.

备注/Memo

备注/Memo:
收稿日期：2020-09-02
基金项目：中央高校基本科研业务费创新团队培育项目（300102329401）
作者简介：崔文飞（1995-），男，山东潍坊人，硕士研究生，研究方向为图像处理、智能交通等；边东岩 (1964-)，女， 河北高阳人，高级讲师， 研究方向为机械设计与制造；王会峰（1971-），男，山西永济人，工学博士， 教授， 硕士研究生导师，研究方向为机器视觉与图像处理、交通基础设施智能网联检测等。
更新日期/Last Update: 2020-12-18