文章摘要
基于图像相似度的车位识别算法设计要点探究
Exploration of Design Points for Parking Space Recognition Algorithm Based on Image Similarity
投稿时间:2024-11-15  
DOI:10.16018/j.cnki.cn32-1650/n.202503009
中文关键词: 图像相似度  图像感知哈希  图像纹理特征
英文关键词: image similarity  image perceptual hashing  image texture features
基金项目:福建省中青年教师教育科研项目(JAT220529)
作者单位
王宇 集美大学诚毅学院,福建厦门 361000 
摘要点击次数: 41
全文下载次数: 17
中文摘要:
      为提高城市停车管理效率,基于图像相似度方法,设计了一种新型组合判别算法,结合纹理特征与图像感知哈希值,通过加权融合进行车位判断。实验证明,与传统算法相比,基于欧几里得距离、汉明距离、结构相似度的融合方法能够有效提高车位识别准确率,在夜间低光照环境中,传统算法准确率为78.5%,而基于图像相似度的算法则提升至87.3%,提升幅度达到8.8%。
英文摘要:
      In order to improve the efficiency of urban parking management. Based on the method of image similarity, a new combination discrimination algorithm is designed, which combines texture features and image perceptual hash value to judge parking spaces through weighted fusion. Experiments show that, compared with traditional algorithms, the fusion method based on Euclidean distance, Hamming distance and structural similarity can effectively improve the accuracy of parking space recognition. In low-light environment at night, the accuracy of traditional algorithms is 78.5%, while that of algorithms based on image similarity is 87.3%, with an improvement range of 8.8%.
查看全文   查看/发表评论  下载PDF阅读器
关闭