吉林大学学报(工学版) ›› 2025, Vol. 55 ›› Issue (10): 3346-3351.doi: 10.13229/j.cnki.jdxbgxb.20240998

• 计算机科学与技术 • 上一篇    

基于机器视觉的图像视觉显著目标快速识别算法

赵男男1(),邓超1,温梓呈1,陈金舰2()   

  1. 1.广东海洋大学 计算机科学与工程学院,广东 阳江 529500
    2.辽宁工程技术大学 机械工程学院,辽宁 阜新 123000
  • 收稿日期:2024-09-12 出版日期:2025-10-01 发布日期:2026-02-03
  • 通讯作者: 陈金舰 E-mail:znn_923@163.com;g3sg1@163.com
  • 作者简介:赵男男(1982-),女,副教授. 研究方向:模式识别,计算机应用. E-mail:znn_923@163.com
  • 基金资助:
    2022广东省科技专项基金项目(SDZX2022009)

Fast recognition algorithm for salient objects in image vision based on machine vision

Nan-nan ZHAO1(),Chao DENG1,Zi-cheng WEN1,Jin-jian CHEN2()   

  1. 1.School of Computer Science and Engineering,Guangdong Ocean University,Yangjiang 529500,China
    2.School of Mechanical Engineering,Liaoning Technical University,Fuxin 123000,China
  • Received:2024-09-12 Online:2025-10-01 Published:2026-02-03
  • Contact: Jin-jian CHEN E-mail:znn_923@163.com;g3sg1@163.com

摘要:

在图像视觉显著目标快速识别中,受到光照条件变化、背景复杂性等因素干扰,图像中存在大量噪声干扰显著目标的检测,导致识别鲁棒性差。为此,本文引入机器视觉技术,利用傅里叶变换滤波技术对原始图像进行增强处理,增强对光照变化、背景复杂性等因素的鲁棒性,提升对噪声的抵抗力,提升目标识别的鲁棒性。采用机器视觉中的傅里叶变换滤波技术对原始图像展开处理,生成梯度图,完成原始图像增强。通过多个邻近像素的线性模型计算斜率差的变化趋势,根据斜率差分布的测量值确定最优阈值,引入形态学迭代腐蚀方法有效区分目标区域与噪声区域,实现图像的高清晰度分割。采用多尺度分析策略将图像划分为多个数量不等的超像素区域,计算各超像素内像素的颜色均值,实现图像抽象化表示。基于显著特征的特性,分别对各个尺度下超像素的显著度均值展开计算,并对其展开融合处理,得到图像视觉显著目标识别结果。结果表明:本文算法的区域对比度可达0.597 7,区域一致性可达0.913 2,在不同类型的噪声下目标识别召回率可达99%,本文算法具有较好的一致性,表明本文方法能够有效提升识别结果鲁棒性。

关键词: 机器视觉, 图像视觉显著目标, 快速识别, 傅里叶变换滤波, 梯度图

Abstract:

In the rapid recognition of visually significant targets in images, the presence of a large amount of noise in the image can interfere with the detection of significant targets due to factors such as changes in lighting conditions and background complexity, resulting in poor recognition robustness. To this end, this article introduces machine vision technology and uses Fourier transform filtering technology to enhance the original image, improve its robustness to factors such as lighting changes and background complexity, enhance its resistance to noise, and improve the robustness of target recognition. According to the Fourier transform filtering technique in machine vision, the original image is unfolded and processed to generate a gradient map, completing the enhancement of the original image. By using a linear model of multiple adjacent pixels to calculate the trend of slope difference, the optimal threshold is determined based on the measured values of slope difference distribution. The morphological iterative erosion method is introduced to effectively distinguish the target area from the noise area, achieving high-definition segmentation of the image. Adopting a multi-scale analysis strategy to divide the image into multiple superpixel regions of varying numbers, calculating the color mean of pixels within each superpixel, and achieving abstract representation of the image. Based on the characteristics of salient features, the mean saliency of superpixels at various scales is calculated and fused to obtain the visual salient object recognition results of the image. The results show that the CM of the proposed algorithm can reach 0.597 7, UM can reach 0.913 2, and the target recognition recall rate can reach 99% under different types of noise. The proposed algorithm has good consistency, indicating that the proposed method can effectively improve the robustness of recognition results.

Key words: machine vision, visual salient targets in images, quick identification, Fourier transform filtering, gradient map

中图分类号: 

  • TP391.41

图1

基于机器视觉的图像分割操作流程图"

图2

实验现场图"

图3

数据集原图"

图4

图像分割效果"

表1

不同算法的CM和UM测试结果比较"

测试样本编号测试指标测试算法
本文算法Hough图像分割算法基于PCNN的图像分割算法
01CM0.578 50.547 20.522 8
UM0.879 00.856 60.827 1
02CM0.586 00.531 10.505 3
UM0.899 60.863 60.846 9
03CM0.569 40.541 30.514 8
UM0.913 20.856 30.831 1
04CM0.597 70.500 50.497 7
UM0.912 20.901 70.900 3

表2

识别鲁棒性"

噪声类型目标识别召回率/%

本文

算法

Hough图像分割算法基于PCNN的图像分割算法
高斯噪声996872
椒盐噪声987670
脉冲噪声998179
[1] 廖辉传, 赵海霞. 基于分类器决策融合的红外图像目标识别方法[J]. 红外与激光工程, 2022, 51(8): 282-287.
Liao Hui-chuan, Zhao Hai-xia. Infrared image target recognition method based on decision fusion of classifiers[J]. Infrared and Laser Engineering, 2022, 51 (8): 282-287.
[2] 徐志京, 孙久武, 霍煜豪. 多特征区域的细粒度船舶图像目标识别方法[J]. 计算机工程与应用, 2022, 58(10): 224-230.
Xu Zhi-jing, Sun Jiu-wu, Huo Yu-hao. Fine-grained ship image target recognition method based on multiple feature regions[J]. Computer Engineering and Applications, 2022, 58(10): 224-230.
[3] 尤锟, 张全成, 李凯勇. 基于三维激光点云的运动图像动态目标识别方法[J].应用激光, 2022, 42(4): 147-153.
You Kun, Zhang Quan-cheng, Li Kai-yong. Dynamic object recognition method in moving image based on three-dimensional laser point cloud[J]. Applied Laser, 2022, 42 (4): 147-153.
[4] 严荣慧, 谢海成, 花敏恒, 等. 基于运动与表象特征的广域船舶目标识别方法[J]. 中国舰船研究, 2022, 17(1): 227-234.
Yan Rong-hui, Xie Hai-cheng, Hua Min-heng, et al. Wide-area ship target recognition method based on motion and appearance features[J]. Chinese Journal of Ship Research, 2022, 17(1): 227-234.
[5] 张官荣, 赵玉, 陈相, 等. 基于CFAR和CNN的SAR图像目标识别技术[J]. 电光与控制, 2022, 29(7): 119-125.
Zhang Guan-rong, Zhao Yu, Chen Xiang, et al. SAR image target recognition technology based on CFAR and CNN[J]. Electronics Optics & Control, 2022, 29(7): 119-125.
[6] Guo L, Yang R, Zhong Z, et al. Target recognition method of small UAV remote sensing image based on fuzzy clustering[J].Neural Computing & Applications,2022,34(15):12299-12315.
[7] 李向荣, 陈永康, 王志刚, 等. 一种基于视觉的库区可疑目标识别算法研究[J]. 北京理工大学学报, 2022, 42(4): 424-429.
Li Xiang-rong, Chen Yong-kang, Wang Zhi-gang, et al. Research on a vision-based suspicious target recognition algorithm in reservoir area[J]. Transactions of Beijing Institute of Technology, 2022, 42(4): 424-429.
[8] 赵鹏宇, 王宗彦, 何晓, 等. 基于机器视觉的Delta机器人目标识别算法研究[J]. 工具技术, 2022, 56(3): 113-117.
Zhao Peng-yu, Wang Zong-yan, He Xiao, et al. Research on target recognition and grabber of delta robot based on machine vision[J]. Tool Engineering, 2022, 56(3): 113-117.
[9] 袁小平, 何祥, 王小倩, 等. 基于多层级特征自适应融合的图像分割算法[J]. 浙江大学学报: 工学版, 2022, 56(10): 1958-1966.
Yuan Xiao-ping, He Xiang, Wang Xiao-qian, et al. Image segmentation algorithm based on multi-level feature adaptive fusion[J]. Journal of Zhejiang University(Engineering Science), 2022, 56(10): 1958-1966.
[10] 倪波, 蔡贤涛. Parzen窗算法下图像视觉显著目标识别仿真[J].计算机仿真, 2023, 40(11): 161-164, 266.
Ni Bo, Cai Xian-tao. Simulation of visual significant object recognition in images using Parzen window algorithm[J]. Computer Simulation, 2023, 40(11): 161-164, 266.
[1] 才华,朱瑞昆,付强,王伟刚,马智勇,孙俊喜. 基于隐式关键点互联的人体姿态估计矫正器算法[J]. 吉林大学学报(工学版), 2025, 55(3): 1061-1071.
[2] 李彦清,唐琪,张悦,徐畅,修航. 基于显微机器视觉的大尺寸钢球直径精密分组方法[J]. 吉林大学学报(工学版), 2025, 55(11): 3554-3563.
[3] 李延风,刘名扬,胡嘉明,孙华栋,孟婕妤,王奥颖,张涵玥,杨华民,韩开旭. 基于梯度转移和自编码器的红外与可见光图像融合[J]. 吉林大学学报(工学版), 2024, 54(6): 1777-1787.
[4] 张自超,陈建. 基于双目仿鹰眼视觉与超分辨的果园三维点云重建[J]. 吉林大学学报(工学版), 2024, 54(5): 1469-1481.
[5] 龙关旭,张修石,辛公锋,王涛,杨干. 融合机器视觉的桥梁动态称重方法[J]. 吉林大学学报(工学版), 2024, 54(1): 188-197.
[6] 刘思远,侯跃谦,寇莹,任真,胡正乙,赵雪微,葛云鹏. 基于线结构光视觉的平面度误差测量方法[J]. 吉林大学学报(工学版), 2023, 53(12): 3358-3366.
[7] 马月坤,郝益锋. 考虑特征稀疏特性的短文本命名实体快速识别方法[J]. 吉林大学学报(工学版), 2023, 53(12): 3529-3535.
[8] 包昊菁,刘思远,任真,张云辉,胡正乙,葛宇鹏. 基于机器视觉的链轮尺寸测量方法[J]. 吉林大学学报(工学版), 2023, 53(10): 2795-2806.
[9] 李景彬,杨禹锟,温宝琴,坎杂,孙雯,杨朔. 基于根茬检测的秋后残膜回收导航路径提取方法[J]. 吉林大学学报(工学版), 2021, 51(4): 1528-1539.
[10] 刘恩泽,吴文福. 基于机器视觉的农作物表面多特征决策融合病变判断算法[J]. 吉林大学学报(工学版), 2018, 48(6): 1873-1878.
[11] 林金花, 王延杰, 王璐, 姚禹. 全局相机姿态优化下的快速表面重建[J]. 吉林大学学报(工学版), 2018, 48(3): 909-918.
[12] 张飞, 单忠德, 任永新, 聂军刚, 刘丰. 缸盖缺陷检测系统线阵相机现场标定方法[J]. 吉林大学学报(工学版), 2017, 47(2): 540-545.
[13] 周晓东, 张雅超, 谭庆昌, 张维君. 基于结构光视觉技术的圆柱度测量新方法[J]. 吉林大学学报(工学版), 2017, 47(2): 524-529.
[14] 张波,王文军,魏民国,成波. 基于机器视觉的驾驶人使用手持电话行为检测[J]. 吉林大学学报(工学版), 2015, 45(5): 1688-1695.
[15] 张保华, 黄文倩, 李江波, 赵春江, 刘成良, 黄丹枫. 基于I-RELIEF和SVM的畸形马铃薯在线分选[J]. 吉林大学学报(工学版), 2014, 44(6): 1811-1817.
Viewed
Full text


Abstract

Cited

  Shared   
  Discussed   
No Suggested Reading articles found!