计算机与现代化 ›› 2022, Vol. 0 ›› Issue (01): 54-60.

• 图像处理 • 上一篇    下一篇

基于显著性目标检测的弹着点定位算法

  

  1. (1.上海师范大学信息与机电工程学院,上海200234; 
    2.中国科学院上海微系统与信息技术研究所无线传感网与通信重点实验室,上海201800)
  • 出版日期:2022-01-24 发布日期:2022-01-24
  • 作者简介:周璇(1997—),女,浙江衢州人,硕士研究生,研究方向:计算机视觉,目标检测,E-mail: 1125153438@qq.com; 通信作者:朱苏磊(1975—),女,副教授,硕士生导师,硕士,研究方向:图像处理,嵌入式开发,阵列信号处理,E-mail: suleizhu@163.com; 何为(1980—),男,研究员,博士生导师,博士,研究方向:目标定位,位置信息服务,E-mail: wei.he@mail.sim.ac.cn。
  • 基金资助:
    国家重点研发计划项目(2018YFC1505204)

Impact Point Detecting Algorithm Based on Salient Object Detection

  1. (1. College of Information, Mechanical and Electrical Engineering, Shanghai Normal University, Shanghai 200234, China;
    2. Key Laboratory of Wireless Sensor Network and Communication, Shanghai Institute of Microsystem and Information Technology,
     Chinese Academy of Sciences, Shanghai 201800, China)
  • Online:2022-01-24 Published:2022-01-24

摘要: 针对目前炮弹定位方法安全隐患大、人工测量效率低、精度差的问题,本文提出一种基于显著性目标检测网络BASNet(Boundary-Aware Salient Object Detection)的弹着点定位方法。采用改进的BASNet网络,结合注意力机制模块CBAM(Convolutional Block Attention Module)、金字塔池化模块PPM(Pyramid Pooling Module)与深度可分离卷积,对炮弹火焰进行显著性检测,提取弹着点图像坐标。实验结果表明,该方法在自制的炮弹火焰数据集上的检测精度F值达到0.914,MAE为0.006,推理速度为3.86 fps,优于BASNet、U2Net等显著性目标检测网络。该方法提取的弹着点图像坐标与真实坐标误差为5.92个像素值,相比于BASNet网络减少近4.85个像素值。综合可知,该算法增强了网络对显著性物体内部的检测精度,提高了模型推理效率,减少了图像弹着点坐标误差,适用于靶场小范围炮弹火焰烟雾的检测,能够满足靶场应用的实测需求。

关键词: 炮弹定位, 显著性目标检测, 注意力机制, 金字塔池化, 深度可分离卷积

Abstract: Aiming at the problems of hidden danger, low efficiency of manual measurement and poor accuracy of current projectile flame location methods, an improved projectile flame detection algorithm based on salient target detection network BASNet (Boundary-Aware Salient Object Detection) is proposed in this paper. Using the improved BASNet network, combined with attention mechanism module-CBAM (Convolutional Block Attention Module), pyramid pooling module-PPM (Pyramid Pooling Module) and depth separable convolution, to detect the projectile fire and extract the coordinates of impact point on the image. The experimental results show that detection accuracy of F measure reaches 0.914, the mean absolute error-MAE reaches 0.006 and detection speed reaches 3.86 fps,better than other salient object detection network like BASNet, U2Net. The error between the coordinates of impact point image extracted by this method and the real coordinates is 5.92 pixels, which is 4.85 pixels less than BASNet. In conclusion, the improved network retains the effective shallow semantic information, enhances the detection accuracy of the network for the significant objects, improves the efficiency of model reasoning, and is suitable for the detection of small-scale projectile fire smoke in the shooting range, which can meet the actual needs of the application in the shooting range.

Key words: projectile positioning, salient object detection, attention mechanism, pyramid pooling, depth separable convolution