云平台和神经网络的舰船目标检测
|
摘要
为了解决当前舰船目标检测过程中存在的检测误差、检测实时性差的缺点,设计了一种云平台和神经网络的舰船目标检测方法。首先采用混合高斯模型对舰船目标所在区域进行获取,然后采用粒子滤波算法对舰船目标进行跟踪和检测,并采用神经网络对舰船目标粒子滤波算法的权值进行优化和更新操作,解决粒子滤波算法的缺陷,最后基于云平台对舰船目标检测方法进行了设计,并进行了舰船目标检测仿真模拟实验。结果表明,本文方法可以对各种环境中的舰船目标进行准确的检测,提高了舰船目标检测的鲁棒性,而且舰船目标检测实时性也得到了明显的改善,克服了当前舰船目标检测方法存在的缺陷,是一种有效的舰船目标检测方法。
In order to solve the shortcomings of the current ship target detection process, such as detection error and poor real-time performance, a ship target detection method based on cloud platform and neural network is designed. Firstly,the region where the ship target is located is acquired by using Mixture Gauss, then the ship target is tracked and detected by particle filter tracking method, and the weights of the ship target particle filter tracking method are optimized and updated by using neural network to solve the defects of the particle filter tracking method. Finally, the ship target detection method is set up based on cloud platform. The simulation experiments of ship target detection are carried out. The results show that this method can accurately detect ship targets in various environments, improve the robustness of ship target detection, and improve the real-time performance of ship target detection. It overcomes the shortcomings of current ship target detection methods and is an effective method of ship target detection..
In order to solve the shortcomings of the current ship target detection process, such as detection error and poor real-time performance, a ship target detection method based on cloud platform and neural network is designed. Firstly,the region where the ship target is located is acquired by using Mixture Gauss, then the ship target is tracked and detected by particle filter tracking method, and the weights of the ship target particle filter tracking method are optimized and updated by using neural network to solve the defects of the particle filter tracking method. Finally, the ship target detection method is set up based on cloud platform. The simulation experiments of ship target detection are carried out. The results show that this method can accurately detect ship targets in various environments, improve the robustness of ship target detection, and improve the real-time performance of ship target detection. It overcomes the shortcomings of current ship target detection methods and is an effective method of ship target detection..
引文
[1]赵其杰,屠大维,高健,等.基于Kalman滤波的视觉预测目标跟踪及其应用[J].光学精密工程,2008,12(5):7-12.
[2]万琴,王耀南.一种多运动目标检测?跟踪方法研究与实现[J].计算机应用研究,2007,24(1):199-202.
[3]蒋巍,张健,曾浩.基于智能视频监控系统的运动目标检测和跟踪[J].电视技术,2012,36(5):110-114.
[4]黄鑫娟,周洁敏,刘伯扬.自适应混合高斯背景模型的运动目标检测方法[J].计算机应用,2010,30(1):71-74.
[5]陈真.基于蚁群优化算法的云计算资源分配[J].青岛科技大学学报(自然科学版),2012,33(6):619-623.
[2]万琴,王耀南.一种多运动目标检测?跟踪方法研究与实现[J].计算机应用研究,2007,24(1):199-202.
[3]蒋巍,张健,曾浩.基于智能视频监控系统的运动目标检测和跟踪[J].电视技术,2012,36(5):110-114.
[4]黄鑫娟,周洁敏,刘伯扬.自适应混合高斯背景模型的运动目标检测方法[J].计算机应用,2010,30(1):71-74.
[5]陈真.基于蚁群优化算法的云计算资源分配[J].青岛科技大学学报(自然科学版),2012,33(6):619-623.