基于机器学习的船舶吃水动态检测系统设计
|
摘要
传统的吃水检测系统存在检测抗扰性差的弊端,为了解决传统检测方法的缺点,动态检测系统得到应用。设计基于机器学习的通航船舶吃水动态检测系统,基于机器学习技术,在现有激光测量器上利用DSP测量技术,构成DSP激光多点测量单元,完成硬件单元构建。采用NPS多波束声呐图源滤波算法,对硬件采集图像做多波束滤波运算,完成图像滤波运算,解决对检测抗扰性差的问题。对比实验证明,提出的基于机器学习的通航船舶吃水动态检测系统设计,能够准确、实时、动态、自主检测船舶吃水数据。各项测试结果均优于传统吃水检测系统。
Traditional draught detection system has the disadvantages of poor detection immunity. In order to solve the shortcomings of traditional detection methods, dynamic detection system has been applied. Based on machine learning technology, a dynamic detection system for navigable ship draft is designed. On the basis of machine learning technology, a DSP laser multi-point measurement unit is constructed by using the DSP measurement technology on the existing laser measuring device, and the hardware unit is completed. The NPS multi-beam sonar image source filtering algorithm is used to perform multi-beam filtering operation on the hardware image acquisition, and the image filtering operation is completed to solve the problem of poor detection immunity. The comparative experiments prove that the design of dynamic detection system for navigable ship draft based on machine learning can accurately, real-time, dynamically and independently detect the ship draft data. The test results are superior to the traditional draught detection system.
Traditional draught detection system has the disadvantages of poor detection immunity. In order to solve the shortcomings of traditional detection methods, dynamic detection system has been applied. Based on machine learning technology, a dynamic detection system for navigable ship draft is designed. On the basis of machine learning technology, a DSP laser multi-point measurement unit is constructed by using the DSP measurement technology on the existing laser measuring device, and the hardware unit is completed. The NPS multi-beam sonar image source filtering algorithm is used to perform multi-beam filtering operation on the hardware image acquisition, and the image filtering operation is completed to solve the problem of poor detection immunity. The comparative experiments prove that the design of dynamic detection system for navigable ship draft based on machine learning can accurately, real-time, dynamically and independently detect the ship draft data. The test results are superior to the traditional draught detection system.
引文
[1]阎羡功.一种船舶吃水深度检测系统的设计[J].舰船科学技术, 2016, 38(2A):49–51.
[2]张杭琦.海底声呐数据的可视化去噪算法仿真[J].计算机仿真, 2017, 34(11):176–179.
[3]张森,尚根锋,普杰信,等.海洋资源水下目标图像准确提取仿真[J].计算机仿真, 2018, 35(8):188–193+260.
[4]黄允浒,韩燕,吐尔洪江·阿布都克力木.一种基于àTrous算法的水下图像增强算法[J].电子设计工程, 2018, 26(16):6–11+18.
[2]张杭琦.海底声呐数据的可视化去噪算法仿真[J].计算机仿真, 2017, 34(11):176–179.
[3]张森,尚根锋,普杰信,等.海洋资源水下目标图像准确提取仿真[J].计算机仿真, 2018, 35(8):188–193+260.
[4]黄允浒,韩燕,吐尔洪江·阿布都克力木.一种基于àTrous算法的水下图像增强算法[J].电子设计工程, 2018, 26(16):6–11+18.