| 摘要: |
| 针对于海杂波噪声背景下传统方法难以检测出小目标的问题，提出了基于改进小波阈值函数的降噪方法和IBWO-KELM(Improved Beluga Whale Optimization-Kernel Extreme Learning Machine)检测方法。首先，利用变分模态分解(Variational Mode Decomposition，VMD)结合改进的小波阈值函数对海杂波数据进行降噪处理；其次，对白鲸优化(Beluga Whale Optimization，BWO)算法的初始化、探索和开发3个方面进行多策略改进，并利用改进的BWO算法对核极限学习机(Kernel Extreme Learning Machine，KELM)的正则化系数和核参数进行优化，从而获得一个最佳的预测模型；最后，根据预测的绝对误差实现海面小目标检测。采用IPIX雷达数据进行实验，在#17、#280和#54号海杂波数据中平均检测效率相较于RCMDE-XGBoost检测方法提升了15%。 |
| 关键词: 小目标检测 海杂波 白鲸优化算法 核极限学习机 |
| DOI：10.20079/j.issn.1001-893x.240723001 |
|
| 基金项目:国家自然科学基金资助项目（61901195） |
|
| Small Target Detection Method in Sea Clutter Background Based on IBWO-KELM |
| LU Huiyang,XI Caiping |
| (College of Automation,Jiangsu University of Science and Technology,Zhenjiang 212100,China) |
| Abstract: |
| In response to the problem that traditional methods are difficult to detect small targets in the background of sea clutter noise,a denoising method based on improved wavelet thresholding function and Improved Beluga Whale Optimization-Kernel Extreme Learning Machine(IBWO-KELM) detection method are proposed.Firstly,the variational mode decomposition(VMD) combined with the improved wavelet threshold function is used to denoise the sea clutter data.Secondly,the BWO algorithm is improved in three aspects of initialization,exploration and development,the regularization coefficient and kernel parameter of the kernel extreme learning machine(KELM) are optimized by the improved BWO algorithm,so as to obtain a best prediction model.Finally,the small target detection on the sea surface is realized according to the predicted absolute error.IPIX radar data is used for experiments,and the average detection efficiency of sea clutter data #17,#280 and #54 is increased by 15% compared with that of the RCMDE-XGBoost detection method,which verifies the effectiveness of the IBWO-KELM detection method. |
| Key words: small target detection sea clutter waves beluga whale optimization algorithm kernel extreme learning machine |
《电讯技术》编辑部 