为了实现有效的海上监管和响应，提高舰船监管效率，降低人力成本，提出基于遗传算法优化支持向量机的舰船目标识别分类方法。以HU矩为舰船目标的特征描述子，在舰船目标图像内，提取具备旋转、尺度与平移不变性的舰船目标特征矩；利用遗传算法，优化支持向量机的惩罚因子与核参数；在参数优化后的支持向量机内，输入舰船目标特征矩样本，输出舰船目标识别分类结果。实验证明，该方法可有效提取舰船目标特征矩；经过参数优化后的支持向量机，可有效降低计算复杂度，加快检测目标识别分类效率，具备较优的舰船目标识别分类性能。该方法均可精准识别分类舰船目标。
In order to realize effective maritime supervision and response, improve ship supervision efficiency and reduce labor cost, the ship target recognition and classification method of genetic algorithm optimization support vector machine is studied. Taking HU moment as the characteristic descriptor of ship target, the characteristic moment of ship target with rotation, scale and translation invariance is extracted from ship target image. The penalty factor and kernel parameters of SVM are optimized by genetic algorithm. In the support vector machine after parameter optimization, the characteristic moment samples of ship target are input and the recognition and classification results of ship target are output. Experimental results show that this method can extract the characteristic moments of ship target effectively. After parameter optimization, support vector machine can effectively reduce the computational complexity, speed up the detection target recognition and classification efficiency, and has better ship target recognition and classification performance. This method can accurately identify and classify ship targets.
2024,46(4): 174-178 收稿日期：2023-12-29
DOI：10.3404/j.issn.1672-7649.2024.04.033
分类号：TP391.4
作者简介：杨永平(1980-),男,硕士,讲师,研究方向为计算机网络及机器学习
参考文献：
[1] 简涛, 王哲昊, 王海鹏, 等. 基于损失加权修正的舰船目标HRRP小样本元学习识别方法[J]. 信号处理, 2022, 38(12): 2460-2468.
[2] 关欣, 国佳恩, 衣晓. 基于低秩双线性池化注意力网络的舰船目标识别[J]. 系统工程与电子技术, 2023, 45(5): 1305-1314.
[3] 薛安克, 毛克成, 张乐. 多分类器联合虚警可控的海上小目标检测方法[J]. 电子与信息学报, 2023, 45(7): 2528-2536.
[4] 顾鹏, 王碧垚, 黄黔川, 等. 基于KNN和雷达辐射源脉间参数的舰船目标个体识别方法[J]. 中国电子科学研究院学报, 2022, 17(2): 186-192.
[5] 张云, 化青龙, 姜义成, 等. 基于混合型复数域卷积神经网络的三维转动舰船目标识别[J]. 电子学报, 2022, 50(5): 1042-1049.
[6] 蔡一杰, 陈俊杰, 王君, 等. 基于遗传算法优化支持向量机的船用柴油机气门漏气故障智能诊断方法[J]. 内燃机工程, 2022, 43(2): 71-76,84.
[7] 张拯宁, 张林, 王钺, 等. 基于层间分类一致性准则的舰船目标细粒度识别方法[J]. 中国空间科学技术, 2023, 43(3): 93-104.
[8] 柳碧辉, 王培元. 基于改进Faster R-CNN的舰船目标三维识别算法[J]. 火力与指挥控制, 2022, 47(5): 42-52.
