首页|基于集成学习和深度学习的海上目标联合分类识别方法

基于集成学习和深度学习的海上目标联合分类识别方法

扫码查看
为了提高复杂多元海况环境下海上目标的分类准确率,联合利用集成学习和深度学习,选择随机森林模型、深度森林模型、CNN模型和BiGRU模型,提出评估因子合成模型进行海上目标分类识别.实例验证表明,该评估因子合成模型提高了海上目标的识别准确率,可清晰认知海上形成的整体动态局势,从而助力于海域决策的及时性与准确性.
Joint classification and recognition method of maritime targets based on integrated learning and deep learning
In order to improve the classification accuracy of marine targets in complex and multivariate sea state environment,integrated learning and deep learning are combined to select random forest model,deep forest model,CNN model and BiGRU model,and propose an evaluation factor synthesis model for marine target classifi-cation and recognition.Example verifications show that this evaluation factor synthesis model can improve the recognition accuracy of maritime targets,and can have a clear cognition of the overall dynamic situation at sea,so as to help the timeliness and accuracy of maritime decision-making.

maritime targetevaluation factor synthesis modelintegrated learningdeep learning

张晨、靳俊峰

展开 >

中国电子科技集团公司第三十八研究所,合肥 230088

孔径阵列与空间探测安徽省重点实验室,合肥 230088

海上目标 评估因子合成模型 集成学习 深度学习

2024

空天预警研究学报
空军预警学院

空天预警研究学报

影响因子:0.39
ISSN:2097-180X
年,卷(期):2024.38(6)