摘要
在舰船遥感图像目标检测中,传统模型对海面背景干扰的抑制能力较弱,导致舰船目标区域周围存在太多附属冗余背景,降低了舰船目标区域完整性。针对这一问题,提出最小二乘支持向量机的舰船遥感图像目标检测研究。挖掘图像样本中的隐含规则,分析背景局限性,利用背景色与前景色间的灰度值差异,提取并标记潜在目标区域,过滤冗余背景信息,实现最小二乘支持向量机的舰船遥感图像目标检测。为测试该方法的目标检测效果,利用仿真实验,对比该方法与传统方法间的差异,综合主观及客观实验结果可知,应用最小二乘支持向量机后,有效抑制了背景干扰,能够获取到更加清晰完整的舰船目标区域轮廓。
In the ship remote sensing image target detection,the traditional model has a weak ability to suppress the sea background interference,resulting in too many redundant background around the ship target area,reducing the integrity of the ship target area.To solve this problem,the least square support vector machine(LSSVM)is proposed to detect the ship remote sensing image.Mining the hidden rules in the image samples,analyzing the limitations of the background,using the gray value difference between the background color and the foreground color,extracting and marking the potential target area,filtering the redundant background information,and realizing the ship remote sensing image target detection based on the least square support vector machine.In order to test the target detection effect of the method,the simulation experiment is used to compare the difference between the method and the traditional method.According to the subjective and objective experimental results,the least square support vector machine can effectively suppress the background interference and obtain a clearer and complete contour of the ship target area.
作者
庞峰
PANG Feng(Shanxi Police College,Taiyuan 030401,China)
出处
《舰船科学技术》
北大核心
2020年第18期79-81,共3页
Ship Science and Technology
基金
山西省“1331”工程公安技术学重点学科建设项目(1331KBC)
2020山西警察学院院级教学改革创新项目重点课题(YJ202012)
关键词
最小二乘支持向量机
目标检测
目标提取
遥感图像
least squares support vector machine
target detection
target extraction
remote sensing image
