摘要
由于激光雷达图像目标检测的过程中,没有将背景和目标区分开,导致图像目标检测的精度较低,检测时间较长,检测效果较差,为此提出基于视觉传达的激光雷达图像目标检测方法。通过置信滤波方法对激光雷达图像实行预处理,划分激光雷达图像的背景区域和目标区域,将目标区域输入到基于视觉传达技术的视觉特征提取器中,提取图像目标区域特征,采用多核学习理论优化支持向量机,将提取的图像特征输入优化后的支持向量机中,实现激光雷达雷达图像的目标检测。实验结果表明,所提方法的检测效果好、检测精度高、检测时间短。
In the process of lidar image target detection, the background and target are not separated, resulting in low accuracy, long detection time and poor detection effect. Therefore, a lidar image target detection method based on visual communication is proposed. Through the confidence filtering method, the lidar image is preprocessed, the background area and target area of the lidar image are divided, the target area is input into the visual feature extractor based on visual communication technology, the image target area features are extracted, the multi-core learning theory is used to optimize the support vector machine, and the extracted image features are input into the optimized support vector machine to realize the target detection of lidar image. The experimental results show that the proposed method has good detection effect, high detection accuracy and short detection time.
作者
张娜
王静
王春霞
ZHANG Na;WANG Jing;WANG Chunxia(Jiaozuo University,Jiaozuo Henan 454000,China)
出处
《激光杂志》
CAS
北大核心
2023年第2期113-117,共5页
Laser Journal
基金
教育部规划基金(No.21A11522001)
河南省教育科学规划课题(No.2021YB0697)
河南省科技计划项目(No.212400410356)。
作者简介
张娜(1982-),女,硕士,讲师,研究方向:视觉传达、美术学。
