基于YOLOv8的目标检测算法改进研究
Research on the Improvement of Target Detection Algorithm Based on YOLOv8
肖富坤1
作者信息
- 1. 沈阳航空航天大学,辽宁 沈阳 110136
- 折叠
摘要
现有的目标检测网络往往存在结构复杂、参数量巨大的问题.因此,研究结构简单、参数量少的高精度目标检测算法,其重要性和价值不言而喻.YOLO系列算法作为深度学习时代具有代表性的单阶段目标检测算法,具有准确、高效和易于部署等特点,为目标检测提供了非常好的理论及技术基础.因此,文章在YOLOv8 的基础上对其卷积网络进行改进,实验结果证明,改进算法成功提升了检测精度.
Abstract
The existing target detection network often has the problems of complex structure and huge parameters.Therefore,it is difficult to study the high-precision target detection algorithm with simple structure and a few parameters.As a representative single-stage target detection algorithm in the era of Deep Learning,YOLO series algorithms have the characteristics of accuracy,high efficiency and easy deployment,which provides a very good theoretical and technical basis for target detection.Therefore,this paper improves its convolutional network based on YOLOv8 and successfully improves the detection accuracy.
关键词
目标检测/高精度/YOLOv8/卷积网络Key words
target detection/high accuracy/YOLOv8/convolutional network引用本文复制引用
出版年
2024
国家科技期刊平台
NSTL
万方数据