基于深度学习目标检测与跟踪技术的研究

Deep learning in object detection and recognition

目标的检测与跟踪技术在计算机视觉领域有着广泛的应用,比如在视频监控,无人驾驶,机器人等领域都有着举足轻重的价值。随着深度学习算法与技术的飞速发展,更是带动了该技术在性能、速度等方面取得了质的飞跃。然而随着社会的发展,需求的不断提高,我们在研究算法高效性的同时还要考虑算法所训练出的模型在实际应用上的性能与速度。本篇文章主要研究基于深度学习的目标检测与跟踪技术,该技术主要包括一个离线训练的检测模型,一个优化的跟踪器,以及一个学习模块来组成在线跟踪系统。通过研究出一种更快、性能更好的算法以及模型的压缩来达到使其训练出的模型在手机等嵌入式设备上实时运行的目的。

The technology of Object detection and tracking has a wide range of applications in the field of computer vision,such as video surveillance,driverless,robotics and other areas.It has a decisive value.Not only the deep learning technology has improved dramatically,but also led to a qualitative leap in performance,speed,and other aspects of the technology.However,with the development of society and the continuous improvement of demand,we must consider the performance and the speed of the trained model by the algorithm in practical applications about the efficiency.This article mainly studies the object detection and tracking technology based on deep learning.The technology mainly includes an offline training detection model,an optimized tracker,and a learning module to form an online tracking system.Through the research of a faster,better-performing algorithm and the compression of the model,the purpose of making the trained model run in real-time on embedded devices such as mobile phones.