基于多特征融合的防爆自动导引车避障方法研究

Research on Obstacle Avoidance Method of Explosion-proof Automatic Guided Vehicle based on Multi-feature Fusion

为解决在工厂环境中防爆自动导引车(Automated Guided Vehicle,AGV)难以在不同光照条件下实现行人检测的问题,本文提出将“结构光+双目视觉相机”的图像采集方案应用于爆炸性危险环境。针对传统的基于RGB(代表红、绿、蓝三个通道的颜色)图像的HOG-LBP(Histogram of Oriented Gradient-Local Binary Pattern)方法受光照影响较大且对特征利用率低的情况,设计了基于RGB-D(RGB-Depth)图像的HOG-CLBP(Histogram of Oriented Gradient-Completed Local Binary Pattern)框架,并提出了一种改进的多特征融合方法,通过支持向量机分类实现实时行人目标检测。实验验证该方法可以有效解决在不同光照条件下实现对行人目标的检测,从而实现避障功能。

This paper proposes an approach to applying image acquisition with"structured light+binocular vision camera"to an explosively dangerous environment since it is difficult for explosion-proof Automatic Guided Vehicle(AGV)to detect pedestrians under different lighting conditions in factory environment.Traditional HOG-LBP(Histogram of Oriented Gradient-Local Binary Pattern)method based on RBG(red,blue,green)channels of images is greatly affected by illumination and has low utilization rate.Therefore,a HOG-CLBP(Histogram of Oriented Gradient-Completed Local Binary Pattern)framework based on RGB-D(RGB-Depth)images is designed,and an improved multi-feature fusion method is proposed to realize real-time pedestrian target detection through support vector machine classification.Experiments verify that this method can effectively solve the problems of pedestrian targets under different lighting conditions to achieve the obstacle avoidance function.