自动引导车视觉导航的灰度投影稳像算法

Gray Projection Image Stabilization Algorithm of AGV Visual Navigation

自动引导车(AGV)在运行过程中由于路面不平等原因产生随机抖动,引起摄像头所获得图像产生位移导致后续路径视觉跟踪不稳定;针对此问题,本研究搭建了AGV视觉抖动模型,在图像中划分出路径区域作为目标区域进行后续处理以减少背景中其它物体带来的干扰,然后通过快速灰度投影算法检测平移向量,利用运动补偿校正视频图像从而达到稳像目的。仿真结果表明:相对传统的投影曲线匹配法,本研究所用算法显著提高了图像平移向量检测精度并大大提高了运算效率,满足AGV对稳像系统高效性和实时性的要求。

Due to random jitter caused by the road irregularity in the process of AGV movement,the images obtained by camera displacement causes subsequent visual tracking unstable.To solve this problem,the paper builts AGV visual jitter model,which carves out a path area in the image as the target area for subsequent processing to reduce background interference caused by other objects.Then we use quickly gray projection algorithm to detect the translation vector,and uses motion compensation correction to achieve image stabilization purposes.Compared to the traditional projection curve matching method,simulation results show that the algorithm herein used can significantly improve the detection accuracy of the translation vector images,and the computational efficiency,and meet the requirements of AGV image stabilization system on efficiency and real-time.