复杂光照条件下视觉导引AGV路径提取方法

发布时间：2018-04-14 11:02

本文选题：视觉导引 + 路径提取　；参考：《农业机械学报》2017年10期

【摘要】：针对复杂光照条件下视觉导引AGV的路径提取问题,提出一种基于光照色彩模型的自适应图像照度分区阈值分割方法。首先研究光照照度与图像亮度分量的关系,通过统计复杂光照条件下的图像色彩分布建立光照色彩模型。其次根据光照色彩模型将导引路径图像划分为不同照度区域。然后在RGB色彩空间对低照度区域进行图像增强以还原路径色彩信息,在高亮光区域对色度分量Cb与Cr进行差分运算以抑制共模照度干扰,最后对不同照度区域分别进行自适应阈值分割。大量实验结果表明,在光照环境中同时存在高反光和暗阴影的运行路面,该路径提取方法具有较强的复杂光照适应性,可显著降低在高反光和暗阴影区域的欠分割及过分割误差,对导引路径的识别率为98%。
[Abstract]:Aiming at the path extraction problem of AGV guided by vision under complex illumination conditions, an adaptive image illumination partition threshold segmentation method based on illumination color model is proposed.Firstly, the relationship between illumination intensity and image luminance component is studied, and the illumination color model is established by statistical analysis of image color distribution under complex illumination conditions.Secondly, according to the illumination color model, the guided path image is divided into different illumination regions.Then the low illuminance region is enhanced in the RGB color space to restore the path color information, and the chrominance component CB and Cr are differentially calculated in the high luminous region to suppress the common-mode illumination interference.Finally, adaptive threshold segmentation is carried out for different illumination regions.A large number of experimental results show that there are both high reflectance and dark shadows in the running road surface in the illumination environment, and the path extraction method has a strong complex illumination adaptability.The under-segmentation and over-segmentation errors in high reflection and dark shadow areas can be significantly reduced, and the recognition rate of guidance path is 98%.
【作者单位】：南京航空航天大学机电学院;
【基金】：国家自然科学基金项目(61105114) 中国博士后科学基金项目(2015M580421) 江苏省科技支撑计划项目(BE2014137) 江苏省博士后科研计划项目(1501103C) 江苏省产学研前瞻性联合研究项目(BY2015003-11) 中央高校基本科研业务费项目(NS2016050)
【分类号】：TP242;TP391.41

【相似文献】