摘要
针对林区道路图像检测难的问题,提出一种基于多元宇宙优化算法的林区道路图像检测方法。首先输入林区道路的RGB彩色图像,对图像进行2G-R-B变换,得到林区道路的灰度图像;其次设定多元宇宙优化算法的目标函数,对图像进行多元宇宙优化(MOV)算法处理,得到合适的阈值和对应的分割结果;最后对二值图像进行数学形态学处理,得到最优分割图,实现林间道路检测。实验结果表明,该方法具有较强的去噪能力和分割能力,能够很好地检测林区道路,SD、Dice、ER、NR的平均值分别为87.06%、92.84%、14.64%、1.155%。
In view of the difficulty in road image detection in forest areas,a method for forest road image detection based on multivariate universal optimization algorithm was proposed.First,select a forest road'RGB color images to be input,and perform 2G-R-B image transformation to obtain the gray image of the forest road.Secondly,set the objective function of the multivariate universal optimization algorithm,process the image using the multi-cosm optimization(MOV)algorithm to obtain an appropriate threshold and corresponding segmentation results.Finally,perform mathematical image processing of the binary image to obtain the optimal segmentation map to realize the forest road detection.The experimental results show that the method has strong denoising ability and segmentation ability,and can detect forest roads well,with the average values of SD,Dice,ER and NR being 87.06%,92.84%,14.64%and 1.155%respectively.
作者
黄元
付义
康益堃
黄思瑜
耿浩宇
程玉柱
HUANG Yuan;FU Yi;KANG Yi-kun;HUANG Si-yu;GENG Hao-yu;CHENG Yu-zhu(School of Mechanical and Electronic Engineering,Nanjing Forestry University,Nanjing Jiangsu 210037,China;School of Civil Engineering,Nanjing Forestry University,Nanjing Jiangsu 210037,China)
出处
《林业机械与木工设备》
2020年第2期16-19,共4页
Forestry Machinery & Woodworking Equipment
基金
南京林业大学大学生创新创业计划训练项目(2019NFUSPITP0459)
关键词
图像分割
多元宇宙优化算法
阈值
OTSU
image segmentation
multivariate universal optimization algorithm
threshold
Otsu
