摘要
道路能见度是影响交通安全的重要因素之一。针对低能见度雾天的道路图像特点,提出一种基于支持向量机回归的车载视频道路能见度检测方法。首先通过区域增长算法确定特征提取区域;然后提取能够反映能见度大小的大气透射率特征和图像熵特征,利用所提取的特征训练支持向量回归模型;最后利用训练好的模型检测图像的能见度。实验结果表明,该方法在400 m以内的能见度检测中总体准确率可达到90.1%,可满足交通行业实际应用的需要。
Road visibility is one of the important factors which affects traffic safety.In this paper,a vehicle⁃mounted video road visibility detection method based on support vector machine regression is proposed according to the characteristics of road images with low visibility in foggy days.Initially,the feature extraction region is determined by employing a region growth algorithm.Subsequently,the key features including atmospheric transmittance and image entropy are extracted.These features are then utilized to train a support vector regression(SVR)model.Finally,the trained model is employed to detect the visibility of the road image.The experimental results show that the overall accuracy of the proposed method in visibility detection within 400 m can reach 90.1%,so the proposed method can meet the needs of practical application in the transportation industry.
作者
周俊
买买提江·吐尔逊
ZHOU Jun;TUERXUN Maimaitijiang(College of Intelligent Manufacturing and Modern Industry,Xinjiang University,Urumqi 830017,China;Xinjiang Key Laboratory of Green Construction and Intelligent Traffic Control of Transportation Infrastructure,Urumqi 830017,China;School of Traffic and Transportation Engineering,Xinjiang University,Urumqi 830017,China)
出处
《现代电子技术》
北大核心
2024年第23期154-158,共5页
Modern Electronics Technique
