摘要
为提高汽车行驶安全性,结合DM642芯片设计一套基于图像处理的车载浓雾天气识别系统.首先通过特征组合法确定晴朗天和浓雾天图像的最佳识别特征是暗通道均值和灰度均值,然后利用支持向量机(SVM)对分类模型进行训练,最后将识别算法以及分类模型移植到DM642中进行硬件测试.试验结果表明,该算法在硬件上能有效地区分晴朗天与浓雾天图像,识别准确率达到95%.
In order to improve the safety of vehicle driving,a vehicle dense fog weather recognition system based on image processing was designed with DM642 chip.Firstly,the best features of sunny and foggy images were determined by feature combination method,which were dark channel mean and gray scale mean.Then,support vector machine(SVM)was used to train the classification model.Finally,the recognition algorithm and classification model were transplanted into DM642 for hardware test.The experimental results show that the algorithm can effectively distinguish between sunny and foggy images in hardware,and the recognition accuracy is up to 95%.
作者
玉雄侯
杨蓉
杨晓虎
黄俊明
YU Xionghou;YANG Rong;YANG Xiaohu;HUANG Junming(School of Mechanical Engineering, Guangxi University, Nanning 530004, China)
出处
《上海工程技术大学学报》
CAS
2020年第2期119-126,共8页
Journal of Shanghai University of Engineering Science
基金
国家自然科学基金资助项目(61703116)
广西科技基地和人才专项资助项目(AD19110019)
广西创新驱动发展专项资助项目(AA18242045-3)
广西自然科学基金资助项目(2017GXNSFBA198228)
广西教育厅科研资助项目(2018KY0024)
南宁市科技局重点研发计划资助项目(20192065)。
关键词
浓雾天气识别
DM642芯片
暗通道均值
灰度均值
支持向量机
dense fog weather recognition
DM642 chip
dark channel mean
gray scale mean
support vector machine(SVM)
