融合式空间塔式算子和HIK-SVM的交通标志识别研究

发布时间：2018-06-10 17:55

本文选题：智能交通 + 交通标志识别　；参考：《交通运输系统工程与信息》2017年01期

【摘要】：在交通标志识别问题上,提出了一种基于融合式的空间塔式算子和直方图交叉核支持向量机(HIK-SVM)的分类方法.在该方法中,通过提取图像的灰度塔式词袋直方图(Gray-PHOW)特征、颜色塔式词袋直方图(Color-PHOW)特征和塔式边缘方向梯度直方图(PHOG)特征来对交通标志的外观、颜色和轮廓信息进行描述.通过提取空间塔式直方图特征,能很好地对图像各种特征的空间分布状况进行描述.提取到图像的外观、颜色、轮廓和特征的空间分布信息后,对其进行融合,最后得到的融合式的空间塔式特征具有很强的鲁棒性.将该融合式特征送入HIK-SVM进行训练和分类,取得了极其高的识别效果.
[Abstract]:In the problem of traffic sign recognition, a classification method based on fusion space tower operator and histogram cross kernel support vector machine (HIK-SVM) is proposed. In this method, the appearance, color and contour information of traffic signs are described by extracting gray-PHOW (gray-PHOW) features, Color-PHOW (Color-PHOW) features and gradient histogram (PHOGG) features of tower edge direction. By extracting spatial tower histogram features, the spatial distribution of various image features can be described well. After extracting the spatial distribution information of the appearance, color, contour and feature of the image, the image is fused. Finally, the fused spatial tower features have strong robustness. The fusion feature is fed into HIK-SVM for training and classification, and the recognition effect is very high.
【作者单位】：武汉理工大学自动化学院;
【基金】：国家自然科学基金(61374151)~~
【分类号】：TP391.41

【参考文献】