论文部分内容阅读
针对我国圆形禁令交通标志,提出一种快速、稳定的标志自动检测方法。首先,基于RGB空间对标志图像进行颜色分割,针对三通道差值的固定阈值分割方法对图像亮度的适应性较差的问题,提出基于基础通道选择的相对通道差值计算及阈值曲线拟合的方法进行自适应颜色分割,并针对标志与背景同色的情况,提出梯度过滤的方法将标志与同色背景分离;然后对分割图像进行边缘提取,基于标志椭圆特征,用最小二乘椭圆拟合后验偏差估计法来对边缘进行筛选,并提取最终标志边缘。实验表明,该方法能快速准确地对自然环境下不同亮度的圆形禁令标志自动检测,具有实时性和鲁棒性,在智能交通中具有良好的应用前景。
In view of the circular ban traffic sign in our country, a fast and stable automatic signage detection method is proposed. First of all, based on the RGB space, the color of the logo image is segmented. Aiming at the poor adaptability of the fixed threshold segmentation method of three-channel difference to the brightness of the image, the relative channel difference calculation and threshold curve fitting based on the basic channel selection Aiming at the situation that the sign and background are the same color, a gradient filtering method is proposed to separate the sign from the same color background. Then the edge of the segmented image is extracted. Based on the sign ellipse, the least square ellipse is fitted to the posterior Deviation estimation method to filter the edge, and extract the final logo edge. Experiments show that this method can quickly and accurately detect circular ban signs of different brightness in natural environment automatically, which has real-time and robustness and has good application prospect in intelligent transportation.