自适应车牌抓拍识别算法研究

Self-adaption recognition algorithm of plate number

现有图像分割技术对复杂的环境适应性并不强,在有外界因素干扰的情况下,简单的图像分割技术根本无法满足工程实践的精度要求,针对此种情况,提出一种自适应的图像分割方法,利用小波分析剔除图像噪声,在充分保留图像原始信息的情况下再对图像进行分割处理,该分割方法能够根据不同的环境状况,不断调整X、Y方向的阈值检测范围,获取不同方向上图像最大的梯度变化率,通过图像梯度变化率以及X、Y方向的阈值变化确定最佳的分割区域。最后,以复杂环境下车牌照图例作为算例对所提算法进行验证,测试结果表明,自适应图像分割算法在一定程度上能够适应复杂的环境条件。

The simple image segmentation technology can＇t meet the engineering practical accuracy when the interference factors exist. Aiming at this situation, this paper proposes an adaptive image segmentation method.First, the method adopts wavelet analysis to eliminate image noise. In the situation of fully retaining the original image information, image segmentation processing is performed. This method can constantly adjust the direction of X, Y threshold detection range according to the different environmental conditions and obtain different direction image＇s largest gradient rate.We can determine the best segmentation areas through the image gradient rate and the threshold change of X, Y direction. The adaptive image segmentation algorithm can adapt the complex environment conditions