分块思想在汽车牌照粗定位中的应用

分块思想,是相对于模板匹配而言的一种图像处理思想。它是采用相对静止的方式进行图像匹配,首先根据待匹配图像的某些特征选择块的尺寸,对整幅的图进行分块,利用某些特征值来进行匹配。在以往的车牌定位算法中,有很多文献提出利用模板匹配方法进行车牌定位的思想,但是由于模板匹配方法本身存在计算量大的缺点,直接明显地影响到车牌定位速度,无法满足车牌照自动识别系统的实时性要求,所以本文提出把分块思想应用到车牌照粗定位过程中,利用车牌区域字符自身的一些纹理特征,采用固定分块划分出若干大小相同的特征区域,在各区域内通过水平差分突出垂直方向纹理特征,寻找水平差分累加值最大的区域即为车牌的大致位置,这样就大大提高了车牌定位速度,同时也为后续的车牌精确定位奠定了基础。

Section Thinking Applied in Car License Plate Approximate Location

In contrast with template matching, section thinking remains similar and difference, which is adapted to complete matching image in static way. First, the section with a certain size is decided by means of some features on the image required matching. Second, the whole image is divided into many small section with same size, and the last, some feature values are found to distinguish these small sections for final matching. In the former license plate location algorithms, some articles proposed template matching method in the process of license plate location, but template matching method itself lies in the disadvantage over the calculation, which clearly and directly influence the location speed, and can't live up to real time requirement on the license plate recognition system. So the paper presented the idea, that is, section thinking was applied in car license plate rough location. The whole image was curved up some number of blocks with the same size, and then vertical texture features were enhanced by means of the difference in the horizontal direction, and finally in the way of seeking the maximum among difference accumulating values in the blocks, approximate region of license plate was found. The results show that the method can effectively improve location speed, at the same time, the rough location results are the better base of the next step, exact location.