【摘要】 为了克服车牌字符的倾斜和相似字符间的误识别对字符识别带来的影响,提出了一种基于不变矩的匹配算法。利用不变矩的旋转不变性克服字符倾斜带来的影响。对不变矩算法进行改进,增加了5个新的不变矩量,包含了更多的细节特征,并用原点矩代替中心矩,减少平移不变性带来的误差,解决了相似字符间的误识别。实验证明了方法的正确性和可行性。更多还原

【Abstract】 In order to overcome the impacts of character slants and error recognition of similar character,an improved matching algorithm based on moment invariants was proposed in this paper.The affect of character slants was avoided by using the rotation invariance.To get more detail features,five new moment invariants were added,and the error of translation invariance was reduced by replacing raw moment with central moment.The error recognition of similar character was reduced in improved algorithm.Experiments show that the method is correct and feasible.更多还原