摘要
在分析阿拉伯数字以及大写英文字母的形状特征后,提出了一种适于现场可编程门阵列(FPGA)实现的图像实时识别算法.该算法能够充分描述数字和字母图像的特征向量,然后经过预处理和定位后提取出各特征向量,最后计算出各特征向量间的距离并比较大小就能对图像进行实时识别.算法对图像平移和小角度旋转具有不变性,并且具有简单、识别准确率高、计算复杂度低、存储量小的优点.FPGA仿真结果表明,该算法只需要8 000逻辑门就能实时识别出50×75像素大小的数字和字母图像,证明了算法具有较高的可行性和高效性.
A real time recognization algorithm of Arabic numerals and English letters based on field-programmable gate array （FPGA） was proposed after analyzing the features of the numerals and letters. The algorithm proposes a group of vectors that can describe the characters of the numerals and letters; then, drows the characters of them after preprocessing and location; finally, calculates and compares the distance of vectors to recognize them in real time. The algorithm can recognize Arabic numerals and English letters with a little rotation and position offset. Besides, it is simple, but the accuracy is high. In addition, both the calculation complexity and the storage needed are small. The FPGA results show that it needs only 8 000 gates to recognize pictures with sizes of 50 × 75 pixels, which proves the feasibility and effectivity of the algorithm.
出处
《上海交通大学学报》
EI
CAS
CSCD
北大核心
2006年第1期12-15,共4页
Journal of Shanghai Jiaotong University
关键词
图像识别
数字识别
英文字母识别
现场可编程门阵列
image recognization
Arabic numerals recognization
English letters recognization
field-programmable gate array （FPGA）
