基于多通道Gabor滤波的手机按键识别方法

摘要

为实现手机按键的自动化检测,提出一种手机按键识别方法.利用多通道Gabor滤波器组对图像进行滤波,提取局部能量极值作为特征向量,再通过特征向量的匹配进行识别.采用基于按键边框长宽比估算目标图像旋转角度的方法,提高识别方法的效率.实验结果表明,该识别方法在各种干扰的情况下识别率仍能达到90％以上,具有一定的实用性.%In order to achieve automatic detection of the phone keys, this paper proposes a method to recognize the phone keys. It uses the multichannel Gabor filter to filter the images, detectes local energy extremum as the feature vectors, and recognizes the keys by the feature vectors. It proposes a method with estimating the rotation angle of the target image based on frame aspect ratio of the keys, and improves the efficiency of matching algorithm. Experimental result shows that the identification rate of recognition method proposed here is over 90% in the case of various disturbances and it is available.