离线建模法在虚拟键盘中的应用

Offline modeling method in the application of the virtual keyboard

针对基于图像处理技术的虚拟键盘存在的按键定位准确度低的问题,提出了一种基于图像处理的离线建模的方法。该方法将摄像头和键盘距离保持固定,通过摄像头采集键盘图像,轮廓查找算法定位键盘的每一行,再进行一些简单的计算即可建立键盘在采集的图像中的位置信息。后续程序只需检测手指按下的位置,通过坐标转换即可定位用户按下的目标按键,实现输入功能。本文在VC6.0编程环境下,用开源机器视觉库(OpenCV),编写了软件算法,实验表明该方法能够高准确度定位键盘位置,为虚拟键盘的研究提供了新的设计方法。

For the low accuracy in the key identification of virtual keyboard based on image processing technique, an offiine modeling method based on image processing was presented. In this method, the distance between the camera and keyboard was constant, the camera captured the keyboard image, each row was located by using contours finding algorithm, and then the keyboard location information in the captured image was got by some simple computation. In the follow-up steps, the target key clicked by users can easily be located by detecting the finger press position and coordinate transformation, and then the input function was implemented. We used the Open Source Computer Vision Library in the VC6.0 programming environment, and wrote the software algorithm. Finally, in the practical test, it showed that the keyboard position can be located with high accuracy, which provides a new reference approach for the virtual keyboard research.