【摘要】 对面部和手部图像进行正确定位是当前图像处理领域中的重要课题 .在机器人手势识别实验中 ,我们试图利用一个BP神经网络对连续手势进行识别 .由于实时性的要求 ,我们希望找到一种尽量简单的图像划分方法 .本文通过对采集到的彩色图像进行色度变换和色调过滤 ,成功的实现了从复杂背景的彩色图像中划分出手部图像 ,并提取其轮廓以供处理 .作为计算机图像识别的预处理过程 ,这种划分方法算法简单 ,易于实现 ,十分适合于实时彩色图像的划分与识别 .更多还原

【Abstract】 To locate correctly the image of the face and hand is an important subject in the area of image processing nowadays. In the experiment of hand recognition for a robot, we tried to use a BP neural network to recognize the hand image sequence. Due to the request of real time processing, we expected to find the simplest method to divide the target image from the source one. In this paper, a new scheme is proposed to divide hand image from complex background and to take out its outline for processing, in which, hue transform and hue filtration play the key role. As the pre processing for image recognition, this method has been proved to be very suitable for the dividing and recognition of real time color images, since its simple algorithm is easy to be realized.更多还原