摘要
手势识别成为了一种新型的人机交互手段,对于社会生产和日常生活都有着非常重要的研究意义。以OpenCV为基础,通过HSV色彩模型来分割手势,用Canny图像边缘检测的方法来绘制手势轮廓,建立在图像的Hu不变矩来进行模板匹配,对常见的数字静态手势进行实时的分类识别。结果表明,基于OpenCV的静态手势识别技术达到预期效果,准确率达92%。
Gesture recognition has become a novel method of human-computer interaction,which has very important research significance for social production and everyday life.On the basis of OpenCV,this research makes use of HSV color model to segment gestures,uses Canny image edge detection to draw gesture profiles,and performs template matching based on Hu invariant moments of images,and conducts real-time classification and recognition of usual digital static gestures.The results show that the static gesture recognition technology based on OpenCV achieves the expected effect and the precision rate reaches 92%.
作者
于蕾
YU Lei(School of Information Engineering and Media,Hefei Polytechnic University,Hefei 230011,Anhui)
出处
《攀枝花学院学报》
2022年第5期86-92,共7页
Journal of Panzhihua University
基金
安徽省自然重点项目“基于多层半监督单类极限学习机(ml-s2ocell)的运输图像特定目标检测关键技术研究”(KJ2021A1399)
合肥职业技术学院自然重点项目“基于深度学习的运输图像特定目标检测关键技术研究”(2021KJA10)。
