A platform for research: civil engineering, architecture and urbanism
A platform for research: civil engineering, architecture and urbanism
Vehicle window control method based on gesture recognition
The invention relates to the field of automobile intelligent cabin control, in particular to an automobile window control method based on gesture recognition, which comprises the following steps: acquiring gesture image data in an automobile by using a camera and an infrared sensor, dynamically matching gesture images by using a DTW time sequence matching algorithm, and classifying by using a kNN classification strategy so as to control the opening and closing of an automobile window. An interaction mode based on gesture recognition is adopted, instant response is achieved, and user experience is more friendly. The distance between different gesture templates is measured by using the DTW algorithm, and then classification is performed through the k-nearest neighbor classifier, so that the recognition accuracy is higher, the time distortion is more robust, the calculation efficiency is higher, the individual difference is more adaptive, the k-DTW focuses on matching the existing templates, and large-scale data labeling for training is not needed.
本发明涉及汽车智慧座舱控制领域，具体涉及一种基于手势识别的车窗控制方法，利用摄像头、红外传感器采集车内手势图像数据,采用DTW时间序列匹配算法对手势图像进行动态匹配，利用kNN分类策略进行分类,进而控制车窗的开启和关闭。本发明采用基于手势识别的交互方式,即时响应,用户体验更友好。利用DTW算法衡量不同手势模板之间的距离,再通过k近邻分类器进行分类，识别准确率更高、时间扭曲更加鲁棒、计算效率更高、个体差异更加适应，k‑DTW侧重匹配已有模板,不需要大规模标注数据进行训练。
Vehicle window control method based on gesture recognition
The invention relates to the field of automobile intelligent cabin control, in particular to an automobile window control method based on gesture recognition, which comprises the following steps: acquiring gesture image data in an automobile by using a camera and an infrared sensor, dynamically matching gesture images by using a DTW time sequence matching algorithm, and classifying by using a kNN classification strategy so as to control the opening and closing of an automobile window. An interaction mode based on gesture recognition is adopted, instant response is achieved, and user experience is more friendly. The distance between different gesture templates is measured by using the DTW algorithm, and then classification is performed through the k-nearest neighbor classifier, so that the recognition accuracy is higher, the time distortion is more robust, the calculation efficiency is higher, the individual difference is more adaptive, the k-DTW focuses on matching the existing templates, and large-scale data labeling for training is not needed.
本发明涉及汽车智慧座舱控制领域，具体涉及一种基于手势识别的车窗控制方法，利用摄像头、红外传感器采集车内手势图像数据,采用DTW时间序列匹配算法对手势图像进行动态匹配，利用kNN分类策略进行分类,进而控制车窗的开启和关闭。本发明采用基于手势识别的交互方式,即时响应,用户体验更友好。利用DTW算法衡量不同手势模板之间的距离,再通过k近邻分类器进行分类，识别准确率更高、时间扭曲更加鲁棒、计算效率更高、个体差异更加适应，k‑DTW侧重匹配已有模板,不需要大规模标注数据进行训练。
Vehicle window control method based on gesture recognition
一种基于手势识别的车窗控制方法
XU LINHAO (author) / LIANG CI (author) / ZHANG ZHENGXUAN (author) / CAO RONGGE (author) / HUA DEZHENG (author) / HAO JINGBIN (author) / LIU XINHUA (author) / LIU XIAOFAN (author) / ZHOU HAO (author)
2023-12-19
Patent
Electronic Resource
Chinese
System and method for controlling opening of vehicle window through gesture recognition
| European Patent Office | 2021
Closestool control system and method based on gesture recognition
| European Patent Office | 2024
UWB-based vehicle window control method and vehicle window control system
| European Patent Office | 2023