成交(房地产)
计算机科学
书桌
人工智能
卷积神经网络
预处理器
过程(计算)
人工神经网络
深度学习
机器学习
火车
计算机视觉
实时计算
操作系统
政治学
法学
地理
地图学
作者
Zhuoran Xu,Jiafei Zong,Ling Ding,Wei Zhang,Guoguang Cao
标识
DOI:10.1109/cac51589.2020.9327314
摘要
In this paper, a kind of deep learning - based learning table automatic closing anti - clamping technology is proposed. Based on image processing, this technology builds a machine vision model of deep convolutional neural network, collects samples in sections for video analysis, conducts batch image preprocessing on massive data, trains corresponding models respectively, and then conducts real-time monitoring on the closing process of the learning desk. If there is an obstacle, the information will be fed back to the system to stop the closing of the learning desk. If not, continue to perform the closing action until the learning table is completely closed. This method through the actual test, has high response speed and accuracy of the obstacle detection, detection blind area is smaller, the advantages of non-contact and active prevention clip. At the same time, it has strong adaptability, low requirements on hardware, and has a very good application prospect.
科研通智能强力驱动
Strongly Powered by AbleSci AI