A Head-Mounted Assistive Device for Visually Impaired People with Warning System from Object Detection and Depth Estimation

辅助装置 计算机科学 计算机视觉 人工智能 触觉技术 方向(向量空间) 主管(地质) 视力受损 信号(编程语言) 人机交互 物理医学与康复 数学 医学 几何学 地貌学 程序设计语言 地质学
作者
Boonthicha Sae-jia,Rodolfo Lian Paderon,Thatchai Srimuninnimit
出处
期刊:Journal of physics [IOP Publishing]
卷期号:2550 (1): 012034-012034
标识
DOI:10.1088/1742-6596/2550/1/012034
摘要

Abstract People with visual impairment use white cane as their traditional method for perceiving the surroundings. However, the utilization of a cane is limited by its length and orientation. In Thailand, the obstacles on paveway in daily life are not located only on the floor but also above knee level which sometimes could be harmful to pedestrians, especially blind people. A head-mounted assistive device is developed to be an enhancement used with a cane for the visually impaired to comprehend their environment both lower and higher the knee level. The assistive device is designed to be compact and light-weight. It could also send the tactile feedback as a warning from vibration motors mounted on the device. To generate a warning signal, YOLOv4 is used to detect the location of obstacles and depth map from the stereo camera is used to estimate the distance mapping into 4 defined ranges: dangerous, very close, close and fine. The results indicate that the head-mounted assistive device has the ability to perceive obstacles locating farther than 0.9 m. The prediction returned 9.23%, 14.63% and 7.86% error when estimating the depth of obstacles at 1.3 m., 2.8 m. and 4.2 m. respectively. The average execution time for the device to return the command controlling vibration motors is 0.13 second and the maximum estimated time for the motor to send the haptic feedback is 1.05 second.

科研通智能强力驱动
Strongly Powered by AbleSci AI
科研通是完全免费的文献互助平台,具备全网最快的应助速度,最高的求助完成率。 对每一个文献求助,科研通都将尽心尽力,给求助人一个满意的交代。
实时播报
黎建东完成签到,获得积分10
刚刚
CQMEDCHEM完成签到,获得积分10
1秒前
123完成签到,获得积分10
1秒前
Orange应助IU冰冰采纳,获得30
1秒前
科研通AI2S应助IU冰冰采纳,获得10
1秒前
科研通AI6.4应助IU冰冰采纳,获得100
1秒前
妙芙发布了新的文献求助10
2秒前
3秒前
xjz完成签到,获得积分10
4秒前
Lzqqqqq发布了新的文献求助10
4秒前
6秒前
英姑应助xky3371采纳,获得10
7秒前
甜甜薯片关注了科研通微信公众号
7秒前
7秒前
xiaowan完成签到,获得积分10
9秒前
闫格关注了科研通微信公众号
11秒前
11秒前
木质素发布了新的文献求助10
12秒前
Hang完成签到,获得积分10
12秒前
山山而川完成签到 ,获得积分10
13秒前
Mircale完成签到,获得积分10
13秒前
NexusExplorer应助09nankai采纳,获得10
14秒前
14秒前
LayeredSly完成签到,获得积分10
15秒前
15秒前
123发布了新的文献求助10
16秒前
16秒前
喜悦的乞完成签到 ,获得积分10
17秒前
Ava应助香菜采纳,获得10
17秒前
17秒前
19秒前
19秒前
20秒前
英姑应助留胡子的语海采纳,获得10
20秒前
大模型应助碧蓝傲蕾采纳,获得10
20秒前
carrier_hc完成签到,获得积分10
21秒前
shuaiBsen发布了新的文献求助10
22秒前
田様应助沉默的鱼采纳,获得10
22秒前
22秒前
22秒前
高分求助中
Malcolm Fraser : a biography 680
Signals, Systems, and Signal Processing 610
天津市智库成果选编 600
Climate change and sports: Statistics report on climate change and sports 500
Forced degradation and stability indicating LC method for Letrozole: A stress testing guide 500
全相对论原子结构与含时波包动力学的理论研究--清华大学 500
Organic Reactions Volume 118 400
热门求助领域 (近24小时)
化学 材料科学 医学 生物 纳米技术 工程类 有机化学 化学工程 生物化学 计算机科学 物理 内科学 复合材料 催化作用 物理化学 光电子学 电极 细胞生物学 基因 无机化学
热门帖子
关注 科研通微信公众号,转发送积分 6455829
求助须知:如何正确求助?哪些是违规求助? 8266393
关于积分的说明 17618581
捐赠科研通 5522196
什么是DOI,文献DOI怎么找? 2905004
邀请新用户注册赠送积分活动 1881750
关于科研通互助平台的介绍 1724922