已入深夜,您辛苦了!由于当前在线用户较少,发布求助请尽量完整地填写文献信息,科研通机器人24小时在线,伴您度过漫漫科研夜!祝你早点完成任务,早点休息,好梦!

Research on Positioning and dimension measurement of truck compartment based on Semantic segmentation in automatic loading system

卡车 坐标系 职位(财务) 计算机视觉 像素 人工智能 过程(计算) 计算机科学 分割 点(几何) 维数(图论) 模拟 工程类 汽车工程 数学 纯数学 财务 操作系统 几何学 经济
作者
Fengying Ma,Bowen Zhang,Peng Ji,Yuhe Sun,Jie Li,Xuexin Li,Wei Yan,Fengying Wang,Gaofeng Mu
标识
DOI:10.1145/3614008.3614022
摘要

Given the current situation that the domestic loading machine system needs to adjust the truck body to stop at the specified position with the specified attitude before automatic loading, to solve the problem that the parking position or body attitude of the truck is not consistent with the posture during teaching because the position or attitude of the truck cannot be adjusted to the specified posture, the automatic loading system cannot complete the automatic loading. In this paper, the positioning and dimension measurement of the truck compartment in the process of automatic loading system are studied. Firstly, two industrial cameras are used to collect the images of the truck from two different fields of vision, and the images are stitched together to obtain the complete images of the truck and vehicle. Then FCN-ResNet50 network with the CBAM attention mechanism is used to segment the truck image. This method can accurately detect the position of the truck in different postures. Secondly, the reference point is selected through the feature plate to determine the length of unit pixel of the image. Finally, the segmentation image of the compartment area is used to locate the compartment area and four corners of the compartment. Combining the length of the unit pixel point and the image coordinate position of the four corner points, the size of the compartment is calculated through the coordinate mapping between the image coordinate system and the world coordinate system. The system is verified by truck model experiment, evaluated by MIoU value and measured truck size results. The experimental results show that the MIoU value of the segmentation image of the carriage can reach 97.5%, and the measured size of the carriage area is within the allowable error range of the actual size of the truck model, which eliminates the restriction that the truck needs to stop in strict accordance with the designated position.

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
无花果的应助被万事顺遂采纳,获得10
刚刚
刚刚
JamesPei的应助被金仕王采纳,获得10
1秒前
1秒前
ding的应助被sxystc采纳,获得10
2秒前
3秒前
犯懒完成签到,获得积分10
4秒前
5秒前
chcmuer发布了新的文献求助10
6秒前
pxy666关注了科研通微信公众号
6秒前
科研小能手完成签到,获得积分10
6秒前
7秒前
尾状叶完成签到 ,获得积分0
7秒前
8秒前
9秒前
无心的怜南完成签到,获得积分20
10秒前
tttwwx发布了新的文献求助10
10秒前
11秒前
ccccc完成签到,获得积分10
12秒前
执行完成签到,获得积分10
13秒前
14秒前
zero1122完成签到 ,获得积分10
14秒前
15秒前
科研通AI6.2的应助被庞steven采纳,获得10
17秒前
pxy666发布了新的文献求助10
18秒前
gpz完成签到,获得积分20
19秒前
19秒前
xinqisusu完成签到 ,获得积分10
20秒前
万事顺遂发布了新的文献求助10
21秒前
wang关注了科研通微信公众号
21秒前
lance发布了新的文献求助30
22秒前
名井南来北往完成签到 ,获得积分10
22秒前
gpz发布了新的文献求助10
25秒前
26秒前
书竹完成签到,获得积分10
29秒前
若即若離发布了新的文献求助30
30秒前
alice完成签到,获得积分10
32秒前
可靠老头完成签到 ,获得积分10
33秒前
35秒前
好眠哈密瓜完成签到 ,获得积分10
36秒前
高分求助中
(应助此贴封号)【重要!!请各用户(尤其是新用户)详细阅读】【科研通的精品贴汇总】 10000
自動車の空力技術 800
Organizational Behavior 510
Management and the Arts 510
Geschichtliche Grundbegriffe (GGB), Band 5: Pro–Soz 300
Die Religion in Geschichte und Gegenwart (RGG), 4. Auflage, Band 7: R–S 300
Die Religion in Geschichte und Gegenwart (RGG), 4. Auflage, Band 1: A–B 300
热门求助领域 (近24小时)
化学 材料科学 医学 生物 纳米技术 计算机科学 化学工程 工程类 有机化学 物理 复合材料 生物化学 内科学 细胞生物学 基因 遗传学 免疫学 冶金 光电子学 癌症研究
热门帖子
关注 科研通微信公众号,转发送积分 7791825
求助须知:如何正确求助?哪些是违规求助? 9329108
关于积分的说明 20426761
捐赠科研通 7381474
什么是DOI,文献DOI怎么找? 3323563
关于科研通互助平台的介绍 2471374
邀请新用户注册赠送积分活动 2340568