亚博网页版

信息学部-人工智能交叉学科论坛

发布时间: 2021-07-05浏览次数: 605

为(wei)进一步推进亚(ya)博网页(ye)版(ban)“双(shuang)一流(liu)”建(jian)设,促进学(xue)(xue)校人工(gong)智(zhi)能(neng)与(yu)相关学(xue)(xue)科(ke)的交叉融合,增强学(xue)(xue)科(ke)间学(xue)(xue)术(shu)交流(liu)与(yu)合作(zuo),形成“人工(gong)智(zhi)能(neng)+X”的复合发展(zhan)模(mo)式,亚(ya)博网页(ye)版(ban)科(ke)学(xue)(xue)技术(shu)研究院(yuan)(yuan)、人工(gong)智(zhi)能(neng)学(xue)(xue)院(yuan)(yuan)、控(kong)制科(ke)学(xue)(xue)与(yu)工(gong)程学(xue)(xue)院(yuan)(yuan)联合举办2021年亚(ya)博网页(ye)版(ban)信息(xi)学(xue)(xue)部-人工(gong)智(zhi)能(neng)交叉学(xue)(xue)科(ke)论(lun)坛,诚邀(yao)各(ge)位老(lao)师参加!

会议时间: 2021年(nian)7月10日(星期(qi)六)上午9:00

会(hui)议地点(dian): 人工(gong)智(zhi)能学院4D322会(hui)议室

会议日程:

日期

时间

内容

主持人

7月10日(ri)

9:00-9:05

刘义副校长代(dai)表(biao)科(ke)学(xue)技术研究(jiu)院致辞(ci)

宋(song)丽(li)梅

9:05-9:10

姜勇副校长(zhang)代表信息学(xue)部致辞

9:10-9:15

合影

 

9:15-9:45

学术(shu)报告: 三维智能视觉检测(ce)与机器人

报 告 人(ren): 宋(song)丽梅(mei) 教授

陈炜

9:45-10:15

学术(shu)报告(gao): 可穿戴传感(gan)器人体(ti)运动能量俘(fu)获自供电技术(shu)的研究

报 告 人: 姚明辉 教授(shou)

10:15-10:45

学(xue)术报告: 面向自动驾驶(shi)构建模型实现夜间(jian)交(jiao)通(tong)雾霾图像的复原

报 告 人: 汤(tang)春明 教授

10:45-11:15

学术报告: 智能(neng)检测技术在复合材料寿命预测及风险(xian)评估(gu)过程中的应(ying)用

报 告(gao) 人: 张(zhang)荣华(hua) 副教(jiao)授(shou)

11:15-11:45

交流研讨(tao)

会议(yi)结束

会议联系人(ren): 杜娟娟 022-83955821,武志超(chao) 022-83955426


报(bao)告题(ti)目: 三维智(zhi)能视(shi)觉检测(ce)与(yu)机(ji)器人(ren)

报(bao)告人: 宋(song)丽(li)梅 教授

报告摘要: 三维(wei)(wei)智(zhi)(zhi)(zhi)能(neng)(neng)视(shi)觉(jue)检测技术可以(yi)帮助机(ji)器(qi)人(ren)(ren)(ren)(ren)(ren)进(jin)行智(zhi)(zhi)(zhi)能(neng)(neng)的(de)检测、智(zhi)(zhi)(zhi)能(neng)(neng)的(de)识(shi)别(bie)、智(zhi)(zhi)(zhi)能(neng)(neng)的(de)定(ding)位、智(zhi)(zhi)(zhi)能(neng)(neng)的(de)跟(gen)踪以(yi)及智(zhi)(zhi)(zhi)能(neng)(neng)的(de)制造。本(ben)报(bao)告(gao)将介绍(shao)实验(yan)室自(zi)主(zhu)(zhu)开发的(de)三维(wei)(wei)智(zhi)(zhi)(zhi)能(neng)(neng)视(shi)觉(jue)检测设备以(yi)及在多个工业场景应用的(de)案例,同时也将介绍(shao)实验(yan)室自(zi)主(zhu)(zhu)开发的(de)智(zhi)(zhi)(zhi)能(neng)(neng)人(ren)(ren)(ren)(ren)(ren)脸识(shi)别(bie)、手势识(shi)别(bie)、姿态识(shi)别(bie)、垃圾分类、人(ren)(ren)(ren)(ren)(ren)体(ti)三维(wei)(wei)扫描、动作捕捉、虚(xu)拟现实等人(ren)(ren)(ren)(ren)(ren)工智(zhi)(zhi)(zhi)能(neng)(neng)相关(guan)设备。三维(wei)(wei)智(zhi)(zhi)(zhi)能(neng)(neng)视(shi)觉(jue)检测技术与机(ji)器(qi)人(ren)(ren)(ren)(ren)(ren)结合最大的(de)亮(liang)点(dian)是让机(ji)器(qi)人(ren)(ren)(ren)(ren)(ren)有(you)了高精(jing)度的(de)三维(wei)(wei)立体(ti)眼睛和大脑,免去人(ren)(ren)(ren)(ren)(ren)类为(wei)机(ji)器(qi)人(ren)(ren)(ren)(ren)(ren)编程(cheng)的(de)过程(cheng),由自(zi)动化向智(zhi)(zhi)(zhi)能(neng)(neng)化转(zhuan)变。

 

报告题(ti)目: 可穿(chuan)戴传感器人体运动能量俘(fu)获自供电(dian)技术的研(yan)究

报告(gao)人: 姚(yao)明辉 教授

报告(gao)摘要: 可(ke)穿(chuan)(chuan)戴(dai)(dai)柔(rou)性(xing)材(cai)料(liao)与器(qi)(qi)件的(de)(de)(de)(de)研(yan)(yan)究(jiu)始于上世纪九(jiu)十年(nian)代(dai),早(zao)期(qi)的(de)(de)(de)(de)研(yan)(yan)究(jiu)大都是亚博网页版柔(rou)性(xing)材(cai)料(liao)的(de)(de)(de)(de)研(yan)(yan)究(jiu),近几年(nian)逐渐融入(ru)了能(neng)(neng)(neng)(neng)(neng)(neng)(neng)(neng)(neng)量(liang)(liang)俘(fu)(fu)(fu)获(huo)和高(gao)效(xiao)转(zhuan)换的(de)(de)(de)(de)研(yan)(yan)究(jiu)。对能(neng)(neng)(neng)(neng)(neng)(neng)(neng)(neng)(neng)量(liang)(liang)俘(fu)(fu)(fu)获(huo)技(ji)术(shu)(shu)的(de)(de)(de)(de)研(yan)(yan)究(jiu)始于上世纪九(jiu)十年(nian)代(dai)末,该(gai)思想的(de)(de)(de)(de)初衷是为低(di)(di)功耗电(dian)子器(qi)(qi)件长期(qi)供(gong)(gong)(gong)能(neng)(neng)(neng)(neng)(neng)(neng)(neng)(neng)(neng),不需(xu)要定期(qi)更换电(dian)池(chi)。而(er)可(ke)穿(chuan)(chuan)戴(dai)(dai)传(chuan)(chuan)感器(qi)(qi)能(neng)(neng)(neng)(neng)(neng)(neng)(neng)(neng)(neng)耗通常在微(wei)瓦至(zhi)毫瓦量(liang)(liang)级,因此(ci),通过能(neng)(neng)(neng)(neng)(neng)(neng)(neng)(neng)(neng)量(liang)(liang)俘(fu)(fu)(fu)获(huo)技(ji)术(shu)(shu)可(ke)实现对其供(gong)(gong)(gong)电(dian)。所(suo)以(yi),近年(nian)来可(ke)穿(chuan)(chuan)戴(dai)(dai)传(chuan)(chuan)感器(qi)(qi)自(zi)(zi)供(gong)(gong)(gong)电(dian)技(ji)术(shu)(shu)逐渐兴起,它(ta)能(neng)(neng)(neng)(neng)(neng)(neng)(neng)(neng)(neng)够推动(dong)(dong)(dong)(dong)人(ren)与可(ke)穿(chuan)(chuan)戴(dai)(dai)电(dian)子产品和自(zi)(zi)驱动(dong)(dong)(dong)(dong)产品的(de)(de)(de)(de)深(shen)度融合,引领技(ji)术(shu)(shu)革新并深(shen)刻(ke)改变(bian)人(ren)类社会。可(ke)穿(chuan)(chuan)戴(dai)(dai)传(chuan)(chuan)感器(qi)(qi)自(zi)(zi)供(gong)(gong)(gong)电(dian)技(ji)术(shu)(shu)的(de)(de)(de)(de)开发设计(ji)涉及(ji)到(dao)力(li)学、机械(xie)、材(cai)料(liao)、电(dian)子、生(sheng)命(ming)、纺织(zhi)等诸多学科(ke),是多学科(ke)交(jiao)叉的(de)(de)(de)(de)研(yan)(yan)究(jiu)领域。根(gen)据能(neng)(neng)(neng)(neng)(neng)(neng)(neng)(neng)(neng)源(yuan)(yuan)转(zhuan)换材(cai)料(liao)的(de)(de)(de)(de)不同,人(ren)体(ti)(ti)(ti)能(neng)(neng)(neng)(neng)(neng)(neng)(neng)(neng)(neng)量(liang)(liang)俘(fu)(fu)(fu)获(huo)的(de)(de)(de)(de)方式有运动(dong)(dong)(dong)(dong)俘(fu)(fu)(fu)能(neng)(neng)(neng)(neng)(neng)(neng)(neng)(neng)(neng)、摩擦俘(fu)(fu)(fu)能(neng)(neng)(neng)(neng)(neng)(neng)(neng)(neng)(neng)、静电(dian)俘(fu)(fu)(fu)能(neng)(neng)(neng)(neng)(neng)(neng)(neng)(neng)(neng)、热俘(fu)(fu)(fu)能(neng)(neng)(neng)(neng)(neng)(neng)(neng)(neng)(neng)和生(sheng)化俘(fu)(fu)(fu)能(neng)(neng)(neng)(neng)(neng)(neng)(neng)(neng)(neng)。与人(ren)体(ti)(ti)(ti)的(de)(de)(de)(de)其它(ta)能(neng)(neng)(neng)(neng)(neng)(neng)(neng)(neng)(neng)源(yuan)(yuan)相比较(jiao),人(ren)体(ti)(ti)(ti)的(de)(de)(de)(de)运动(dong)(dong)(dong)(dong)能(neng)(neng)(neng)(neng)(neng)(neng)(neng)(neng)(neng)源(yuan)(yuan)最(zui)丰富、转(zhuan)换率(lv)(lv)最(zui)高(gao)。与机械(xie)能(neng)(neng)(neng)(neng)(neng)(neng)(neng)(neng)(neng)、风(feng)能(neng)(neng)(neng)(neng)(neng)(neng)(neng)(neng)(neng)、水波能(neng)(neng)(neng)(neng)(neng)(neng)(neng)(neng)(neng)等能(neng)(neng)(neng)(neng)(neng)(neng)(neng)(neng)(neng)量(liang)(liang)的(de)(de)(de)(de)俘(fu)(fu)(fu)获(huo)原理不同,人(ren)体(ti)(ti)(ti)运动(dong)(dong)(dong)(dong)能(neng)(neng)(neng)(neng)(neng)(neng)(neng)(neng)(neng)量(liang)(liang)俘(fu)(fu)(fu)获(huo)具(ju)有不同的(de)(de)(de)(de)采集原理、特点(dian)和应用范围。人(ren)体(ti)(ti)(ti)运动(dong)(dong)(dong)(dong)能(neng)(neng)(neng)(neng)(neng)(neng)(neng)(neng)(neng)量(liang)(liang)具(ju)有低(di)(di)频、周期(qi)性(xing)非谐(xie)(xie)波等特点(dian)。而(er)传(chuan)(chuan)统的(de)(de)(de)(de)能(neng)(neng)(neng)(neng)(neng)(neng)(neng)(neng)(neng)量(liang)(liang)俘(fu)(fu)(fu)获(huo)器(qi)(qi)的(de)(de)(de)(de)振(zhen)源(yuan)(yuan)具(ju)有高(gao)频、谐(xie)(xie)振(zhen)等特点(dian)。因此(ci)传(chuan)(chuan)统的(de)(de)(de)(de)谐(xie)(xie)振(zhen)俘(fu)(fu)(fu)能(neng)(neng)(neng)(neng)(neng)(neng)(neng)(neng)(neng)器(qi)(qi)无法(fa)与人(ren)体(ti)(ti)(ti)的(de)(de)(de)(de)低(di)(di)频运动(dong)(dong)(dong)(dong)相匹(pi)配,因此(ci)较(jiao)难获(huo)得高(gao)效(xiao)的(de)(de)(de)(de)能(neng)(neng)(neng)(neng)(neng)(neng)(neng)(neng)(neng)量(liang)(liang)俘(fu)(fu)(fu)获(huo)效(xiao)率(lv)(lv)。所(suo)以(yi)可(ke)穿(chuan)(chuan)戴(dai)(dai)传(chuan)(chuan)感器(qi)(qi)自(zi)(zi)供(gong)(gong)(gong)电(dian)技(ji)术(shu)(shu)源(yuan)(yuan)于多学科(ke)领域交(jiao)叉的(de)(de)(de)(de)共性(xing)难题是如(ru)何高(gao)效(xiao)俘(fu)(fu)(fu)能(neng)(neng)(neng)(neng)(neng)(neng)(neng)(neng)(neng),如(ru)何提(ti)高(gao)能(neng)(neng)(neng)(neng)(neng)(neng)(neng)(neng)(neng)量(liang)(liang)转(zhuan)换效(xiao)率(lv)(lv)。

 

报告题目: 面向自动驾驶构建(jian)模型实现夜间交通雾霾图像的复原

报告(gao)人: 汤春明 教授(shou)

报告摘(zhai)要: 对(dui)(dui)(dui)于(yu)白天(tian)雾(wu)霾(mai)图(tu)像,最常采用的(de)(de)(de)(de)(de)复(fu)原(yuan)方法是(shi)利用物理模型(xing),即大气散射模型(xing)进行复(fu)原(yuan)。由于(yu)该(gai)模型(xing)中的(de)(de)(de)(de)(de)关键(jian)参(can)数是(shi)针对(dui)(dui)(dui)太阳光(guang)照下的(de)(de)(de)(de)(de)环境(jing)进行估计的(de)(de)(de)(de)(de),因此该(gai)模型(xing)对(dui)(dui)(dui)于(yu)夜(ye)间(jian)(jian)(jian)户外场景的(de)(de)(de)(de)(de)去雾(wu)效果很(hen)(hen)不(bu)理想(xiang)。我(wo)们从提(ti)取夜(ye)间(jian)(jian)(jian)雾(wu)霾(mai)交(jiao)通(tong)视频中的(de)(de)(de)(de)(de)车(che)灯开(kai)始研究,到(dao)暗视野(ye)图(tu)像的(de)(de)(de)(de)(de)整体(ti)(ti)复(fu)原(yuan),直至目前(qian)应用于(yu)夜(ye)间(jian)(jian)(jian)自(zi)动(dong)(dong)驾驶目标识别前(qian)的(de)(de)(de)(de)(de)图(tu)像复(fu)原(yuan)。现(xian)在自(zi)动(dong)(dong)驾驶技(ji)术在夜(ye)间(jian)(jian)(jian)与(yu)恶劣天(tian)候等能见度不(bu)良的(de)(de)(de)(de)(de)环境(jing)下与(yu)日间(jian)(jian)(jian)的(de)(de)(de)(de)(de)差距较大: 日间(jian)(jian)(jian)自(zi)动(dong)(dong)驾驶系统已有Level 3的(de)(de)(de)(de)(de)产品(pin)上市,研发中的(de)(de)(de)(de)(de)技(ji)术更达Level 4~5;而夜(ye)间(jian)(jian)(jian)自(zi)动(dong)(dong)驾驶技(ji)术还在相当(dang)于(yu)Level 1的(de)(de)(de)(de)(de)自(zi)动(dong)(dong)刹车(che)。最大原(yuan)因是(shi)夜(ye)间(jian)(jian)(jian)雾(wu)霾(mai)会(hui)对(dui)(dui)(dui)可见光(guang)图(tu)像中目标的(de)(de)(de)(de)(de)检测和(he)进一(yi)步(bu)识别带来很(hen)(hen)大的(de)(de)(de)(de)(de)干扰(rao)。我(wo)们的(de)(de)(de)(de)(de)研究从经(jing)典的(de)(de)(de)(de)(de)模型(xing)参(can)数预(yu)测到(dao)应用深度学习(xi)方法,以构建轻量型(xing)夜(ye)间(jian)(jian)(jian)去雾(wu)模型(xing),缩短训练时(shi)间(jian)(jian)(jian)从而快速(su)恢复(fu)图(tu)像/视频质量为目标,最终(zhong)提(ti)高目标检测率,以解(jie)决自(zi)动(dong)(dong)驾驶领域非(fei)常迫(po)切和(he)急(ji)于(yu)解(jie)决的(de)(de)(de)(de)(de)前(qian)沿问题。

 

报告题目: 智能(neng)检测技术在复合材料寿(shou)命预(yu)测及风险评估过程中的应用

报(bao)告人(ren): 张荣(rong)华(hua) 副教(jiao)授(shou)

报告(gao)摘要: 碳纤维增强复合材料(Carbon Fiber Reinforced Polymer, CFRP)在(zai)大型(xing)基础设(she)施(shi)主承(cheng)力(li)(li)结(jie)构(gou)中的(de)(de)(de)(de)(de)应用(yong)(yong)可(ke)显著减轻设(she)备(bei)整体质量,提高(gao)能源(yuan)利用(yong)(yong)效率,降低(di)成本。然(ran)而,相(xiang)比于传(chuan)统金属材料,CFRP具有(you)各向(xiang)异性、异质性等特点,在(zai)循环(huan)载荷作(zuo)用(yong)(yong)下其疲劳(lao)演化过(guo)程分(fen)布(bu)于不同(tong)几何尺(chi)(chi)度,导致(zhi)其使用(yong)(yong)寿(shou)命的(de)(de)(de)(de)(de)估(gu)算及(ji)在(zai)役期间的(de)(de)(de)(de)(de)结(jie)构(gou)健康监(jian)测(ce)成为限制(zhi)其广泛应用(yong)(yong)的(de)(de)(de)(de)(de)瓶颈(jing)。采用(yong)(yong)具有(you)非接触、低(di)功耗、无辐射的(de)(de)(de)(de)(de)电磁检(jian)测(ce)方法(fa),设(she)计(ji)并(bing)优化可(ke)分(fen)布(bu)于非平(ping)面(mian)结(jie)构(gou)的(de)(de)(de)(de)(de)角(jiao)度敏感电磁传(chuan)感器阵列,通(tong)过(guo)构(gou)建(jian)采集数(shu)据在(zai)多物理(li)表(biao)征层级的(de)(de)(de)(de)(de)变(bian)分(fen)自编码模型(xing),重建(jian)CFPR不同(tong)几何尺(chi)(chi)度的(de)(de)(de)(de)(de)电磁特征映射,获取CFPR的(de)(de)(de)(de)(de)疲劳(lao)演化过(guo)程并(bing)准确估(gu)计(ji)其剩余寿(shou)命,从而实现对CFRP主承(cheng)力(li)(li)结(jie)构(gou)体在(zai)线健康监(jian)测(ce)和闭环(huan)健康管理(li)。研(yan)究(jiu)成果可(ke)有(you)效降低(di)在(zai)役CFRP主承(cheng)力(li)(li)结(jie)构(gou)的(de)(de)(de)(de)(de)安全风(feng)险(xian),缩(suo)短CFRP结(jie)构(gou)件设(she)计(ji)周期,降低(di)设(she)计(ji)成本,促(cu)进(jin)其在(zai)国民(min)经济相(xiang)关领域更广泛应用(yong)(yong)。

(撰(zhuan)稿人: 杜(du)娟(juan)娟(juan);审稿人: 陈炜(wei)


亚博网页版