APPLICATION OF WEARABLE CAMERAS IN STUDYING INDIVIDUAL BEHAVIORS IN BUILT ENVIRONMENTS

Zhaoxi ZHANG, Ying LONG
2019 Landscape Architecture Frontiers  
景观设计学 / 论文 LANDSCAPE ARCHITECTURE FRONTIERS / PAPERS 023 了城市街区环境要素对人群活动的影响;徐磊青 [6] 、潘海啸 [7] 、陈泳 [8] 等 中国学者则聚焦于人在环境中的行为变化,通过观察、问卷、访谈等 方式记录了人们出行和接触环境的情况,分析了城市环境对人们行为 的影响。尽管目前此类研究已经取得一定成果,但关注个体使用者在 城市空间中行为特征的研究仍然较少-受人力和时间成本限制,研 究者难以客观、长期、连续地对某一个体进行行为追踪和数据收集。 随着技术的发展,基于大数据和开放数据共同构成的新数据环境已为 新时代城市空间研究提供了有力支持 [9] [如高德或百度等导航地图中的兴 趣点数据(POIs)、公交卡数据 [10] 、街景数据 [11] 等],但个体数据的获 取仍然存在困难。探索获取个体数据的有效方法将有助于弥补目前在 研究个体行为与建成环境互动关系上的短板。 2 新设备发展 2.1 可穿戴设备 随着数字化和智能化的普及,新技术和新设备不断涌现。智能手 环、智能手表等常见可穿戴设备已经可以通过人机交互记录佩戴者的
more » ... 状态和使用情况。目前应用于个体监测研究中的可穿戴设备大致 分为两类:一种以状态监测为主,如阿米尔·穆阿雷米等 [12] 利用可穿戴 胸带监测个体睡眠和心理压力情况,彼得·阿斯皮纳尔等 [13] 利用心电 传感器监测个体情绪变化;另一种以行为记录为主,如朴重勋等 [14] 利 用计步器分析佩戴者的机体活动,乔治娜·布朗等 [15] 利用穿戴式相机 SenseCam帮助健忘症患者记录日常活动。1945年,范内瓦·布什曾提 第一步 Step 1 个体图片收集 Individual data collection 个体佩戴实验 Wearing experiment 新设备 New device 新技术 New technology 研究思路 Research method 第三步 Step 3 行为空间 Behavior and space 第二步 Step 2 信息识别 Information identification 人工识别 Manual image identification 调用计算机 视觉分析API Image recognition with Computer Vision API 利用Matlab进行 色彩识别 Color calculation in Matlab 行为特征 Behavior 生活方式与时间分配 Lifestyle and time use 空间移动 Space 场所驻留与路径跟随 Places and tracks 行为信息 Behavior 地点信息 Place 时间信息 Time 频率信息 Frequency 叙事性 Narrative descriptions 空间性 Spatiality 连续性 Continuity 规律性 Regularity 建成环境 Built environment 个体时空行为 Individual spatiotemporal behavior 第四步 Step 4 未来讨论 Future discussion 从感知到量化 From perception to quantification 从群体到个体 From group to individual 1 © 张昭希,龙灜 1. 4 © 张昭希,龙灜 5 © 张昭希,龙灜 4. 利用Matlab识别出的蓝 色及绿色比例 5. 人工识别和Matlab色彩 识别的结果对比 4. The ratio of blue color and green color identified through the color calculation in Matlab 5. Comparison of the information acquisition by manual image identification and the color calculation in Matlab 连续出现的人行道、树木、天空 Elements such like sidewalk, trees, and sky presented in continuous photos 行为判断:户外步行/通勤 Behavior identified: walking outdoors / commuting 连续出现的电脑、纸笔 Elements such as computer, paper, and pen presented in continuous photos 行为判断:工作 Behavior identified: working 连续出现的餐具、食物 Elements such as cutlery and food presented in continuous photos 行为判断:就餐 Behavior identified: mealing 连续出现的人物 Elements such as people presented in continuous photos 行为判断:社交 Behavior identified: social 连续出现的货架、商品 Elements such as shop shelves and commodities presented in continuous photos 行为判断:购物/休闲 Behavior identified: shopping / relaxing 7 © 张昭希,龙灜 6. 个体活动时间轴(2018 年9月3日) 7. 测试者的行为信息被概 括 为 工 作 、 通 勤 、 就 餐、社交和休闲5类。 6. The volunteer's timeline on September 3, 2018. 7. The volunteer's behaviors included working, commuting, mealing, social activities, and relaxing. 设备电量不足,晚间通勤未记录完整 Commuting at night was not collected because the battery had run out 设备电量不足,晚间通勤未记录完整 Commuting at night was not collected because the battery had run out 设备电量不足,晚间通勤未记录完整 Commuting at night was not collected because the battery had run out 设备电量不足,晚间通勤未记录完整 Commuting at night was not collected because the battery had run out 设备电量不足,未记录完整 No more information was collected because the battery had run out 设备电量不足,未记录完整 No more information was collected because the battery had run out 设备电量不足,未记录完整 No more information was collected because the battery had run out 设备电量不足,未记录完整 No more information was collected because the battery had run out 设备电量不足,未记录完整 No more information was collected because the battery had run out 设备电量不足,未记录完整 No more information was collected because the battery had run out 9 © 张昭希,龙灜 建成环境 Built environment 行为活动 Behavior 城市大数据 Urban big data 个体大数据 Big data of individuals 新研究 New studies 新视角 New perspectives 公共健康 Public health 群体层面 Group 个体层面 Individual 地点 Location 事件 Scene 时间 Time 生活方式 Lifestyle 13 © 张昭希,龙灜 13. 未来研究将从关注群体 层面转为更加关注个体 层面。 13. Researchers in the future are expected to pay more attention from group study to individual study.
doi:10.15302/j-laf-20190203 fatcat:sgdnwwqonnbavh5r2hrzkmj52m