運用影像辨識法探討行人步道綠視率與熱影像溫度之相關性分析

The Relationship between Visible Green Index and Temperatures of Infrared Thermograph on Pedestrian Using Image Recognition Technology

張效通(Hsiao-Tung Chang) ； 劉錫睿(Xi-Rui Liu) ； 張慕恩(Mu-En Chang) ； 王常暢(Chang-Chang Wang)

《建築學報》 114_S期：建築物理環境控制專刊 (2020/12) Pp. 39-57

https://doi.org/10.3966/101632122020120114009

摘要

綠視率為綠化面積在行人正常視野面積所占的比例，已經成為衡量公共綠化的三維指標。為探討單位面積下綠視率愈高，則平均溫度是否也愈低；因此，本文透過實景影像及熱影像雙鏡頭的無人機低空飛行模擬人體正視視角下拍攝行人步道影像，進而應用Visual Basic撰寫影像識別程式，將每張照片分成16區塊，解析每區塊之綠視率及熱影像溫度等量化數值。研究成果顯示，綠視率越高熱像平均溫度值則越低、綠視率越低則熱像溫度越高。提高綠視率的效益確能緩解都市熱環境溫度，亦影響人的舒適度感受。

關鍵字

綠視率 ； 熱影像 ； 都市熱環境 ； 行人步道 ； 影像辨識法 ； 無人機

並列摘要

The visible green index is the proportion of green area in pedestrian's normal field of vision, which has become a three-dimensional indicator to measure public greening. To explore whether the higher the visible green index per unit area, the lower the average temperature is. Therefore, through the real image and infrared thermograph by UAV flying in low altitude with two lenses, which simulates the human body to take the pedestrian image from the perspective of the front view. And then, that uses Visual Basic to write the image recognition program, divides each picture into 16 blocks, and analyzes the visible green index and temperatures of infrared thermography values of each block. The research results show that the higher the visible green index, the lower the temperatures of infrared thermograph is, On the contrary, the lower visible green index is, the higher the temperature of infrared thermograph. It can also alleviate the urban thermal environment temperature and affect the comfort of people.

並列關鍵字

Visible Green Index ； Infrared Thermograph ； Urban Thermal Environment ； Pedestrian ； Image Recognition Technology ； Unmanned Aerial Vehicle (UAV)

參考文獻

豊田市(2018)。豊田市緑の基本計画[EB/OL]。引用於2018 年 05 月 03 日，取自http://www.city.toyota.aichi.jp/shisei/gyoseikeikaku/toshiseibi/1007512.html。Toyota (2018). Toyota City Green Basic Plan. [EB/OL]. Japan. on May 3, 2018. From, http://www.city.toyota.aichi.jp/shisei/gyoseikeikaku/toshiseibi/1007512.html

Aoki, Y. (1991). Evaluation methods for landscapes with greenery. Landscape Research, 16(3), 3-6.

Aoki, Y. (2006). Trends of researches on visual greenery since 1974 in Japan. Environmental Information Science, 34(4), 46-49.

Arup (2016). Cities Alive: Towards a Walking World. Arup, London.

Bobiak, J. P., & McGeorge, G. (2015). Assessing hyperspectral image content of pharmaceutical products using the herfindahl–hirschman index and Ripley@@$$s k-functions. Applied Spectroscopy, 69(8), 955-965.