PARK. M.Y., CHOI. J.Y., KANG J,E. (2023). Analysis of street green rate in busan using deep learning model. IAIA23 Annual Conferences (2023. 05. 08. - 2023. 05. 11.)
<Abstract>
This study sought to determine the status of securing street green by analyzing the street green rate for National Route 7 and National Route 2, which pass through the center of Busan, as a new green area indicator. Also this study conduct a Pearson correlation analysis between street green rate and land price, number of buildings, population density which explain development degree. The process of calculating green rate is following. First, using the national standard link, link information on the Route 2 and 7 is collected, and then latitude and longitude coordinates are collected by designating the specific location at 50M intervals. After that, a macro program was created to collect a road view near the latitude hardness, and 866 photos of the road view were collected. The collected road views were extracted only from green areas using a deep learning model called MIT Driving Scene Segmentation. The street green rate was calculated using (number of pixels divided into green areas/total number of pixels)*100 expression, and the green rate was visualized through ArcMap. As a result of the analysis, the average street green rate was 10.3%, which was lower than the target level of previous studies. As a result of conducting a Pearson correlation analysis, land price showed a weak correlation with a correlation coefficient of ?0.464 and the others were not correlated. This study is meaningful in that it used the street green rate as a new stereoscopic indicator that can determine the green area rate of citizens, and employed a deep learning model that has high accuracy.
사사 : 기후변화특성화대학원