研究内容

Research contents

文章聚焦于女性在地铁车站及周边建筑环境中感知不安全的视觉因素。

先前研究已表明女性在公共空间的安全感与周围建筑环境设计有关，但对地铁系统内这一关系的研究较少。

通过新方法分析米兰地铁1号线通勤核心区内视觉因素对女性安全感的影响，涵盖车站及其周边区域，为优化地铁环境提供数据支持。

The article focuses on the visual factors that contribute to women's perceived insecurity in subway stations and the surrounding built environment.

Previous research has shown that women's sense of safety in public spaces is related to the design of the surrounding built environment, but there has been less research on this relationship within subway systems.

A new method is used to analyze the impact of visual factors on women's sense of security in the core commuting area of Milan Metro Line 1, covering the station and its surrounding areas, providing data support for optimizing the subway environment.

研究方法

Research methods

缓冲区数据：使用谷歌街景图像（GSV）获取地铁车站周边400米范围内的图像。

内部区域数据：通过手动摄影获取地铁车站内部关键区域的图像。

语义分割：利用ADE20数据集对缓冲区图像进行语义分割，计算各视觉元素的面积比。

专家审核与机器学习：结合专家审核和机器学习模型（如随机森林）对图像进行评分，评估女性对安全感的感知。

回归模型构建：基于视觉元素面积比与女性感知分数，构建回归模型分析二者之间的关系。

Buffer data: Use Google Street View imagery (GSV) to obtain images within 400 meters of the subway station.

Internal area data: Obtain images of key areas inside subway stations through manual photography.

Semantic segmentation: Use the ADE20K data set to perform semantic segmentation on the buffer image and calculate the area ratio of each visual element.

Expert review and machine learning: Combining expert review and machine learning models (such as random forests) to score images to assess women's perception of 

safety.

Regression model construction: Based on the visual element area ratio and female perception score, a regression model was constructed to analyze the relationship 

between the two.

研究结果

Research conclusions

缓冲区负面影响因素：墙壁、未铺设区域对女性安全感有负面影响。

缓冲区正面影响因素：树木、汽车、路灯、窗户以及更宽的人行道和道路对女性安全感有正面影响。

内部区域负面影响因素：维护不良的基础设施（如破损的路面）、散落的垃圾对女性安全感有负面影响。

内部区域正面影响因素：其他人的存在、更宽的平台、安全亭对女性安全感有正面影响。

基于回归模型，建立了多维评价系统，可对米兰地铁车站及周边环境的安全感进行多维度评估，为优化策略的制定提供依据。

Negative factors affecting the buffer zone: Walls and unpaved areas have a negative impact on women’s sense of security.

Positive buffer factors: Trees, cars, streetlights, windows, and wider sidewalks and roads have a positive impact on women’s sense of security.

Negative factors affecting internal areas: Poorly maintained infrastructure (such as broken pavement) and scattered garbage have a negative impact on women's sense of 

security.

Factors that positively influence the interior area: the presence of other people, wider platforms, and safety booths have a positive influence on women’s sense of security.

Based on the regression model, a multi-dimensional evaluation system was established to conduct a multi-dimensional evaluation of the sense of security at Milan subway 

stations and the surrounding environment, providing a basis for the formulation of optimization strategies.