Computer vision technologies are increasingly used in indoor and outdoor environment research because they can extract rich spatial, environmental, and human-related information from images and videos. Compared with traditional sensors, computer vision provides a non-contact, scalable, and information-rich method for understanding how people interact with buildings, cities, and environmental conditions.

In indoor environments, computer vision can support occupant detection, people counting, posture recognition, activity analysis, and thermal comfort assessment. These applications are especially useful for smart buildings and HVAC control, where real-time information about occupancy, location, behavior, and comfort state can help improve energy efficiency while maintaining occupant comfort. For example, camera-based or thermal image-based systems can detect whether a space is occupied, estimate the number of occupants, identify sitting or standing postures, and provide inputs for demand-driven ventilation, heating, and cooling strategies.

In outdoor environments, computer vision is widely applied in urban and environmental studies. It can be used to analyze pedestrian movement, traffic flow, land use, vegetation coverage, sky view factor, building façade conditions, shading, street quality, and public space usage. With street-view images, drone images, satellite images, and surveillance videos, researchers can evaluate urban form, microclimate conditions, walkability, environmental exposure, and people’s perception of urban spaces at a larger scale.

A major strength of computer vision is its ability to transform visual data into measurable indicators. Deep learning models such as CNNs, YOLO, Mask R-CNN, and vision transformers allow researchers to detect objects, segment scenes, track movement, and classify environmental features automatically. This makes computer vision a powerful tool for both building-scale and urban-scale environmental analysis.

However, several challenges remain. In indoor applications, privacy protection, occlusion, lighting conditions, camera placement, and model accuracy are important issues. In outdoor applications, weather changes, image quality, seasonal variation, data bias, and generalization across different urban contexts can affect performance. Therefore, computer vision should often be combined with other sensing methods, such as environmental sensors, thermal cameras, LiDAR, GPS, WiFi sensing, or occupant feedback.

Overall, computer vision provides a promising approach for indoor and outdoor environment research. It helps researchers better understand human behavior, spatial characteristics, and environmental performance. Its applications can contribute to smarter buildings, more responsive HVAC systems, improved urban design, sustainable planning, and more human-centered environmental research.