Integration and differentiation: comparison of photography behaviors using unmanned aerial vehicle data in China and Europe

The increasing popularity of consumer-grade unmanned aerial vehicles (UAVs), or drones, has revolutionized aerial photography. This study addresses the growing trend of recreational UAV photography by analyzing a large dataset of user-generated content (UGC) from SkyPixel, a major online platform for sharing drone imagery. The study's primary focus is on understanding the differences and similarities in photography behaviors between Chinese and European users. This research is timely and important due to the rapid expansion of the consumer UAV market and the lack of existing research specifically on the recreational use of UAVs in tourism. By leveraging advancements in artificial intelligence (AI), particularly computer vision, sentiment analysis, and spatial analysis techniques, this research aims to generate insights into the photographic preferences, emotional responses, and spatial distribution patterns of UAV photographers in these two regions. Understanding these patterns can inform tourism destination planning and promotion strategies, as well as offer valuable insights into the unique perspectives and impacts of UAV technology on tourism experiences and visual representations of destinations.

Existing research on tourism photography has largely focused on traditional methods using cameras or mobile phones. Early studies utilized postcards or visitor-employed photography (VEP), which involved researchers guiding tourists to take photos. However, VEP is costly and subject to bias. The rise of the Internet and social media platforms like Flickr has opened new avenues for research using user-generated content (UGC). Studies using Flickr data have analyzed the contents of photos, exploring differences in photography behavior and destination image perceptions. However, UAV photography presents unique characteristics. The aerial perspective offers a previously unavailable viewpoint, allowing for broader context and visual impact. This study bridges a gap by applying AI-powered image analysis to a large dataset of UAV photographs, extending existing research beyond traditional photography methods and incorporating spatial analysis to explore geographic patterns.

This study used data from SkyPixel, DJI's global aerial photography platform. A web crawler collected UAV photographs, captions, and geolocation data. The dataset consisted of 1230 photos: 466 from China and 764 from Europe. The researchers employed several analytical methods: 1. **Computer Vision Analysis:** Microsoft Azure's cloud-based computer vision API was used to extract keywords describing the content of each photo. This AI-powered analysis identified objects, actions, and scenes within the images, generating a list of keywords for each photograph. 2. **Social Network Analysis:** Gephi software was used to analyze the co-occurrence of keywords, creating networks to visualize the relationships between different content categories in photos from China and Europe. This helps understand the connections between image elements, for instance, identifying frequent combinations of keywords like "natural landscape" and "water". 3. **Text Sentiment Analysis:** Microsoft Azure's text analytics API analyzed the emotional tone of photo captions. This sentiment analysis provided a numerical score indicating the positivity or negativity of each caption, distinguishing between positive, negative, and neutral sentiments. 4. **Spatial Analysis:** Kernel density estimation (KDE) was applied to the geolocation data to visualize the spatial distribution of UAV photography hotspots in both China and Europe. This technique highlights areas with high concentrations of photographs, identifying popular locations for drone photography.

The analysis revealed several key findings: **Photo Content Analysis:** Both Chinese and European photographers predominantly captured natural landscapes (27.31% and 29.84% respectively). However, significant differences emerged in other categories. Chinese photos more frequently featured human landscapes, traffic (especially large vehicles), and food, indicating a focus on human activity within the environment. European photos, conversely, emphasized characters participating in outdoor sports and nature scenes, reflecting a preference for showcasing dynamic activities in natural settings. **Photo Network Feature Analysis:** Network analysis showed that both regions' core photography themes included natural landscapes, daily behaviors, adjectives, and transportation. However, Europe displayed stronger connections between sports and characters, while China had a more diverse range of themes on the edges of the network, indicating a broader range of associated content in Chinese UAV photography. **Sentiment Content Analysis:** Chinese photographers exhibited largely neutral emotions (0.5), while European photographers showed a wider distribution across the emotional spectrum, with a notable proportion expressing positive emotions (0.7-1). Negative sentiment in both datasets frequently correlated with harsh weather conditions (snow, ice) and barren landscapes. Positive sentiment was associated with scenes of people engaging in recreational activities near water bodies. **Spatial Analysis:** Hotspots for UAV photography were concentrated in established tourist areas, particularly near bodies of water, in both regions. However, spatial distribution differed. China showed a wider spread of data points, concentrated along the southeast coast, reflective of population density and economic development. Europe's distribution was more focused on Central and Southern Europe, with high density areas in Germany and Italy. This difference may be attributed to variations in UAV regulations, which are often more restrictive in certain European countries.

The findings suggest that while the overarching theme of natural landscapes unites UAV photographers across China and Europe, cultural and regulatory factors influence the specific content and emotional expression within their photography. The greater emphasis on human activity in Chinese photos aligns with the region's bustling cities and higher population density, while the European preference for outdoor sports and untouched nature may reflect differing cultural values and recreational preferences. The difference in emotional expression may also be attributed to cultural norms surrounding emotional expression, with Europeans showing greater openness compared to Chinese photographers. The spatial distribution patterns highlight the role of tourism infrastructure and regulations in shaping the landscape of UAV photography. The study's results underscore the potential of AI-powered techniques in analyzing large UGC datasets to provide valuable insights into tourism behaviors and trends.

This study provides a novel comparative analysis of UAV photography behaviors in China and Europe, highlighting both commonalities and differences influenced by culture and regulations. The core themes of natural landscapes, transportation, and human activities emerge across both regions, but distinct preferences shape image content and emotional expression. The findings are valuable for destination marketing organizations (DMOs) and tourism planners to tailor strategies for this growing segment of travelers and photographers. Future research could investigate how evolving UAV regulations and technology will continue to shape photographic preferences and behavior.

This study's reliance on SkyPixel data may introduce bias, as it doesn't represent the entire population of UAV users. The sentiment analysis primarily focused on photo captions, potentially missing nuances in emotional expression. Further qualitative research, including interviews with photographers, could enrich the understanding of motivations and interpretations of images. Finally, the study's spatial analysis is limited by the distribution of data points within SkyPixel; areas with limited UAV activity might not be fully represented.