Mapping street trees using google street view and artificial intelligence.
| dc.coverage | Europe | |
| dc.creator | Velasquez, L. F., Echeverria, L., Etxegarai, M., et al. | |
| dc.date | 2023-04-27T13:56:40Z | |
| dc.date | 2023-04-27T13:56:40Z | |
| dc.date | 2022 | |
| dc.date | 2022-08-06T10:58:25.0000000Z | |
| dc.date.accessioned | 2026-06-28T01:50:46Z | |
| dc.description | Urban forests provide ecosystem services to the increasingly urban society, contributing to human wellbeing in cities. While individual-tree characterization and identification are crucial for an efficient and accurate evaluation of urban biodiversity and ecosystem services, the available information on urban trees is rather limited worldwide. Currently, forest inventories require a large amount of human and economic resources, which limit their applicability. Some open data sources such as Google Street View (GSV) offer ground-level images in most urban areas around the world, which, coupled with computer vision techniques and artificial intelligence, provide a promising alternative to fieldwork to conduct urban forest inventories with the aim of characterizing tree diversity and structure and related ecosystem services. Our research aimed at using open data sources such as GSV to map and inventory street trees through an artificial intelligence engine. We developed an automatic transfer learning-based method that allows us to identify urban trees in the images, and the use of remote sensing techniques for geopositioning validation to properly map them with high accuracy (>75%). This method was validated with GSV and Google Maps images as well as with the ground-sourced forest inventory in an urban area (city of Lleida, Spain). This research highlights the potential of artificial intelligence in forest science to generate accurate and efficient mapping of trees, particularly in urban forest ecosystems. Keywords: Innovation, Adaptive and integrated management, Human health and well-being, Monitoring and data collection, Landscape management ID: 3482699 | |
| dc.format | 8p. | |
| dc.format | application/pdf | |
| dc.identifier | https://openknowledge.fao.org/handle/20.500.14283/CC1359EN | |
| dc.identifier | http://www.fao.org/3/cc1359en/cc1359en.pdf | |
| dc.identifier.uri | http://hdl.handle.net/123456789/347776 | |
| dc.language | English | |
| dc.publisher | FAO ; | |
| dc.rights | Non-FAO | |
| dc.title | Mapping street trees using google street view and artificial intelligence. | |
| dc.title | XV World Forestry Congress, 2-6 May 2022 | |
| dc.type | Article |