Mapping of Allergenic Tree Species in Highly Urbanized Area Using PlanetScope Imagery—A Case Study of Zagreb, Croatia
Mapping and identifying allergenic tree species in densely urbanized regions is vital for understanding their distribution and prevalence. However, accurately detecting individual allergenic tree species in urban green spaces remains challenging due to their smaller site and patchiness. To overcome...
| Main Authors: | , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2023-06-01
|
| Series: | Forests |
| Subjects: | |
| Online Access: | https://www.mdpi.com/1999-4907/14/6/1193 |
| _version_ | 1797594708473544704 |
|---|---|
| author | Mateo Gašparović Dino Dobrinić Ivan Pilaš |
| author_facet | Mateo Gašparović Dino Dobrinić Ivan Pilaš |
| author_sort | Mateo Gašparović |
| collection | DOAJ |
| description | Mapping and identifying allergenic tree species in densely urbanized regions is vital for understanding their distribution and prevalence. However, accurately detecting individual allergenic tree species in urban green spaces remains challenging due to their smaller site and patchiness. To overcome these issues, PlanetScope (PS) satellite imagery offers significant benefits compared with moderate or high-resolution RS imagery due to its daily temporal resolution and 3 m spatial resolution. Therefore, the primary objectives of this research were to: assess the feasibility of mapping allergenic tree species in the highly urbanized area using high-resolution PS imagery; evaluate and compare the performance of the most important machine learning and feature selection methods for accurate detection of individual allergenic tree species. The research incorporated three classification scenarios based on ground truth data: The first scenario (CS1) used single-date PS imagery with vegetation indices (VI), while the second and third scenarios (CS2 and CS3) used multitemporal PS imagery with VI, and GLCM and VI, respectively. The study demonstrated the feasibility of using multitemporal eight-band PlanetScope imagery to detect allergenic tree species, with the XGB method outperforming others with an overall accuracy of 73.13% in CS3. However, the classification accuracy varied between the scenarios and species, revealing limitations including the inherent heterogeneity of urban green spaces. Future research should integrate high-resolution satellite imagery with aerial photography or LiDAR data along with deep learning methods. This approach has the potential to classify dominant tree species in highly complex urban environments with increased accuracy, which is essential for urban planning and public health. |
| first_indexed | 2024-03-11T02:27:06Z |
| format | Article |
| id | doaj.art-0fa0ce4095cf4cbfbadc70e2c53cd03d |
| institution | Directory Open Access Journal |
| issn | 1999-4907 |
| language | English |
| last_indexed | 2024-03-11T02:27:06Z |
| publishDate | 2023-06-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Forests |
| spelling | doaj.art-0fa0ce4095cf4cbfbadc70e2c53cd03d2023-11-18T10:27:38ZengMDPI AGForests1999-49072023-06-01146119310.3390/f14061193Mapping of Allergenic Tree Species in Highly Urbanized Area Using PlanetScope Imagery—A Case Study of Zagreb, CroatiaMateo Gašparović0Dino Dobrinić1Ivan Pilaš2Chair of Photogrammetry and Remote Sensing, Faculty of Geodesy, University of Zagreb, 10000 Zagreb, CroatiaChair of Geoinformatics, Faculty of Geodesy, University of Zagreb, 10000 Zagreb, CroatiaCroatian Forest Research Institute, Division of Ecology, Cvjetno naselje 41, 10450 Jastrebarsko, CroatiaMapping and identifying allergenic tree species in densely urbanized regions is vital for understanding their distribution and prevalence. However, accurately detecting individual allergenic tree species in urban green spaces remains challenging due to their smaller site and patchiness. To overcome these issues, PlanetScope (PS) satellite imagery offers significant benefits compared with moderate or high-resolution RS imagery due to its daily temporal resolution and 3 m spatial resolution. Therefore, the primary objectives of this research were to: assess the feasibility of mapping allergenic tree species in the highly urbanized area using high-resolution PS imagery; evaluate and compare the performance of the most important machine learning and feature selection methods for accurate detection of individual allergenic tree species. The research incorporated three classification scenarios based on ground truth data: The first scenario (CS1) used single-date PS imagery with vegetation indices (VI), while the second and third scenarios (CS2 and CS3) used multitemporal PS imagery with VI, and GLCM and VI, respectively. The study demonstrated the feasibility of using multitemporal eight-band PlanetScope imagery to detect allergenic tree species, with the XGB method outperforming others with an overall accuracy of 73.13% in CS3. However, the classification accuracy varied between the scenarios and species, revealing limitations including the inherent heterogeneity of urban green spaces. Future research should integrate high-resolution satellite imagery with aerial photography or LiDAR data along with deep learning methods. This approach has the potential to classify dominant tree species in highly complex urban environments with increased accuracy, which is essential for urban planning and public health.https://www.mdpi.com/1999-4907/14/6/1193urban green arearandom forestneural networkextreme gradient boostingPlanetScopefeature selection |
| spellingShingle | Mateo Gašparović Dino Dobrinić Ivan Pilaš Mapping of Allergenic Tree Species in Highly Urbanized Area Using PlanetScope Imagery—A Case Study of Zagreb, Croatia Forests urban green area random forest neural network extreme gradient boosting PlanetScope feature selection |
| title | Mapping of Allergenic Tree Species in Highly Urbanized Area Using PlanetScope Imagery—A Case Study of Zagreb, Croatia |
| title_full | Mapping of Allergenic Tree Species in Highly Urbanized Area Using PlanetScope Imagery—A Case Study of Zagreb, Croatia |
| title_fullStr | Mapping of Allergenic Tree Species in Highly Urbanized Area Using PlanetScope Imagery—A Case Study of Zagreb, Croatia |
| title_full_unstemmed | Mapping of Allergenic Tree Species in Highly Urbanized Area Using PlanetScope Imagery—A Case Study of Zagreb, Croatia |
| title_short | Mapping of Allergenic Tree Species in Highly Urbanized Area Using PlanetScope Imagery—A Case Study of Zagreb, Croatia |
| title_sort | mapping of allergenic tree species in highly urbanized area using planetscope imagery a case study of zagreb croatia |
| topic | urban green area random forest neural network extreme gradient boosting PlanetScope feature selection |
| url | https://www.mdpi.com/1999-4907/14/6/1193 |
| work_keys_str_mv | AT mateogasparovic mappingofallergenictreespeciesinhighlyurbanizedareausingplanetscopeimageryacasestudyofzagrebcroatia AT dinodobrinic mappingofallergenictreespeciesinhighlyurbanizedareausingplanetscopeimageryacasestudyofzagrebcroatia AT ivanpilas mappingofallergenictreespeciesinhighlyurbanizedareausingplanetscopeimageryacasestudyofzagrebcroatia |