| Description | |||
| Abstract: Represents the urban forest tree canopy. The Urban Forest was defined from raster-based satellite imagery and processed into the City's Geodatabase as a vector feature class layer. The Urban Forest Project addresses a unique enhancement of City’s information related to urban vegetation, especially trees. In particular, this project introduces the science of Remote Sensing technology to the City of Richmond. Source Imagery provided by Space Imaging. Product Name: Geo 4m MS (4 meter resolution, Multispectral imagery). From Satellite: IKONOS. Date: 200307. | |||
| Purpose: Information derived from satellite imagery, such as the Urban Forest layer, is used for environmental sciences, general planning, and NOT for highly accurate mapping of trees or tree stands at fine scales. Therefore, this data should NOT be used at scales greater than 1:10,000. Be sure to read the Supplemental Information about the methodology and development of the Urban Forest. | |||
| Supplemental Information: Methodology: The primary data source of the project is IKONOS imagery. The imagery has 4-meter resolution and the following multispectral range: Band 1 0.45 – 0.52 micrometers Band 2 0.51 – 0.60 micrometers Band 3 0.63 – 0.70 micrometers Band 4 0.76 – 0.85 micrometers There were two main methods to derived tree canopy from the imagery – 1) unsupervised and 2) supervised classifications. Unsupervised classification is a process whereby the software computes, analyzes, and groups imagery cells together based upon same or similar light reflectance values that were recorded by the satellite sensor. Supervised classification is a process of more 'hands-on' approach to analyzing reflectance values. This is briefly discussed below; training samples. ERDAS Imagine software was used to classify the imagery. In order for the IKONOS satellite to capture complete coverage of the spatial extent of the City of Richmond, 3 orbiting passes had to be made. This resulted in 3 strips of imagery data. Step 1: The 3 original strips of imagery were delivered in TIFF format. These were converted into the native format of ERDAS Imagine software. Step 2: The 3 imagery strips were mosaiced together into a single image. A "training sample" is a defined geographic area within which there is a known type of land cover, such as trees, concrete, water, ect... From a training sample, you analyze the reflectance values of imagery cells (or pixels); these are measured in micrometers. Based upon these statistical samples, the technician can determine reflectance that would typically depict that land cover type. In our study, we were interested in locating training sample areas across the City which were obviously dominated by tree land cover. These areas were easy to delineate from orthophotography and from field visits. Step 3: 50 Training Samples were defined and light reflectance values were statistically analyzed. Step 4: Through a series of trial-and-error classifications of the data, a final classified image of land cover was developed. Step 5: The image was exported from the ERDAS Imagine format to a GIS "GRID" format, for work in Workstation and ArcGIS software. Step 6: The GRID was clipped to the spatial extent of the City of Richmond. Step 7: Through the science of map algebra techniques in GIS, the cells representing urban forest were extracted into a new GRID representing only that land cover type. Step 8: Finally, the GRID was converted from this 'raster' format, to a 'vector' format (shapefile) for loading into the geodatabase. Conclusions: Although there are many places that do not represent the tree cover precisely, generally, the ‘Urban Forest’ feature class derived from the imagery delineated tree canopies quite well. There are areas where pixel 'noise' disturbed the classification algorithm. For example, downtown areas or some reactance surfaces mixed the reflectance values and on some edges of ground features. Users of this data set will notice that as they zoom in very closely, the resolution of the original satellite imagery prevented it from being a dead-on accurate match with our more accurate GIS Basemap (derived from 1:100 scale). Information derived from satellite imagery, such as the Urban Forest layer, is used for environmental sciences, general planning, and NOT for highly accurate mapping of trees or tree stands at fine scales. Therefore, this data is best used for mapping of the urban forest at scales of 1:10,000. | |||
| Contact | |||
| Custodian: | Zbigniew Brodzik | ||
| Department: | Information Technology | ||
| Position: | GIS Project Manager | ||
| Address: | 900 E Broad St, Rm G2 | ||
| City: | Richmond | ||
| State: | Virginia | ||
| Zip Code: | 23219 | ||
| Email: | Zbigniew.Brodzik@richmondgov.com | ||
| Time Period of Content | |||
| Date: | 200307 | ||
| Status | |||
| Progress: | Complete | ||
| Update Frequency: | None planned | ||
| Spatial Domain | |||
| West Coordinate: | -77.602305 | ||
| East Coordinate: | -77.384506 | ||
| North Coordinate: | 37.604073 | ||
| South Coordinate: | 37.446261 | ||
| Spatial Data Information | |||
| Data Type: | vector digital data | ||
| Data Format: | SDE Feature Class | ||
| Data Projection: | Lambert Conformal Conic | ||
| Access and Usage Information | |||
| Access Constraints: | none | ||
| Use Constraints: | All GIS layers and datasets are owned by the City of Richmond and cannot be modified, re-used, re-distributed, and/or re-sold without a sub-licensing agreement. All guarantees of validity expire once a dataset leaves the City of Richmond firewall or physical premises. Acknowledgement of the City of Richmond GIS would be appreciated in products derived from these data. | ||
| Entity and Attribute Information | |||
| Entity Name: | vector.ric.veg_Canopy | ||
| Entity Type: | Feature Class | ||
| Entity Count: | 568736 | ||