Google Earth Large Parking Finder
Budget: €250 – €750 EUR
I need a small, reliable program that automatically scans Google Earth imagery across Austria and Germany, flags every large, tree-free parking area it finds, and returns the information in an easy-to-review package. The core workflow I have in mind is: pull or stream tiles from Google Earth, run an image-analysis routine to detect sizable paved surfaces, filter out anything with visible tree cover, then save each approved hit.
For every location the software recognises, I want four data points captured:
• exact GPS coordinates
• overall dimensions (approximate length, width, total square metres)
• a short text description of the surrounding environment (e.g., “industrial estate, no residential housing within 200 m”)
•, when feasible, the registered owner of the plot
Please bundle the results as a folder that contains:
• a clear screenshot of the parking lot pulled from Google Earth
- a picture of streetview or 3d
• a companion file (CSV, JSON, or similar) with the coordinate pair, address, and the other details above
Screenshots and their matching data must share a common ID so I can cross-reference them quickly.
Acceptance criteria
• Algorithm correctly ignores areas with even sparse tree cover.
• At least 90 % precision when tested on a random 100-site sample I provide.
• Code is delivered with a short README explaining setup, dependencies, and how to extend the scan to additional countries later.
Preferred languages are Python or JavaScript, but I am open to whatever lets you leverage Google Earth Engine, Google Maps Static API, or comparable libraries efficiently.
For every location the software recognises, I want four data points captured:
• exact GPS coordinates
• overall dimensions (approximate length, width, total square metres)
• a short text description of the surrounding environment (e.g., “industrial estate, no residential housing within 200 m”)
•, when feasible, the registered owner of the plot
Please bundle the results as a folder that contains:
• a clear screenshot of the parking lot pulled from Google Earth
- a picture of streetview or 3d
• a companion file (CSV, JSON, or similar) with the coordinate pair, address, and the other details above
Screenshots and their matching data must share a common ID so I can cross-reference them quickly.
Acceptance criteria
• Algorithm correctly ignores areas with even sparse tree cover.
• At least 90 % precision when tested on a random 100-site sample I provide.
• Code is delivered with a short README explaining setup, dependencies, and how to extend the scan to additional countries later.
Preferred languages are Python or JavaScript, but I am open to whatever lets you leverage Google Earth Engine, Google Maps Static API, or comparable libraries efficiently.