New Approaches to Automated Object Recognition and Data Visualization Based on Laser Scanning and Photo Panoramas
The initial aim was to create a common geo-data platform for combining and displaying information about urban spaces. The major tasks were (1) to find approaches to visualize urban environment data in the most perceivable way and (2) to recognize objects from point cloud automatically to create a city model later and use it for practical tasks.
We process data from laser scanning and photo panoramas automatically and create a platform that combines information about the real surrounding city space with geo-data.
This fundamentally new three-dimensional form of visualization of urban space provides the following unique features:
▪ A coherent view of the surrounding objects on the map and panorama.
▪ One-click information about objects in information databases.
▪ A united database for every geo-referenced urban objects
▪ A united platform for a collection of all information about a city
▪ Inventory and control over the urban infrastructure objects and buildings.
▪ Measuring objects directly on panoramas (linear and area objects).
▪ Planning of territories by introducing new objects – embedding objects function. (Augmented reality)
▪ Visual monitoring of changes in the selected territory (retrospective analysis).
Unique algorithms for automatic recognition of objects of the urban environment and the formation of geo-information layers. Automatic recognition of road signs, poles, hatches, billboards, building facades, bus stops and other road infrastructure objects. Object Recognition adaptive algorithms help to recognize standard infrastructure objects with a high level of accuracy (up to 95%)