Regional countrywide model;
3D models of almost all large cities;
Geo-database with semantic infrastructure information.
The idea of the project was to develop and optimize the 4G telecommunication network countrywide, which is supported and promoted by the biggest mobile operators of Vietnam.
Within the project it was planned to do a considerable amount of work whereby the high-accuracy geodata have to be analyzed and processed in a shortest period of time: 3D maps of 74 Vietnam cities totaling 5053 sqkm, regional map for the whole Vietnam territory, i.e. 331210 sqkm, which would meet the requirements of 2D city models with detailed building development. Additionally for some biggest cities it was envisioned to capture and accumulate database of infrastructure data, so called Points of Interest (POI).
The geo-database collected under this project gives a holistic and detailed information regarding the built-up territory allowing development and optimization of 4G telecommunication network and further deployment and implementation of 5G.
In order to complete the project and ensure the required accuracy it was bought over 95 stereo-pairs of satellite imageries with 0.3–0.5 m/pixel resolution. Accurate and reliable spatial orientation of them was made based on the set of 200 GCPs (Ground Control Points), which were captured within the GPS survey. The positioning accuracy of the measured points is about 0.1 meter which complies the project demands. Taking into account the precision of orbital data the final spatial orientation of the images is not worth than 1–3 meters.
As an inevitable circumstances, we had to take into account during the satellite imageries processing, was a significant cloud cover owing to the monsoon climate of the country. To handle the high quality and clear satellite images it was necessary to use the special algorithm allowing to search the cloudiness places and automatically replace them with the corresponded cloudless areas taken from others images. Sometimes it was required to order a direct stereo shooting to avoid the cloud cover areas.
All the work under the project was done mostly in manual and on rare occasions in semiautomatic modes plotting totally collecting over 12,000,000 building contours. The vegetation layers and their heights were captured using the recognition algorithms which are based on the artificial intelligence (AI), in particular neutral networks. Due to the enormous amount of data this approach got a remarkable approbation and as a result has been used with high reliability and effectiveness.
Each project stage was controlled at seven various levels to provide the required accuracy and credibility.
The geo-database might be used further across multiple industries where relevant and up-to-date information regarding the built-up territory is needed: urban planning environmental management, transportation, risk mitigation, emergency response, infrastructure rollout, etc.
Visit our web site visicomdata.com to learn more about our geospatial products for Telecom and GIS markets.
The data samples are available for free https://visicomdata.com/samples?lang=en#vietnam