Scene models

Supported data formats

GanosBase Scene provides two native 3D data types: glTF and OSGB. It also supports the GLB binary variant of glTF as a compatible format.

Spatial reference system

A spatial reference system (SRS) defines how a Scene object maps to a location on Earth's surface. GanosBase uses an integer SRID to reference an SRS definition. Each Scene object carries its own SRID value.

For details, see Spatial reference.

scene_columns view

The scene_columns view lists all Scene-type columns in the current database. It follows the Open Geospatial Consortium (OGC) Simple Features Specification for SQL and can be extended.

To list all Scene-type columns in the current database:

SELECT * FROM scene_columns;

Spatial index

GanosBase provides Generalized Search Tree (GiST) indexes for Scene data. GiST operates on bounding boxes, which makes it well suited for spatial relationship queries—such as intersection and inclusion—on large 3D datasets, without requiring a full sequential scan.

To create a GiST index:

CREATE INDEX <index_name>
ON <table_name>
USING GIST(<column_name>);

Oblique photography management and visualization

Oblique photography is the primary source of 3D basemap data in surveying and mapping. GanosBase Scene imports, stores, and processes oblique photography data natively. Store raw data in Object Storage Service (OSS) and build Scene types in the database—this separates storage from compute and avoids data redundancy while keeping spatial queries, analysis, and rendering cost-effective.

Spatial data mining

GanosBase Scene supports cross-model spatial relationship queries. Common use cases include:

• Detecting whether buildings overlap (3D intersection)

• Finding all buildings within a geographic range (spatial inclusion)

• Extracting geometry and surface attributes of a specific structure—such as a building on a campus—for downstream analysis

Campus and urban planning

GanosBase Scene runs advanced spatial analysis entirely in the database, reducing GPU resource consumption:

• Visibility analysis: determine camera placement on a campus, or assess sightlines for event security planning

• Shadow ratio analysis: evaluate solar exposure for building or solar power station site selection

Lightweight preprocessing functions—3D model simplification, and texture clipping and simplification—reduce model complexity before frontend rendering, improving performance for upper-layer applications.

Prerequisites

Before you begin, make sure you have:

• A PolarDB for PostgreSQL (Compatible with Oracle) instance

• Superuser or extension-creation privileges on the target database

Create the extension

CREATE EXTENSION Ganos_scene CASCADE;

CREATE EXTENSION Ganos_scene WITH SCHEMA public CASCADE;

Create a table

CREATE TABLE t_scene (
  id   integer,
  data scene
);

Insert data

Insert three glTF scene objects into the table:

INSERT INTO t_scene (id, data)
VALUES
  (1, '{"type" : "gltf", "affine" : [1,0,0,0,0,1,0,0,0,0,1,0,0,0,0,1], "content" : {"accessors":[{"bufferView":0,"componentType":5126,"count":42,"max":[1.0,1.0,1.0],"min":[-1.0,-1.0,-1.0],"type":"VEC3"},{"bufferView":1,"componentType":5126,"count":42,"max":[1.0,1.0,1.0],"min":[-1.0,-1.0,-1.0],"type":"VEC3"},{"bufferView":2,"componentType":5126,"count":42,"max":[0.0],"min":[0.0],"type":"SCALAR"},{"bufferView":3,"componentType":5121,"count":240,"max":[41],"min":[0],"type":"SCALAR"}],"asset":{"generator":"Ganos Scene","version":"2.0"},"bufferViews":[{"buffer":0,"byteLength":504,"target":34962},{"buffer":0,"byteLength":504,"byteOffset":504,"target":34962},{"buffer":0,"byteLength":168,"byteOffset":1008,"target":34962},{"buffer":0,"byteLength":240,"byteOffset":1176,"target":34963}],"buffers":[{"byteLength":1416,"uri":"data:application/octet-stream;base64,AAAAAAAAgD8AAAAAAAAAP70bTz96N54+AAAAP70bTz96N56+AAAAAEDEWT9Qlga/AAAAAEDEWT9QlgY/QMRZP1CWBj8AAAAAAAAAv70bTz96N56+AAAAv70bTz96N54+QMRZv1CWBj8AAAAAAAAAAAAAgL8AAAAAAAAAP70bT796N56+AAAAP70bT796N54+AAAAAEDEWb9QlgY/AAAAAEDEWb9Qlga/QMRZP1CWBr8AAAAAAAAAv70bT796N54+AAAAv70bT796N56+QMRZv1CWBr8AAAAAAACAPwAAAAAAAAAAvRtPP3o3nr4AAAC/vRtPP3o3nj4AAAC/UJYGPwAAAABAxFm/vRtPP3o3nj4AAAA/vRtPP3o3nr4AAAA/UJYGPwAAAABAxFk/AACAvwAAAAAAAAAAvRtPv3o3nr4AAAA/vRtPv3o3nj4AAAA/UJYGvwAAAABAxFk/vRtPv3o3nj4AAAC/vRtPv3o3nr4AAAC/UJYGvwAAAABAxFm/AAAAAAAAAAAAAIA/ejeevgAAAL+9G08/ejeePgAAAL+9G08/ejeePgAAAD+9G08/ejeevgAAAD+9G08/AAAAAAAAAAAAAIC/ejeevgAAAD+9G0+/ejeePgAAAD+9G0+/ejeePgAAAL+9G0+/ejeevgAAAL+9G0+/AAAAAAAAgD8AAAAAAQAAP70bTz96N54+AQAAP70bTz96N56+AAAAAEDEWT9Rlga/AAAAAEDEWT9RlgY/QMRZP1KWBj8AAAAAAQAAv70bTz96N56+AQAAv70bTz96N54+QMRZv1KWBj8AAAAAAAAAAAAAgL8AAAAAAQAAP70bT796N56+AQAAP70bT796N54+AAAAAEDEWb9RlgY/AAAAAEDEWb9Rlga/QMRZP1KWBr8AAAAAAQAAv70bT796N54+AQAAv70bT796N56+QMRZv1KWBr8AAAAAAACAPwAAAAAAAAAAvRtPP3o3nr4BAAC/vRtPP3o3nj4BAAC/UpYGPwAAAABAxFm/vRtPP3o3nj4BAAA/vRtPP3o3nr4BAAA/UpYGPwAAAABAxFk/AACAvwAAAAAAAAAAvRtPv3o3nr4BAAA/vRtPv3o3nj4BAAA/UpYGvwAAAABAxFk/vRtPv3o3nj4BAAC/vRtPv3o3nr4BAAC/UpYGvwAAAABAxFm/AAAAAAAAAAAAAIA/ejeevgEAAL+9G08/ejeePgEAAL+9G08/ejeePgEAAD+9G08/ejeevgEAAD+9G08/AAAAAAAAAAAAAIC/ejeevgEAAD+9G0+/ejeePgEAAD+9G0+/ejeePgEAAL+9G0+/ejeevgEAAL+9G0+/AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAECAwACBAEABQIBAAYHBAAHAwYACAcGCQoLDAkLDQoJDgsKCQ8QDQkQDA8JERAPEhMUBRIUDhMSFRQTEhYXDhIXBRYSGBcWGRobCBkbERoZHBsaGR0eERkeCB0ZHx4dICEiGCAiHCEgDCIhICMkHCAkGCMgBCQjJSYnFSUnHyYlAycmJSgpHyUpFSglDSkoIxYBBCMBGBYjBQEWBxskBAckCBsHHCQbAhQnAwInBRQCFScUJh0GAyYGHx0mCAYdCxciDAsiDhcLGCIXIRoPDCEPHBohEQ8aKBMKDSgKFRMoDgoTEB4pDRApER4QHyke"}],"meshes":[{"primitives":[{"attributes":{"NORMAL":1,"POSITION":0,"_USER_1":2},"indices":3,"mode":4}]}],"nodes":[{"mesh":0}],"scenes":[{"nodes":[0]}]}}'::scene),
  (2, '{"type" : "gltf", "affine" : [1,0,0,0,0,1,0,0,0,0,1,0,0,0,0,1], "content" : {"accessors":[{"bufferView":0,"componentType":5121,"count":36,"max":[7],"min":[0],"type":"SCALAR"},{"bufferView":1,"componentType":5126,"count":8,"max":[2.5,0.5,0.5],"min":[1.5,-0.5,-0.5],"type":"VEC3"}],"asset":{"generator":"ganos","version":"2.0"},"bufferViews":[{"buffer":0,"byteLength":36,"target":34963},{"buffer":0,"byteLength":96,"byteOffset":36,"target":34962}],"buffers":[{"byteLength":132,"uri":"data:application/octet-stream;base64,AgEAAQIDBAIAAgQGAQQABAEFBgUHBQYEAwYHBgMCBQMHAwUBAADAPwAAAL8AAAC/AAAgQAAAAL8AAAC/AADAPwAAAD8AAAC/AAAgQAAAAD8AAAC/AADAPwAAAL8AAAA/AAAgQAAAAL8AAAA/AADAPwAAAD8AAAA/AAAgQAAAAD8AAAA/"}],"meshes":[{"primitives":[{"attributes":{"POSITION":1},"indices":0,"mode":4}]}],"nodes":[{"mesh":0}],"scenes":[{"nodes":[0]}]}}'::scene),
  (3, '{"type" : "gltf", "affine" : [1,0,0,0,0,1,0,0,0,0,1,0,0,0,0,1], "content" : {"accessors":[{"bufferView":0,"componentType":5121,"count":36,"max":[7],"min":[0],"type":"SCALAR"},{"bufferView":1,"componentType":5126,"count":8,"max":[6.5,0.5,0.5],"min":[5.5,-0.5,-0.5],"type":"VEC3"}],"asset":{"generator":"ganos","version":"2.0"},"bufferViews":[{"buffer":0,"byteLength":36,"target":34963},{"buffer":0,"byteLength":96,"byteOffset":36,"target":34962}],"buffers":[{"byteLength":132,"uri":"data:application/octet-stream;base64,AgEAAQIDBAIAAgQGAQQABAEFBgUHBQYEAwYHBgMCBQMHAwUBAACwQAAAAL8AAAC/AADQQAAAAL8AAAC/AACwQAAAAD8AAAC/AADQQAAAAD8AAAC/AACwQAAAAL8AAAA/AADQQAAAAL8AAAA/AACwQAAAAD8AAAA/AADQQAAAAD8AAAA/"}],"meshes":[{"primitives":[{"attributes":{"POSITION":1},"indices":0,"mode":4}]}],"nodes":[{"mesh":0}],"scenes":[{"nodes":[0]}]}}'::scene);

Query scene data

Retrieve all rows as text:

SELECT id, ST_AsText(data)
FROM t_scene;

Create a spatial index

Create a GiST index on the data column to accelerate spatial queries:

CREATE INDEX idx_t_scene
ON t_scene
USING GIST(data);

Run spatial queries

All examples below use ST_3DMakeCuboid(3, 3, 3) as the query bounding box.

Find objects that intersect with a 3D bounding box:

SELECT id
FROM t_scene
WHERE ST_3DIntersects(data, ST_3DMakeCuboid(3, 3, 3));
-- Returns: 2

Find objects fully contained within a 3D bounding box:

SELECT id
FROM t_scene
WHERE ST_Contains(ST_3DMakeCuboid(3, 3, 3), data);
-- Returns: 1

Run spatial computations

Compute the intersection geometry between a scene object and a bounding box:

SELECT ST_AsText(ST_Intersection(data, ST_3DMakeCuboid(3, 3, 3)))
FROM t_scene
WHERE id = 2;

Merge multiple glTF scene objects into one:

SELECT ST_AsText(ST_Combine(data))
FROM t_scene;

Remove the extension

To remove the extension and all associated objects:

DROP EXTENSION Ganos_Scene CASCADE;