LY/T 3184-2020Technical specification for virtual 3D forest map making (English PDF)

4.1 Coordinates and projection requirements

4.1.1 Coordinate system

Using.2000 geodetic coordinate system.

4.1.2 Projection method

Using Gauss-Krüger projection;

1.5000 scale is divided by 3°;

1.10000 scale is divided by 3°.

4.1.3 Elevation system

The 1985 national elevation datum is used.

4.2 Remote sensing data requirements

4.2.1 Spatial resolution

The resolution of remote sensing images (aerial photos) should be higher than 2.5m×2.5m.

4.2.2 Data requirements

The data should meet the following requirements.

--The cloud cover area is less than 5%, and can cover the main work area;

--The image is clear, rich in layers, uniform in tone and moderate in contrast;

--The lateral viewing angle of data reception should generally be less than 15°, not more than 25° in plain areas, and not more than 20° in mountainous areas;

- There is no banding, speckle noise, row loss, etc. in the data;

--The shooting time of the image data should be selected in the season when the vegetation grows vigorously.

4.3 Topographic map data requirements

Collect the latest electronic version or paper version with a scale not less than 1.10000 national standard basic topographic maps. Topographic map data requirements should be include.

--The paper topographic map should be checked for its surface flatness, whether the symbols, lines, outline points and kilometer nets are complete, and whether the Connection situation Paper topographic maps need to be scanned and corrected. The scanning resolution is 300dpi, stored in TIF format, and raster topographic maps are corrected by the geometric precision correction method is used to calibrate, and the intersection point is selected as the control point of each kilometer grid, and the raster topographic map is corrected to the grid consistent with its coordinate system and projection.

Grid image, and then vectorized.

--The inspection of electronic topographic maps shall be carried out in accordance with GB/T 18316-2008.

4.4 Elevation model data requirements

4.4.1 Elevation model data

Use vectorized topographic maps to generate corresponding scale digital elevation models; or use aerial stereo images to build digital elevation models.

4.4.2 Data quality requirements

Data quality should meet the following requirements.

--There is no crack phenomenon after the adjacent framing DEM is connected;

--The elevation value of the overlapping area should be consistent;

--DEM has complete metadata, and records the coordinate system, precision, grid sampling size and other information of the data in detail;

--1.The grid size of 10000 DEM should not be less than 12.5m, the location of elevation data should reach 0.1m, and the DEM accuracy should reach CH/T 1008 Requirements.

4.5 Basic data requirements for 3D tree modeling

4.5.1 Tree requirements

The selected trees should be representative, have good growth conditions, conform to normal growth rules, and have no undesirable growth phenomena such as broken shoots.

The surrounding weeds should not be higher than 0.5m.

4.5.2 Digital camera requirements

The pixels are greater than 8 million.

4.5.3 Photo quality requirements

The photo quality should meet the following requirements.

--The photo format should be JPEG;

--The photos cover the panoramic view of standing wood, and there will be no cracks after adjacent photos are joined;

--The photos are clear, without stains, overexposure, etc.

5.1 Remote sensing image processing

5.1.1 Image orthorectification

Based on the control point data and DEM, orthorectify panchromatic images and multispectral images.

5.1.2 Geometric precision correction

Eliminate the geometric distortion of the image and realize the geometric integration process of the original image and the standard image or map.

5.1.3 Image Fusion

Using PCA, HSI and other image fusion algorithms, the high-resolution panchromatic band image and the lower-resolution multi-spectral band image are combined Line fusion to highlight forest vegetation information.

5.1.4 Band combination

Select the corresponding bands from the fused multispectral data after orthorectification and combine to form true and false color images.

5.1.5 Image enhancement

Using linear stretching, HIS and other image enhancement algorithms, the fused image is enhanced to make the image rich in color and layered Bright, clear texture, highlighting forest vegetation information.

5.1.6 Image mosaic

Combine multiple adjacent remote sensing images into a large-scale, seamless image.

5.2 Three-dimensional model of trees

5.2.1 Color

The colors of leaves, branches, and trunks of the tree model are based on the actual colors.

5.2.2 Tree model

The tree model should have a branching system, and the leaves should be randomly distributed. The tree model should conform to the normal growth pattern, without broken shoots, broken branches, etc.

Build tree models according to different tree species and age groups.

5.2.3 Trunk branching system

There is a parent-child relationship between each trunk and branches, where the first-level branch is attached to the parent trunk, and each first-level branch can only have one parent Trunk, a parent trunk can have multiple first-level branches, and first-level branches can have no second-level branches.

5.2.4 Leaf branch system

Each small branch structure can contain 3 or 4 leaf structures.

5.2.5 Texture of bark

The textures of the trunk and branches should be consistent and mapped with the real bark texture.

5.3 Virtual 3D terrain model

5.3.1 3D terrain modeling 3D terrain modeling adopts regular grid model and irregular triangulated grid model.

--Regular grid model. a model constructed by dividing the area into a regularly arranged grid according to a certain unit size, which is mostly suitable for Terrain modeling in relatively flat areas.

--Irregular triangle network model. connect topographic feature points according to certain rules into triangles that cover the entire area and do not overlap each other the model built can accurately represent the terrain features, and is mostly suitable for terrain modeling in mountainous and hilly areas with more complex terrain.

5.3.2 Optimization of 3D terrain model

The terrain model is optimized by the method of partition modeling and hierarchical modeling.

--Division modeling. According to the characteristics of the terrain, establish terrain models of different levels of detail for different regions. In areas with complex terrain, establish High-precision terrain model; in flat terrain, build a lower-precision terrain model.

-Hierarchical modeling. According to the viewpoint position, build terrain models of different complexity for the same area. Use LOD (Level of Detail) when the viewpoint is far away from the area, a simpler terrain model is used. When the viewpoint is closer to the area, a more complicated Terrain model representation.

6.1 Requirements for the basic database

Refer to LY/T 1662.7 for database data content, data structure, file naming, data layering, graphics and attribute data content, etc.

-2008 "Digital Forestry Standards and Norms".

6.2 Requirements for auxiliary databases

The auxiliary database mainly includes data dictionary and sub-administrative district index map data.

--Data Dictionary. For data standardization, high efficiency and maintainability in data management, forest resource planning and design are usually adjusted the related attribute field names and field values and data descriptions in the search data are defined in a uniformly prescribed form and a definition database is established.

---Sub-administrative district index map data. In order to improve the efficiency of data retrieval when managing spatial data, a spatial data index needs to be established.

7.1 Basic data of forest diagram

Use the latest forest resource planning and design survey data as the basic data.

7.2 Small class tree species

Based on the dominant tree species in the forest resource planning and design survey data, different age groups should have different models.

7.3 Number of small teams

The number of trees per unit area of the sub-class is used to control the density of trees in the sub-class, and the area of the sub-class is used to control the number of trees in the entire sub-class, and the tree model should be randomly distributed in small classes.

7.4 Base elevation of tree position

The elevation of the location of the tree is obtained from the digital elevation model.

7.5 Making a virtual three-dimensional forest map

7.5.1 Vectorized topographic map

Correct the scale of the latest electronic version or paper version that is not less than 1.10000 national standard topographic map, and then enter it after processing.

Line vectorization to obtain vectorized topographic map.

7.5.2 Digital elevation model data processing

Use vectorized topographic maps to generate corresponding scale digital elevation model data, or directly use aerial stereo images to construct digital elevation the model gets the data.

7.5.3 Remote sensing image processing

Ortho-correction, geometric precision correction, image fusion, band combination, image enhancement, image mosaic and other processing of remote sensing images.

7.5.4 Aerial Photo Processing

Perform geometric precision correction and image mosaic processing on aerial photos.

7.5.5 Virtual 3D terrain modeling

Use the processed remote sensing images (aerial photos) and the generated digital elevation model to superimpose processing to generate a virtual three-dimensional terrain model.

7.5.6 Three-dimensional modeling of trees

Take representative images of trees on the spot, and use 3D modeling software to build 3D models of trees.

7.5.7 Establishment of virtual three-dimensional forest diagram

Using the established virtual three-dimensional terrain model and tree three-dimensional model, each sub-class represents the tree species of the entire sub-class with dominant tree species.

The number of trees per unit area of the sub-class and the area control the density and total number of trees in the sub-class, and the trees in the sub-class should be randomly distributed in the small in the class, through superposition and other processing, a virtual three-dimensional forest diagram is established.