A traffic infrastructure ecological coordination selection and division method and device

Abstract

This invention discloses a method and apparatus for ecologically coordinated planning of transportation infrastructure, comprising: collecting multi-source ecological planning image data and traffic remote sensing image data, and unifying spatial reference through GIS geographic registration; adaptively adjusting segmentation parameters based on image characteristics to perform multi-scale segmentation and obtain vector graphic data of the area to be planned; superimposing vector data to calculate an eco-development index value to quantify the suitability of ecological protection and development; and finally selecting a site based on the index value to achieve coordinated optimization of transportation infrastructure and ecological protection. This method, through multi-source data fusion and quantitative evaluation, improves the scientific rigor and ecological adaptability of the planning results, and is applicable to the ecologically coordinated planning and site selection of basic transportation facilities.

Description

Technical Field

[0001] This invention relates to the field of infrastructure construction, and more specifically, to a method and apparatus for ecological collaborative planning of transportation infrastructure. Background Technology

[0002] Transportation infrastructure construction is a crucial support for regional economic development. However, traditional site selection methods often prioritize engineering feasibility and economic benefits, neglecting systematic consideration of ecological protection. This can easily lead to problems such as encroachment on ecologically sensitive areas and damage to biodiversity. In existing technologies, ecological planning data and transportation remote sensing data are often difficult to analyze collaboratively due to data barriers and a lack of quantitative indicators to balance the needs of ecological protection and development. This results in highly subjective site selection results and low efficiency in conflict resolution. Summary of the Invention

[0003] The purpose of this invention is to provide a method and apparatus for ecological collaborative planning of transportation infrastructure.

[0004] In a first aspect, embodiments of the present invention provide a method for ecologically coordinated planning of transportation infrastructure, comprising:

[0005] Collect multi-source ecological planning image data and traffic remote sensing image data, and perform GIS georegistration on the multi-source ecological planning image data to convert it into registered ecological planning image data with spatial reference;

[0006] Based on the image characteristics of the registered ecological planning data and the traffic remote sensing image data, the segmentation parameters are adaptively adjusted to perform multi-scale segmentation, and vector graphic data corresponding to each of the multiple candidate areas are obtained.

[0007] By overlaying multiple vector graphic data, the ecological-development index values ​​corresponding to each of the multiple candidate areas are calculated.

[0008] Based on the ecological-development index value, the site selection of transportation infrastructure is carried out for the multiple candidate areas, and the results of the transportation infrastructure selection for the multiple candidate areas are obtained.

[0009] In one possible implementation, the step of adaptively adjusting segmentation parameters based on the image characteristics of the registered ecological planning data and the traffic remote sensing image data to perform multi-scale segmentation and obtain vector graphic data corresponding to multiple candidate areas includes:

[0010] Based on the planning map features of the registered ecological planning data, vector graphic data corresponding to the registered ecological planning data of the multiple candidate areas are obtained by optimizing color blocks and text filtering, and by integrating morphological filtering to eliminate legend label interference.

[0011] Based on the remote sensing image features of the traffic remote sensing image data, vector graphic data corresponding to the traffic remote sensing image data of the multiple candidate areas are obtained through texture enhancement and edge detection.

[0012] In one possible implementation, the step of overlaying multiple vector graphic data to calculate the ecological-development index values ​​corresponding to each of the multiple candidate delineation areas includes:

[0013] Through the formula: The ecological-development index value corresponding to each of the candidate areas is calculated;

[0014] in, The ecological-development index value. Let be the weight of the planning legal effect corresponding to the i-th vector graphic data of the candidate area. This represents the base score corresponding to the i-th vector graphic data of the candidate region;

[0015] If there are multiple overlapping vector graphics data in the area to be selected, the minimum base score among the multiple overlapping vector graphics data shall be taken as the base score of the area.

[0016] In one possible implementation, the step of selecting transportation infrastructure sites for the plurality of candidate areas based on the eco-development index value, to obtain the transportation infrastructure selection results for the plurality of candidate areas, includes:

[0017] Based on the ecological-development index value, site selection for land-based shoreline facilities and water-based facilities are carried out in the multiple candidate areas to obtain site selection results for either land-based shoreline facilities or water-based facilities.

[0018] In one possible implementation, the site selection for land-based shoreline facilities is performed on the plurality of candidate areas based on the ecological-development index value, resulting in the land-based shoreline facility site selection and delineation results, including:

[0019] Multiple consecutive candidate areas with an ecological-development index value greater than a first preset index threshold are designated as candidate areas, and candidate areas with an ecological-development index value less than a second preset index threshold are designated as avoidance areas; the avoidance areas are used to determine buffer avoidance areas within a preset range.

[0020] By optimizing the function: Calculate the planning suitability index for each of the proposed areas, and take the proposed area corresponding to the largest planning suitability index as the site selection result for the land shoreline facilities.

[0021] Where TF is the planning suitability index, and k is the number of planned transportation infrastructure facilities. Let EDI be the value of the k-th suitable region. Let k be the length or area of ​​the k-th suitable region. Let α be the construction cost of the kth transportation infrastructure, and let α be the cost coefficient for reconstruction and expansion when existing infrastructure is occupied.

[0022] In one possible implementation, the site selection and delineation results for waterway facilities include waterway delineation results. Based on the ecological-development index value, waterway facility sites are selected from the multiple candidate areas to obtain the waterway facility site selection and delineation results, including:

[0023] Acquire electronic nautical chart depth point cloud data;

[0024] Based on the electronic nautical chart water depth point cloud data, multiple high-precision water depth grids are generated by Kriging interpolation.

[0025] Extract continuous regions that meet the minimum navigable water depth threshold from the multiple high-precision water depth grids, and remove broken waterways whose length or width does not meet the preset distance.

[0026] The continuous region is abstracted as a network node. The node spacing is defined as connected if it is less than a preset spacing threshold. The inverse of the EDI exponent corresponding to the continuous region is used as the edge weight. The Dijkstra algorithm is used to calculate the minimum cumulative weight path from the preset starting point to the ending point.

[0027] An adaptive width corridor is generated along the path with the lowest cumulative weight. For the route segment that crosses the ecologically sensitive area, the route is oscillated within the adaptive width corridor to determine the route selection result.

[0028] In one possible implementation, the site selection and delineation results for the water facilities include anchorage delineation results. Based on the ecological-development index value, water facility site selection is performed on the multiple candidate areas to obtain the water facility site selection and delineation results, including:

[0029] Acquire electronic nautical chart depth point cloud data;

[0030] Based on the electronic nautical chart water depth point cloud data, multiple high-precision water depth grids are generated by Kriging interpolation.

[0031] The continuous regions that meet the minimum anchoring depth threshold are extracted from the multiple high-precision water depth grids, and broken water areas with an area smaller than the preset area threshold are removed to obtain the basic navigation area.

[0032] From the basic navigation area, areas with an ecological-development index greater than a third preset index threshold are selected as preferred anchorage areas, and the anchorage selection results are obtained by sorting them in descending order.

[0033] In one possible implementation, the method further includes:

[0034] A simulation model of the ecological impact of the selected planning scheme is constructed using generative adversarial networks. The vector graphic data corresponding to the selected planning results of the transportation infrastructure is used as the input of the generator. The generated simulated regional ecological landscape is compared with the historical ecological baseline data by a discriminator to calculate the similarity of ecological impact.

[0035] If the similarity is lower than the preset threshold, the boundary range of the delineation scheme is adjusted based on the gradient information fed back by the discriminator until the ecological impact similarity reaches the preset threshold, and the final delineation result is obtained.

[0036] In one possible implementation, the method further includes:

[0037] By constructing an ecological-transportation collaborative sensing network, real-time access is provided to ecological sensor data and traffic flow prediction data of the candidate area.

[0038] Based on the ecological sensor data and the traffic flow prediction data, areas where the short-term fluctuations of the associated indicators exceed a preset range are marked as temporary restricted areas. After the fluctuations of the indicators in the temporary restricted areas stabilize, the corresponding traffic infrastructure selection results are re-evaluated.

[0039] Secondly, embodiments of the present invention provide a transportation infrastructure ecological collaborative planning device, comprising:

[0040] The acquisition module is used to collect multi-source ecological planning image data and traffic remote sensing image data, and to perform GIS georegistration on the multi-source ecological planning image data to convert it into registered ecological planning image data with spatial reference; based on the image characteristics of the registered ecological planning data and the traffic remote sensing image data, the segmentation parameters are adaptively adjusted to perform multi-scale segmentation to obtain vector graphic data corresponding to multiple candidate delineation areas.

[0041] The delineation module is used to overlay multiple vector graphic data to calculate the ecological-development index value corresponding to each of the multiple candidate delineation areas; and to select the location of transportation infrastructure for the multiple candidate delineation areas based on the ecological-development index value to obtain the transportation infrastructure delineation results for the multiple candidate delineation areas.

[0042] Compared to existing technologies, the beneficial effects provided by this invention include: employing the ecological collaborative planning method and apparatus for transportation infrastructure disclosed in this invention, multi-source ecological planning image data and transportation remote sensing image data are collected and unified spatial reference is established through GIS geographic registration; based on image characteristics, segmentation parameters are adaptively adjusted to perform multi-scale segmentation to obtain vector graphic data of the area to be planned; the vector data is overlaid to calculate the eco-development index value, quantifying the suitability for ecological protection and development; finally, site selection is performed based on the index value, achieving synergistic optimization of transportation infrastructure and ecological protection. This method, through multi-source data fusion and quantitative evaluation, improves the scientific rigor and ecological adaptability of the planning results, and is applicable to the planning and site selection of basic transportation facilities. Attached Figure Description

[0043] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the embodiments will be briefly described below. It should be understood that the following drawings only show some embodiments of the present invention and should not be considered as limiting the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0044] Figure 1 A flowchart illustrating the steps of the ecological collaborative planning method for transportation infrastructure provided in this embodiment of the invention;

[0045] Figure 2 A flowchart illustrating the multi-source image data acquisition and registration process provided in this embodiment of the invention;

[0046] Figure 3 A schematic block diagram of the structure of the transportation infrastructure ecological collaborative planning device provided in an embodiment of the present invention;

[0047] Figure 4 A schematic block diagram of the structure of a computer device provided in an embodiment of the present invention. Detailed Implementation

[0048] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. The components of the embodiments of the present invention described and shown in the accompanying drawings can generally be arranged and designed in various different configurations.

[0049] The specific embodiments of the present invention will now be described in detail with reference to the accompanying drawings.

[0050] In order to solve the technical problems mentioned in the background art Figure 1This is a flowchart illustrating the ecological collaborative planning method for transportation infrastructure provided in this embodiment. The following is a detailed description of the ecological collaborative planning method for transportation infrastructure.

[0051] Step S201: Collect multi-source ecological planning image data and traffic remote sensing image data, and perform GIS georegistration on the multi-source ecological planning image data to convert it into registered ecological planning image data with spatial reference.

[0052] Step S202: Based on the image characteristics of the registered ecological planning data and the traffic remote sensing image data, the segmentation parameters are adaptively adjusted to perform multi-scale segmentation, and vector graphic data corresponding to each of the multiple candidate areas are obtained.

[0053] Step S203: Overlay multiple vector graphic data to calculate the ecological-development index value corresponding to each of the multiple candidate areas;

[0054] Step S204: Based on the ecological-development index value, select the location of transportation infrastructure for the multiple candidate areas to obtain the transportation infrastructure selection results for the multiple candidate areas.

[0055] In this embodiment of the invention, an exemplary application scenario is the ecological collaborative planning of a port and port access roads in a coastal city in eastern my country (hereinafter referred to as the "target city"). The server acts as the execution entity, achieving coordinated optimization of transportation facility site selection and ecological protection through multi-source data fusion, intelligent segmentation, and quantitative evaluation. The target area covers land (including urban built-up areas, farmland, and nature reserves), nearshore waters (including tidal flats, aquaculture areas, and port operation areas), and inland waterways, with a total area of ​​approximately 1500 square kilometers. The key areas for planning include the construction of a new 300,000-tonnage deep-water channel, a 20-square-kilometer port logistics park, and a 50-kilometer port access road.

[0056] The server first starts the distributed data acquisition engine, acquires ecological planning and transportation remote sensing data of the target area through the following channels, and completes spatial benchmark unification:

[0057] Data collection phase:

[0058] Ecological planning data: The server collects and organizes relevant planning maps published by the target city, such as the city's overall land space planning map (including the city's coastal zone distribution map, the city's marine functional zoning map, the city's ecological protection red line map, the city's natural protected area distribution map, the city's permanent basic farmland distribution map, etc.), the city-level ecological environment zoning control plan map, and the nearshore marine functional zoning map, etc.

[0059] Traffic remote sensing data: Multispectral data from Landsat-9 (30-meter resolution, cloudless imagery from summer 20XX) and Sentinel-2 (10-meter resolution, imagery from autumn 20XX) were acquired through a satellite remote sensing data service platform, including eight bands: blue, green, red, and near-infrared. Simultaneously, OpenStreetMap online service address was accessed to obtain vector surface data of the current port terminals (including coordinates of 20 berths of 10,000 tons or more), vector line data of the port access road network (design speed 60-80 km / h), and electronic nautical charts of inland waterways (including depth points and navigation mark coordinates).

[0060] GIS georeferencing process:

[0061] The server calls the georeferencing module of the GDAL library to perform coordinate transformation on the non-vector format data, using the 2000 National Geodetic Coordinate System (CGCS2000) and Gauss-Kruger projection (3-degree zone, central meridian 121°E) as the spatial reference.

[0062] For paper-scanned nature reserve planning maps, by identifying three known coordinate control points in the map (e.g., triangulation point P1: X=3852000m, Y=512000m; road intersection P2: X=3853500m, Y=513200m), a quadratic polynomial fitting algorithm is used to convert pixel coordinates into geodetic coordinates, with the error controlled within 0.5 pixels;

[0063] For the raster map of the municipal ecological environment zoning control plan, the resolution was unified to 30 meters by resampling (bilinear interpolation method) to ensure consistency with the spatial resolution of Landsat-9 imagery;

[0064] All vector data (territorial spatial planning, marine functional zoning, etc.) are batch converted to the CGCS2000 coordinate system using coordinate transformation tools, and finally output ecological planning dataset (GeoTIFF format, 30-meter resolution) and traffic remote sensing dataset (GeoTIFF format, 10-30-meter resolution) with unified spatial reference.

[0065] Based on the image characteristics (such as spectral features, texture features, and shape features) of the registered data, the server calls the eCognitionDeveloper multi-scale segmentation algorithm to adaptively adjust parameters to achieve land cover type vectorization.

[0066] Ecological planning map segmentation:

[0067] The overall land use plan map: For the linear features of the ecological protection red line (solid red line) and permanent basic farmland (dashed yellow line), the server is configured with a segmentation scale parameter of 15 (smaller object size), a shape factor of 0.8 (prioritizing the preservation of boundary shapes), and a compactness factor of 0.6. Continuous vector lines are extracted through edge detection and Hough transform. For internal color blocks (such as the yellow planar areas of permanent basic farmland), a region growing algorithm (seed point threshold: NDVI>0.3) is used to extract planar vectors and label them with the "farmland" attribute.

[0068] Mangrove Reserve Planning Map: Considering the spectral differences between the core area (filled with dark green) and the buffer zone (filled with light green), the segmentation scale parameter is set to 30 (for larger object sizes), the spectral factor to 0.9 (prioritizing color-based segmentation), and the shape factor to 0.3. The dark green area is classified as the "core area" vector surface and the light green area as the "buffer zone" vector surface using the maximum likelihood classification method, with the boundary error controlled within 5 meters.

[0069] Traffic remote sensing image segmentation:

[0070] Land transportation facility extraction: For Sentinel-2 imagery, the server employs object-oriented multi-scale segmentation: For existing roads (grayish-white strips, 10-30 meters wide), a segmentation scale of 20 and a texture factor of 0.7 are set (linear texture is extracted based on the gray-level co-occurrence matrix). The road features are identified and vectorized into line features using an SVM classifier (training samples: 1000 road pixels and 2000 non-road pixels); For existing port terminals (grayish areas with rectangular outlines), a segmentation scale of 50 and a shape factor of 0.7 are set (rectangular outlines are prioritized). The area vectors are extracted and labeled with the "port operation area" attribute.

[0071] Sea depth and shoreline extraction: The water depth vector data (containing 100,000 water depth points, coordinates and depth values) from the electronic nautical chart is used to generate a 20-meter resolution water depth raster through inverse distance weighted interpolation. Combined with the NDWI index (Normalized Water Index) of Landsat-9 imagery, the water area is extracted by setting a threshold of NDWI>0.3. The shoreline depression area is filled by morphological closing operation (5×5 rectangular structuring element) to generate a continuous land-sea boundary vector line to distinguish the land area from the sea area.

[0072] Segmentation result verification:

[0073] The server randomly selects 100 vector graphic units (such as ecological protection red line boundary points and highway centerline points) and compares them with the actual GPS measurement data (accuracy 0.5 meters). The planar position error is ≤2 meters, which meets the selection accuracy requirements. The final output is a vector dataset (Shapefile format) containing 12 types of land features, including ecological protection areas, ecological control areas, urban construction areas, farmland protection areas, transportation land and sea use areas, etc.

[0074] The server calculates the EDI value for each 10 m × 10 m grid cell through spatial overlay analysis and a weighted scoring model to quantify the synergy degree of ecological protection and development suitability:

[0075] Construction of the evaluation index system:

[0076] The server loads the preset scoring rule library:

[0077] Ecological constraint indicators: national nature reserves (core area: -10 points, buffer zone: -8 points), provincial ecological protection red lines (-6 points), mangrove wetlands (-5 points);

[0078] Development suitability indicators: within the urban development boundary (+10 points), within 5 km of existing transportation corridors (+8 points), industrial sea areas (+6 points), sea areas with a water depth ≥ 15 m (+5 points);

[0079] Weight assignment: The weight of ecological constraint indicators (0.6) is higher than that of development suitability indicators (0.4), reflecting the principle of ecological priority.

[0080] Spatial overlay and index calculation:

[0081] The server performs the following steps through the raster calculator tool of ArcGIS:

[0082] Convert 12 types of vector feature types into 10 m resolution rasters (e.g., the raster value of "national protected area core area" = -10, the raster value of "urban development boundary" = +10);

[0083] Calculate the EDI value using the weighted summation model: EDI = (Σ scores of ecological constraint indicators × 0.6) + (Σ scores of development suitability indicators × 0.4);

[0084] For conflict areas (e.g., a grid cell belongs to both the "provincial ecological protection red line" (-6 points) and the "urban development boundary" (+10 points)), correct it according to the principle of "taking the minimum value of ecological constraints", that is, the EDI of this cell = (-6 × 0.6) + (10 × 0.4) = -3.6 + 4 = 0.4.

[0085] Index visualization and statistics:

[0086] The server outputs an EDI raster map (value range -10 to +10), where:

[0087] EDI ≤ -5: Ecologically extremely sensitive area (dark blue), accounting for 15% of the total area;

[0088] -5 < EDI < 5: Ecological-development coordination area (light blue to light yellow), accounting for 65% of the total area;

[0089] EDI≥5: Development suitable area (red), accounting for 20% of the total area.

[0090] The server performs site selection optimization based on EDI values ​​and traffic engineering technical standards, categorized by type.

[0091] Port logistics park (land area) planning:

[0092] Suitable area selection: Server constraints are set as follows: EDI≥5 (suitable development area), area≥20 square kilometers, distance from existing port≤10 kilometers, terrain slope<5° (based on DEM data). Three candidate areas are extracted from the EDI raster (Area A: EDI=7.2, area 25km²; Area B: EDI=6.8, area 22km²; Area C: EDI=5.5, area 21km²).

[0093] Multi-factor comparison: The comprehensive score was calculated using the Analytic Hierarchy Process (AHP): ecological impact (weight 0.4, positive correlation with EDI value), construction cost (weight 0.3, area A has the smallest demolition amount and the highest score), and transportation convenience (weight 0.3, area A is close to the existing port access road and has the highest score). Finally, area A had a comprehensive score of 0.85 (area B 0.75, area C 0.68), and was determined as the site for the port logistics park.

[0094] Deepwater channel (sea area) selection:

[0095] Navigation conditions screening: Server constraints are set as follows: EDI≥0 (ecological-development coordination zone), water depth≥15 meters (meeting the navigation requirements of 300,000-ton vessels), width≥300 meters, and straight segment ratio≥80%. Two candidate routes are extracted from the overlay analysis of the sea area EDI grid and water depth grid (Route 1: 45km long, EDI mean 2.5; Route 2: 50km long, EDI mean 3.0).

[0096] Route optimization: The Dijkstra algorithm was used to calculate the path with the least ecological impact. Although route 2 is 5km longer, it has a higher EDI value (less ecological impact) and avoids the 5km range of the mangrove buffer zone (EDI=-5). Therefore, route 2 was finally selected as the deep-water channel.

[0097] Port access road (land area) planning:

[0098] Corridor determination: The server generates 5 candidate corridors (2km wide) in the area where EDI≥0, starting from the port logistics park (Area A) and ending in the main urban area. The path cost (ecological cost: negative correlation with EDI value; engineering cost: including bridge and tunnel length) is calculated using the minimum spanning tree algorithm, and the corridor with the lowest cost is selected.

[0099] Route optimization: Within the selected corridor, a genetic algorithm is used to optimize the specific route: avoiding ecologically sensitive points (such as small wetlands) with EDI≤-3, minimizing the amount of demolition (the proportion of road sections crossing the urban development boundary is ≥80%), and finally outputting the port access highway route (48.5km long, design speed 80km / h).

[0100] Selection results output:

[0101] The server integrates vector data of the port logistics park boundary, deep-water channel axis, and port access road centerline, overlays them onto the target city's GIS base map, and generates a thematic map of ecological collaborative planning for transportation infrastructure (1:50,000 scale). This map includes attributes such as site coordinates, area, and ecological constraints, and outputs a KML format file for planners to review.

[0102] Through the above steps, the server has achieved full automation from multi-source data collection to the output of delineation results. The delineation results comply with the principle of prioritizing ecological protection (avoiding 90% of the ecologically sensitive areas) and meet the needs of transportation engineering (the area of ​​the port logistics park meets the standards and the waterway depth meets the navigation requirements), providing scientific decision support for the planning of transportation infrastructure in the target city.

[0103] In this embodiment of the invention, the step of adaptively adjusting the segmentation parameters based on the image characteristics of the registered ecological planning data and the traffic remote sensing image data to perform multi-scale segmentation and obtain vector graphic data corresponding to each of the multiple candidate areas can be implemented through the following example.

[0104] Based on the planning map features of the registered ecological planning data, vector graphic data corresponding to the registered ecological planning data of the multiple candidate areas are obtained by optimizing color blocks and text filtering, and by integrating morphological filtering to eliminate legend label interference.

[0105] Based on the remote sensing image features of the traffic remote sensing image data, vector graphic data corresponding to the traffic remote sensing image data of the multiple candidate areas are obtained through texture enhancement and edge detection.

[0106] In an embodiment of the present invention, an exemplary method for processing ecological planning data (using a land spatial planning map as an example) is as follows:

[0107] For the registered municipal land spatial planning map (GeoTIFF format, 30-meter resolution), the server first performs color block optimization: through HSV color space conversion, the red blocks of the ecological protection red line (H=0-10°, S=0.6-1.0, V=0.5-1.0) and the blue blocks of the urban development boundary (H=180-240°, S=0.5-1.0, V=0.4-1.0) are color-enhanced to improve the contrast between the color blocks and the background. Next, text filtering is performed: a connected component analysis algorithm is used to identify the text labeled "core area" and "buffer zone" (pixel area <500, aspect ratio >3) in the map, and the text pixels are set to the background color (white) through masking operations to eliminate label interference. Subsequently, the morphological filtering module is invoked, and a 3×3 cross-shaped structural element is used to perform a closing operation on the color block area to fill the holes caused by text removal, so that the color block outlines of the ecological protection red line and the urban development boundary are continuous. Finally, a closed vector surface element is generated through the outline extraction algorithm, and the attributes of "ecological protection red line", "urban development boundary" and "permanent basic farmland" are labeled.

[0108] Traffic remote sensing image processing (using Sentinel-2 imagery as an example):

[0109] For 10-meter resolution traffic remote sensing imagery, the server first performs texture enhancement on existing port access roads (grayish-white striped texture): Texture energy values ​​at azimuth angles of 0°, 45°, 90°, and 135° are calculated using the gray-level co-occurrence matrix. Areas with texture energy higher than a threshold (mean + 1.5 times standard deviation) are marked as potential road areas, enhancing linear texture features. For the port terminal area (grayish rectangular outline), the Laplacian operator is used for edge enhancement, highlighting the gray-level transitions at the rectangular boundaries. Subsequently, the Canny edge detection algorithm (high threshold 0.2, low threshold 0.1) is used to extract the road centerline and port terminal outline. The Hough transform is used to fit the road edge points into straight line vectors, and a polygon approximation algorithm is used to convert the port outline points into planar vectors. Finally, a vector dataset (Shapefile format) containing road line features and port planar features is output.

[0110] Through the above processing, the server eliminated the legend interference of the ecological planning map and enhanced the color block boundaries, extracting 12 types of ecological spatial vectors; at the same time, it enhanced the texture and edge features of transportation facilities in the remote sensing image and extracted 8 types of transportation feature vectors, laying a data foundation for the subsequent calculation of the ecological-development index.

[0111] In this embodiment of the invention, the step of superimposing multiple vector graphic data to calculate the ecological-development index value corresponding to each of the multiple candidate areas can be implemented through the following example.

[0112] Through the formula: The ecological-development index value corresponding to each of the candidate areas is calculated;

[0113] in, The ecological-development index value. Let be the weight of the planning legal effect corresponding to the i-th vector graphic data of the candidate area. This represents the base score corresponding to the i-th vector graphic data of the candidate region;

[0114] If there are multiple overlapping vector graphics data in the area to be selected, the minimum base score among the multiple overlapping vector graphics data shall be taken as the base score of the area.

[0115] In an embodiment of the present invention, for example, the server uses a 10m × 10m grid cell along the coast of the target city as the basic evaluation unit and performs the following calculation process:

[0116] Weights and scores are loaded:

[0117] The server loads the pre-defined "Planning Legal Validity Weighting Library" and "Basic Score Table":

[0118] Ecological protection red line (national level): weight w=1.0, basic score s=-10;

[0119] Urban development boundary (city level): weight w=0.5, base score s=+8;

[0120] Mangrove Nature Reserve Buffer Zone (Provincial Level): Weight w=0.8, Base Score s=-6;

[0121] Existing port operation area (current transportation facilities): weight w=0.6, base score s=+10.

[0122] Spatial Overlay and Conflict Resolution:

[0123] Taking a candidate raster cell as an example (coordinates X=3852000m, Y=512000m), the server detected, using ArcGIS spatial overlay tools, that this cell simultaneously overlaps with three types of vector graphic data:

[0124] National-level ecological protection red line (w=1.0, s=-10);

[0125] City-level urban development boundary (w=0.5, s=+8);

[0126] Provincial-level mangrove buffer zone (w=0.8, s=-6).

[0127] According to the rule "take the minimum base score for overlapping areas", the server extracts the minimum value (-10) among the three s values ​​as the effective base score for that unit.

[0128] Index calculation:

[0129] Server application formula ,in Take the minimum value -10 (corresponding to the national ecological protection red line). =1.0), the calculation shows that the grid cell has extremely strong ecological constraints and is not suitable for transportation infrastructure construction.

[0130] Batch calculation and output:

[0131] The server executes the above process on the entire 1500 square kilometers (1.5×10^8 grid cells) and outputs an EDI raster map (value range -10 to +10). Among them, ecologically sensitive areas (EDI≤-5) account for 18% and development-suitable areas (EDI≥5) account for 22%, providing a quantitative basis for subsequent site selection.

[0132] In this embodiment of the invention, the method of selecting transportation infrastructure sites for the plurality of candidate areas based on the ecological-development index value to obtain the transportation infrastructure selection results for the plurality of candidate areas can be implemented through the following example.

[0133] Based on the ecological-development index value, site selection for land-based shoreline facilities and water-based facilities are carried out in the multiple candidate areas to obtain site selection results for either land-based shoreline facilities or water-based facilities.

[0134] In an embodiment of the invention, exemplarily, the site selection of land-based shoreline facilities (taking a port logistics park as an example): The server loads Eco-Development Index (EDI) raster data (10-meter resolution) and the vector boundary of the candidate area, and sets the screening conditions: EDI ≥ 5 (suitable development area), area ≥ 10 square kilometers, and distance from existing port terminals ≤ 8 kilometers. For three candidate land-based areas (Area A, Area B, and Area C) along the coast of the target city, the server extracts the average EDI value of each area through spatial analysis tools: Area A (EDI = 7.2, shoreline length 3.5 kilometers, demolition cost 230 million yuan), Area B (EDI = 6.8, shoreline length 2.8 kilometers, demolition cost 180 million yuan), and Area C (EDI = 5.5, shoreline length 2.2 kilometers, demolition cost 150 million yuan). The optimization function is called to calculate the comprehensive score (score = EDI × shoreline length / demolition cost). The score of area A is 7.2 × 3.5 / 2.3 ≈ 10.87, which is higher than that of area B (6.8 × 2.8 / 1.8 ≈ 10.58) and area C (5.5 × 2.2 / 1.5 ≈ 8.07). Area A is selected as the site for the port logistics park. The vector boundary and attribute table (including area, shoreline length, and mean EDI) are output.

[0135] Site selection for waterway facilities (taking coastal waterways as an example):

[0136] The server sets filtering criteria for the candidate sea area: EDI ≥ 0 (ecological-development coordination zone), water depth ≥ 12 meters (meeting the navigation requirements of 100,000-ton vessels), and continuous length ≥ 20 kilometers. Three candidate waterways are extracted from the overlay data of electronic charts and EDI raster data: Route 1 (25 km long, mean EDI 2.1, passing through two aquaculture areas), Route 2 (28 km long, mean EDI 3.0, avoiding aquaculture areas), and Route 3 (22 km long, mean EDI 1.8, with some sections having EDI = -2). The server removes Route 3 (EDI < 0) and calculates the ecological impact index (index = route length × (10 - mean EDI) / 10) for Routes 1 and 2. The index for Route 1 is 25 × (10 - 2.1) / 10 = 19.75, and the index for Route 2 is 28 × (10 - 3.0) / 10 = 19.6. Route 2, with the lower index, is selected as the optimal waterway, and the axis vector and 100-meter-wide navigation strip are output.

[0137] In this embodiment of the invention, the site selection of land shoreline facilities is carried out based on the ecological-development index value of the multiple candidate areas to obtain the site selection results of land shoreline facilities. This can be implemented through the following example.

[0138] Multiple consecutive candidate areas with an ecological-development index value greater than a first preset index threshold are designated as candidate areas, and candidate areas with an ecological-development index value less than a second preset index threshold are designated as avoidance areas; the avoidance areas are used to determine buffer avoidance areas within a preset range.

[0139] By optimizing the function: Calculate the planning suitability index for each of the proposed areas, and take the proposed area corresponding to the largest planning suitability index as the site selection result for the land shoreline facilities.

[0140] Where TF is the planning suitability index, and k is the number of planned transportation infrastructure facilities. Let EDI be the value of the k-th suitable region. Let k be the length or area of ​​the k-th suitable region. Let α be the construction cost of the kth transportation infrastructure, and let α be the cost coefficient for reconstruction and expansion when existing infrastructure is occupied.

[0141] In this embodiment of the invention, for example, threshold setting and area selection are as follows: The server first loads 10-meter resolution Eco-Development Index (EDI) raster data of the coastal area of ​​the target city, sets the first preset index threshold (development suitability threshold) to 5, and the second preset index threshold (ecological sensitivity threshold) to -3. Continuous candidate areas with EDI ≥ 5 are extracted using ArcGIS spatial analysis tools, and fragmented patches with an area < 2 square kilometers are merged to obtain three candidate areas (Areas A, B, and C):

[0142] Area A: Distributed along the coastline, covering an area of ​​12 square kilometers, with an average EDI value of 7.5 (mainly industrial reserve land within the urban development boundary), and no overlapping avoidance zones;

[0143] Area B: Adjacent to the existing port operation area, covering an area of ​​10 square kilometers, with an average EDI of 6.8 (including some farmland conversion areas). There is a provincial mangrove nature reserve 500 meters to the north (EDI=-5<-3, determined to be an avoidance zone).

[0144] Area C: Nearshore tidal flats, covering an area of ​​15 square kilometers, with an average EDI of 5.2 (including some salt fields). There are 3 municipal-level ecological public welfare forests scattered within the area (EDI=-4<-3, which are determined to be avoidance areas).

[0145] The server performs buffer zone analysis on the selected avoidance areas: a buffer avoidance area is generated by extending 500 meters outward from the mangrove reserve north of area B, with an overlap area of ​​2 square kilometers with area B. After removing the overlap, the effective area of ​​area B is 8 square kilometers; a buffer avoidance area is generated by extending 300 meters outward from the ecological public welfare forest of area C, with an overlap area of ​​4 square kilometers. After removing the overlap, the effective area of ​​area C is 11 square kilometers.

[0146] Calculation of the suitability index (TF):

[0147] The server sets the planned number of transportation infrastructures to k=1 (to create one new port land expansion area), and calls the optimization function. The parameter values ​​are as follows:

[0148] Area A: =7.5 (regional EDI mean) =12 (area, square kilometers) =8.5 (construction cost, including land acquisition fee of 620 million yuan and earthwork backfilling fee of 230 million yuan), α=1.0 (new area, no existing facilities occupying), and TF=7.5×12 / (1.0×8.5)≈10.59;

[0149] Area B: =6.8 (mean EDI value of effective area). =8 (effective area, square kilometers) =7.2 (construction cost, including land acquisition fee of 500 million yuan and demolition fee of 220 million yuan in the buffer zone), α=0.8 (adjacent to the existing port, some roads can be reused, reconstruction and expansion coefficient of 0.8), calculated to TF=6.8×8 / (0.8×7.2)≈9.44;

[0150] Area C: =5.2 (mean EDI value of effective area). =11 (effective area, square kilometers) =9.8 (construction cost, including 450 million yuan for salt field conversion and 530 million yuan for buffer zone land acquisition), α=1.0 (newly built area), and TF=5.2×11 / (1.0×9.8)≈5.84.

[0151] The result is confirmed:

[0152] The server compares the TF values ​​of the three regions (Area A 10.59 > Area B 9.44 > Area C 5.84), selects Area A with the highest TF value as the site selection result for the port land expansion area, and outputs the vector boundary (including inflection point coordinates) and attribute table (area 12 square kilometers, average EDI 7.5, construction cost 850 million yuan, no overlap of avoidance zones).

[0153] In this embodiment of the invention, the site selection and planning results of the water facilities include the waterway planning results. Based on the ecological-development index value, the site selection of water facilities is carried out on the multiple areas to be selected, and the site selection and planning results of the water facilities are obtained. The following example can be used to implement this.

[0154] Acquire electronic nautical chart depth point cloud data;

[0155] Based on the electronic nautical chart water depth point cloud data, multiple high-precision water depth grids are generated by Kriging interpolation.

[0156] Extract continuous regions that meet the minimum navigable water depth threshold from the multiple high-precision water depth grids, and remove broken waterways whose length or width does not meet the preset distance.

[0157] The continuous region is abstracted as a network node. The node spacing is defined as connected if it is less than a preset spacing threshold. The inverse of the EDI exponent corresponding to the continuous region is used as the edge weight. The Dijkstra algorithm is used to calculate the minimum cumulative weight path from the preset starting point to the ending point.

[0158] An adaptive width corridor is generated along the path with the lowest cumulative weight. For the route segment that crosses the ecologically sensitive area, the route is oscillated within the adaptive width corridor to determine the route selection result.

[0159] In an embodiment of the present invention, for example, the server selects a 100,000-tonnage coastal waterway from the existing port of the target city (starting coordinates X=3,850,000m, Y=510,000m) to the offshore anchorage (ending coordinates X=3,830,000m, Y=530,000m) and executes the following process:

[0160] Electronic chart depth data processing:

[0161] The server acquires water depth point cloud data from electronic nautical charts of the target sea area through a preset database interface. This data includes 150,000 water depth measurement points (format S-57, coordinate system CGCS2000, depth unit in meters, accuracy ±0.5 meters), with a point cloud density of 80 points per square kilometer. The ArcGIS geostatistical analysis module is then used, employing Kriging interpolation (with a spherical variogram model, search radius of 500 meters, and lag distance of 200 meters) to convert the point cloud data into a 10-meter resolution high-precision water depth raster, outputting a GeoTIFF format file (depth range -2 meters to -25 meters, negative values ​​indicating underwater depth).

[0162] Navigation area extraction:

[0163] The server sets a minimum navigable water depth threshold of 12 meters (to meet the navigability requirements of a 100,000-ton bulk carrier at full load). Regions with a depth ≤ -12 meters (i.e., water depth ≥ 12 meters) are extracted from the water depth grid. Continuous water patches are identified through 8-neighbor connectivity analysis. Broken waterways with a length < 20 kilometers or a width < 100 meters (such as short waterways separated by near-shore shoals) are removed, ultimately retaining three continuous navigable areas: Area 1 (22 kilometers long, average width 350 meters), Area 2 (28 kilometers long, average width 420 meters), and Area 3 (25 kilometers long, average width 280 meters).

[0164] Network node construction and path calculation:

[0165] The server abstracts three consecutive navigable areas into a network model: nodes are evenly distributed in each area at 500-meter intervals (generating a total of 110 nodes, including the start point, end point, and turning points within the area), and nodes with a spacing of less than 500 meters are defined as connected. The EDI raster data (10-meter resolution) calculated in step 3 is loaded, and the EDI index of each node's location is extracted (mean 2.1 for area 1, 3.0 for area 2, and 1.8 for area 3). The reciprocal of the EDI index is used as the edge weight (e.g., weight = 1 / 3 ≈ 0.33 when EDI = 3.0, weight ≈ 0.56 when EDI = 1.8). The Dijkstra algorithm is run using a network analysis tool, with the existing port as the start point and the offshore anchorage as the end point, to calculate the path with the lowest cumulative weight: the cumulative weight of the path in area 2 = 28 km × 0.33 ≈ 9.24, lower than that in area 1 (22 × 0.48 ≈ 10.56) and area 3 (25 × 0.56 ≈ 14.00). Area 2 is selected as the base channel direction.

[0166] Adaptive corridor generation and ecologically sensitive area processing:

[0167] The server generates a channel corridor with an initial width of 300 meters along the path with the lowest cumulative weight. The width is dynamically adjusted based on the EDI index within the corridor: segments with EDI ≥ 5 (suitable development areas) are widened to 350 meters; segments with EDI = 0-5 (coordination areas) remain at 300 meters; and segments with EDI < 0 (ecologically sensitive areas) are narrowed to 250 meters. In the middle of the corridor, an ecologically sensitive area (mangrove buffer zone, EDI = -4, 3 km long) is detected. The server generates three alternative swing paths within the corridor (lateral offsets of 500 meters, 1000 meters, and 1500 meters). The ecological impact value of each path is calculated (impact value = offset distance × (10 + EDI) / 10): For an offset of 500 meters, the impact value = 500 × (10 - 4) / 10 = 300; for an offset of 1000 meters, = 1000 × 6 / 10 = 600; and for an offset of 1500 meters, = 1500 × 6 / 10 = 900. The path with the smallest impact value, offset by 500 meters, is selected. The final output of the waterway selection results includes: axis vector (28.5 km long), adaptive width corridor (250-350 m), and attribute table (including the average EDI value, width, and ecologically sensitive area offset distance for each waterway segment).

[0168] In this embodiment of the invention, the site selection and delineation results of the water facilities include the anchorage selection results. The site selection of water facilities is carried out on the multiple candidate areas based on the ecological-development index value to obtain the site selection and delineation results of the water facilities. The following example can be used to implement this.

[0169] Acquire electronic nautical chart depth point cloud data;

[0170] Based on the electronic nautical chart water depth point cloud data, multiple high-precision water depth grids are generated by Kriging interpolation.

[0171] The continuous regions that meet the minimum anchoring depth threshold are extracted from the multiple high-precision water depth grids, and broken water areas with an area smaller than the preset area threshold are removed to obtain the basic navigation area.

[0172] From the basic navigation area, areas with an ecological-development index greater than a third preset index threshold are selected as preferred anchorage areas, and the anchorage selection results are obtained by sorting them in descending order.

[0173] In an embodiment of the present invention, for example, the server selects an anchorage off the coast of a target city (for temporary berthing of 100,000-ton bulk carriers) as a scenario and executes the following process:

[0174] Electronic chart depth data acquisition and processing:

[0175] The server retrieves water depth point cloud data (S-57 format, CGCS2000 coordinate system, depth unit meters) from the electronic nautical chart of the target sea area via a preset data platform interface. This data includes 100,000 water depth measurement points (covering an area of ​​200 square kilometers, point cloud density of 500 points / square kilometer, depth accuracy ±0.5 meters). The ArcGIS geostatistical analysis module is then used, employing Kriging interpolation (with a Gaussian variogram model, search radius of 800 meters, and lag distance of 300 meters) to convert the point cloud data into a 30-meter resolution high-precision water depth raster (GeoTIFF format, depth range -3 meters to -20 meters, negative values ​​indicating underwater depth).

[0176] Basic airspace extraction:

[0177] The server sets a minimum anchoring depth threshold of 10 meters (to meet the anchoring requirements of 100,000-ton vessels, considering a margin of 2 meters). Regions with a depth ≤ -10 meters (i.e., depth ≥ 10 meters) are extracted from the depth grid. Continuous water patches are identified through 8-neighbor connectivity analysis, and fragmented water areas with an area < 2 square kilometers (such as small waterways between isolated shoals) are removed, ultimately retaining 3 basic navigation zones:

[0178] Area A: Nearshore offshore area, covering an area of ​​3.5 square kilometers, with a regular shape (approximately rectangular) and no obvious abrupt changes in water depth;

[0179] Area B: Central sea area, with an area of ​​2.8 square kilometers, is an irregular polygon with two small protrusions on the edge;

[0180] Area C: Offshore waters, 2.2 square kilometers in area, long and narrow, with some edges having a water depth approaching the 10-meter threshold.

[0181] Preferred anchorage areas selection and sorting:

[0182] The server loads the Eco-Development Index (EDI) raster data (10-meter resolution) calculated in step 3 and extracts the average EDI values ​​for three basic navigation areas: Area A 2.5 (Ecological-Development Coordination Area), Area B 1.8 (Ecological-Development Coordination Area), and Area C -0.5 (Slightly Ecologically Sensitive Area). A third preset index threshold is set to 0 (lower limit for Eco-Development Coordination Area), and Areas A and B with EDI > 0 are selected (Area C EDI = -0.5 ≤ 0 is removed). The areas are sorted in descending order of area: Area A (3.5 square kilometers) > Area B (2.8 square kilometers). Area A is selected as the preferred anchorage area, and the vector boundary and attribute table (including area, average water depth of 12.5 meters, average EDI value of 2.5, and coordinate range) are output.

[0183] In this embodiment of the invention, the following implementation methods are also provided.

[0184] A simulation model of the ecological impact of the selected planning scheme is constructed using generative adversarial networks. The vector graphic data corresponding to the selected planning results of the transportation infrastructure is used as the input of the generator. The generated simulated regional ecological landscape is compared with the historical ecological baseline data by a discriminator to calculate the similarity of ecological impact.

[0185] If the similarity is lower than the preset threshold, the boundary range of the delineation scheme is adjusted based on the gradient information fed back by the discriminator until the ecological impact similarity reaches the preset threshold, and the final delineation result is obtained.

[0186] In an embodiment of the present invention, for example, the server takes the aforementioned selected port logistics park area A (originally selected area of ​​12 square kilometers, average EDI value of 7.5, located within the coastal town development boundary of the target city) as the object, and simulates its ecological impact through generative adversarial network (GAN) to optimize the selected boundary to reduce ecological disturbance.

[0187] The server builds an "ecological impact simulation GAN" model, which includes a generator and a discriminator:

[0188] Generator structure: Employs the U-Net architecture, with input being vector graphics data (Shapefile format) of the selected plan, including:

[0189] Original boundary vector of Area A of the port logistics park (polygon, containing coordinates of 12 vertices);

[0190] Vector of surrounding ecologically sensitive areas (mangrove buffer zone: 200 meters east of area A, provincial wetland: 800 meters south of area A);

[0191] Planning infrastructure layout parameters (main road width 20 meters, storage area accounting for 60%, green isolation belt width 50 meters).

[0192] The generator uses spatial feature encoding (converting vector boundaries into a 30-meter resolution binary mask, where 1 represents a selected area and 0 represents a non-selected area) combined with an ecological impact transfer function (such as the "construction intensity-vegetation destruction rate" mapping relationship: for every 10% increase in construction intensity, the vegetation destruction rate increases by 8%) to output a simulated "ecological landscape"—a spatial distribution map (30-meter resolution) containing three key ecological indicators.

[0193] NDVI (Normalized Difference Vegetation Index, representing vegetation cover, with a value range of -1 to 1).

[0194] Soil organic carbon content (characterizing soil fertility, in %).

[0195] Bird habitat suitability index (characterizing biodiversity, value range 0 to 1).

[0196] Discriminator structure:

[0197] Using an improved ResNet-50 architecture, the input consists of a simulated ecological landscape output by the generator and "historical ecological baseline data" (real samples). Features (such as NDVI spatial gradient, soil carbon content texture, and habitat patch connectivity) are extracted through 10 layers of convolution, and the output is a "similarity score" (0 to 1, where 1 indicates that the simulated landscape is completely consistent with the baseline and the ecological impact is acceptable).

[0198] Historical ecological baseline data:

[0199] The server obtains baseline data (as real samples for the discriminator) of the selected area and its surrounding 5-kilometer radius through the following channels:

[0200] Landsat-8 remote sensing images from 2018 to 2022 (3 images per year, mean NDVI 0.42±0.03).

[0201] Data from the ecological and environmental monitoring station (soil organic carbon content 2.1% ± 0.2%, sampled in 2022);

[0202] Bird survey report (Summer 2021, mean habitat suitability index 0.78±0.05, 12 waterbird species recorded).

[0203] The server starts training the GAN model (1000 iterations, batch size=16, learning rate 0.0002). After training, the original delineation scheme vector data of area A is input to generate the first round of simulated ecological landscape:

[0204] NDVI simulation results: The mean NDVI value in the selected area was 0.35 (baseline 0.42, a decrease of 16.7%), mainly due to the removal of native vegetation during the construction of the storage area;

[0205] Soil organic carbon simulation results: The mean value in the selected area is 1.8% (baseline 2.1%, a decrease of 14.3%), due to topsoil loss caused by land leveling;

[0206] Bird habitat suitability index: The average value within 500 meters of the selected area was 0.65 (baseline 0.78, a decrease of 16.7%), because the greenbelt did not completely offset the impact of habitat fragmentation.

[0207] The discriminator compares the simulated scene with historical baseline data, calculates the similarity of each indicator using the Structural Similarity Index (SSIM), and obtains the overall ecological impact similarity after weighted averaging.

[0208] NDVI similarity 0.70 (SSIM value), soil carbon similarity 0.75, habitat index similarity 0.71;

[0209] Overall similarity = (0.70×0.4+0.75×0.3+0.71×0.3) = 0.72 (preset threshold 0.85, below the threshold).

[0210] The server extracts the gradient information output by the discriminator (indicating the area where the simulated scene differs most from the baseline):

[0211] The main areas of difference were: the eastern side of area A (200 meters from the mangrove buffer zone) showed the most significant decrease in NDVI (contributing 40% of the similarity loss), due to the encroachment of the selected boundary into the mangrove "edge effect zone" (the baseline NDVI in this area was 0.55, the simulated value was 0.30, a decrease of 45.5%).

[0212] Secondary areas of difference: Soil carbon content decreased more rapidly in the southern part of Area A (adjacent to the provincial wetland) (contributing 25% of the loss), because the original boundary included a 1.2 square kilometer wetland transition zone (baseline soil carbon 2.5%, simulated value 1.6%, a decrease of 36%).

[0213] Initial boundary adjustment:

[0214] The server shrinks the eastern boundary based on gradient information: the coordinates of the eastern vertex of area A are adjusted westward from (X=3852200m, Y=512000m) to (X=3851900m, Y=512000m), shrinking by 300 meters, increasing the boundary distance from the mangrove buffer zone to 500 meters (beyond the edge effect zone); simultaneously, the southern boundary is shrunk: the coordinates of the southern vertex are adjusted northward from (X=3852000m, Y=511200m) to (X=3852000m, Y=511500m), shrinking by 300 meters, eliminating 0.8 square kilometers of wetland transition zone. After the adjustment, the area of ​​area A is reduced to 10.5 square kilometers.

[0215] Secondary simulation and similarity verification:

[0216] The server re-inputs the adjusted boundary vector into the generator and outputs a secondary simulated ecological landscape:

[0217] The mean NDVI was 0.39 (baseline 0.42, down 7.1%), while the NDVI at the eastern edge rose to 0.48 (baseline 0.55, down 12.7%).

[0218] The average soil organic carbon content was 2.0% (baseline 2.1%, a decrease of 4.8%), while the carbon content in the southern wetland transition zone rebounded to 2.3% (baseline 2.5%, a decrease of 8%).

[0219] The bird habitat suitability index was 0.73 (baseline 0.78, a decrease of 6.4%). The discriminator calculated the overall similarity as (0.82×0.4+0.88×0.3+0.80×0.3)=0.83 (still below the threshold of 0.85). The gradient showed that the difference mainly came from the northwest corner of area A (current farmland area, NDVI baseline 0.48, simulated value 0.40, a decrease of 16.7%).

[0220] Three boundary adjustments:

[0221] The server shrinks the northwest boundary of Area A by 150 meters to the north (removing 0.3 square kilometers of existing high-yield farmland), reducing the area to 10.2 square kilometers. After three simulations:

[0222] The mean NDVI was 0.40 (baseline 0.42, a decrease of 4.8%).

[0223] The average soil organic carbon content was 2.05% (baseline 2.1%, a decrease of 2.4%).

[0224] The bird habitat suitability index was 0.76 (baseline 0.78, a decrease of 2.6%). The discriminator's overall similarity score was (0.89×0.4+0.92×0.3+0.87×0.3)=0.89 (≥threshold 0.85), indicating that the ecological impact met the standards.

[0225] The server locks the adjusted boundary (area of ​​10.2 square kilometers, a 15% reduction from the original plan) and generates the final delineation result: including the optimized vector boundary (coordinates of 15 vertices), the ecological impact assessment report (including similarity scores of each indicator and comparison maps before and after adjustment), and the "ecological compensation recommendation" (to add a 200-meter-wide artificial mangrove belt outside the eastern boundary to offset the remaining ecological impact).

[0226] In this embodiment of the invention, the following implementation methods are also provided.

[0227] By constructing an ecological-transportation collaborative sensing network, real-time access is provided to ecological sensor data and traffic flow prediction data of the candidate area.

[0228] Based on the ecological sensor data and the traffic flow prediction data, areas where the short-term fluctuations of the associated indicators exceed a preset range are marked as temporary restricted areas. After the fluctuations of the indicators in the temporary restricted areas stabilize, the corresponding traffic infrastructure selection results are re-evaluated.

[0229] In this embodiment of the invention, for example, the selected port logistics park (Area A) and the surrounding area of ​​the supporting port access road in the target city are taken as the monitoring objects. The server realizes dynamic evaluation and adjustment of the selection results by constructing an ecological-transportation collaborative perception network.

[0230] Construction of an Eco-Transportation Collaborative Sensing Network:

[0231] The server leads the deployment of a multi-level sensing network:

[0232] Ecological sensor layer: 12 ecological monitoring stations will be deployed within a 5-kilometer radius of Area A of the port logistics park (including land and nearshore waters); 6 atmospheric environmental sensors will be set up on land to monitor PM2.5. 2.5 The system includes: NO2 concentration (accuracy ±5μg / m³, sampling frequency 15 minutes / time); 3 soil moisture sensors (monitoring soil moisture content in the 0-50cm layer, accuracy ±2%, sampling frequency 1 hour / time); and 3 water quality sensors (monitoring dissolved oxygen (DO), pH, and turbidity, DO accuracy ±0.2mg / L, sampling frequency 30 minutes / time), with 2 located near the shoreline in Area A (500 meters from the shore) and 1 kilometer downstream of the planned port access road bridge. The sensors are connected to the server data platform via the LoRaWAN low-power protocol, with a data transmission latency of ≤5 minutes.

[0233] Traffic perception layer: Connects to the "Smart Transportation" platform API interface of the city traffic management bureau to obtain real-time traffic flow data (collected by loop detectors, every 5 minutes, accuracy ±5 vehicles) and vehicle classification data (video recognition, every 10 minutes) of the current port access road (two-way four lanes, designed flow of 5,000 vehicles / day). At the same time, it calls the traffic flow prediction model (based on LSTM neural network, inputting historical flow, weather, holidays and other features to predict the flow of the next hour, with an error rate ≤10%) to obtain the traffic flow prediction data of the planned port access road extension (connecting Area A and the main urban area) (updated every 15 minutes).

[0234] Data fusion layer: The server builds a spatiotemporal database (PostgreSQL + PostGIS) to uniformly store ecological sensor data (with timestamps and latitude and longitude) and traffic flow data (with road segment IDs and timestamps). Data association queries are realized through spatiotemporal indexes (such as "correlation between DO concentration in the area around Zone A and traffic flow on the port access road during the period from 08:00 to 09:00 on October 1, 20XX").

[0235] Real-time data access and indicator fluctuation monitoring:

[0236] The server polls the sensor network every 5 minutes, accessing and parsing the data.

[0237] Ecological indicator monitoring: At 08:30 on October 1, 20XX, the water quality sensor near the shoreline of Area A (coordinates X=3852500m, Y=511800m) uploaded data showing that the dissolved oxygen (DO) concentration decreased from 6.8mg / L at 08:00 to 4.5mg / L at 08:30, a decrease of 2.3mg / L within 24 hours (the preset short-term fluctuation threshold is 2mg / L, with an excess of 0.3mg / L); it was also detected that the sea area within 1 km of the sensor is a planned anchorage candidate area (the edge of Area B in step 7), and the period of DO concentration decrease highly overlapped with the activity time of nearby construction vessels (temporary dredging operations).

[0238] Traffic indicator monitoring: At 09:00 on the same day, the server obtained traffic flow forecast data for the planned port access road extension (K5+000 to K8+000 section): the forecast traffic flow for the next hour (09:00-10:00) is expected to increase from the current 1200 vehicles / hour to 1850 vehicles / hour, an increase of 54.2% (the preset short-term fluctuation threshold is 30%, and the excess is 24.2%). This section is located on the main collection and distribution channel of the selected port logistics park A area. In the original selection plan, the design capacity of this section was 1500 vehicles / hour, and the forecast traffic flow has exceeded the design value.

[0239] Temporary Restricted Area Marking and Processing:

[0240] The server performs the following operations on the region corresponding to the over-range indicator:

[0241] Ecologically Sensitive Area Marking: In response to excessive fluctuations in DO concentration, the server generates a 1.5-kilometer circular buffer zone centered on the water quality sensor (covering the edge of the planned anchorage area B and adjacent sea areas), marking it as a "temporary restricted area". The detailed design of the anchorage in this area (such as anchor numbering and mooring facility layout) is suspended, and early warning information (including sensor location and DO concentration change curve) is pushed to the preset departments.

[0242] Traffic congestion risk zone marking: For the port access road extension section where the traffic flow prediction exceeds the limit, the server extracts the vector line of the road section (3 kilometers in length) and generates a 500-meter-wide corridor (including the affected areas on both sides of the road), marking it as a "temporary restricted zone". The construction drawing design of the road section is suspended, and the traffic organization optimization plan is launched simultaneously (such as adding temporary lanes and adjusting signal timing).

[0243] Indicator stability and reassessment of selection results:

[0244] The server continuously monitors the temporary restriction zone selection index until the fluctuations stabilize.

[0245] Ecological indicators are stable: From 08:00 on October 2nd, DO concentration gradually recovered, reaching 6.5 mg / L by 18:00 on October 2nd. The fluctuation range over 24 hours was ≤0.5 mg / L (≤preset threshold 2 mg / L), indicating stable ecological indicators. The server re-extracted the EDI index for this area (combined with the ecological data after DO recovery, the average EDI value increased from 2.8 to 3.0), and the edge of planned anchorage area B was included in the candidate range, adjusting the total anchorage area to 3.0 square kilometers (originally 2.8 square kilometers).

[0246] Traffic indicators are stable: Through the implementation of temporary lane control (emergency lanes are activated from 09:00 to 10:00), the traffic flow dropped to 1450 vehicles / hour after 10:00 on October 1st (≤ design capacity of 1500 vehicles / hour). The predicted traffic flow fluctuation range over 48 consecutive hours is ≤15% (≤ preset threshold of 30%), indicating that traffic indicators are stable. The server recalculated the planning suitability index for this road section (TF value increased from 0.0032 to 0.0035), maintaining the original selected route, and simultaneously incorporating "adding an emergency lane" into the supplementary clauses of the construction drawing design.

[0247] Ultimately, the server outputs the dynamically adjusted delineation results (including anchorage boundary adjustments and supplementary clauses for port access road design), achieving real-time coordination between ecological protection and transportation development.

[0248] To more clearly describe the solutions provided in the embodiments of the present invention, a more complete implementation method is provided below.

[0249] This system employs a three-tiered architecture: "Intelligent Vectorization of Multi-Source Images → Ecological-Development Index Modeling → Collaborative Decision-Making Constrained by Water Depth." Please refer to the relevant documentation for further details. Figure 2 , Figure 2 The flowchart for multi-source image data acquisition and registration provided in this embodiment of the invention is as follows:

[0250] 1. Multi-source image data acquisition and registration

[0251] (1) Ecological planning data: including the overall land space plan, the coastline protection and utilization plan, the ecological environment zoning control scheme, the nearshore marine environmental functional zoning, the planning of various levels and types of nature reserves such as national parks, nature reserves, scenic spots, forest parks, and wetland parks, and the overall planning of the protection and development of the shoreline of the Yangtze River and other river basins;

[0252] Ecological planning data is converted into GeoTIFF format after GIS georeferencing to provide spatial reference.

[0253] (2) Traffic and geographic data: including electronic nautical charts (including depth point vectors), Landsat, Sentinel and other high-resolution remote sensing images.

[0254] 2. Object-oriented multi-scale segmentation vectorization

[0255] (1) Based on the characteristics of the multi-source images collected in step 1, the segmentation parameters are adaptively adjusted. For example, for land spatial planning, ecological protection areas (dark green) and industrial areas (brown) are segmented based on the uniformity of color blocks; for nature reserve maps of all levels and types, morphological filtering is used to eliminate legend label interference. Different ecological spatial divisions of land and water (sea) areas are obtained, including the land use planning type division of land areas, the ecological spatial division of water (sea) areas, and the division of shorelines with different control requirements.

[0256] (2) For remote sensing image processing: extract the water-land boundary (shoreline) by texture gradient detection, consider the hue, shape, texture and related features of port land, docks, etc., and identify the current transportation infrastructure, such as port docks, by combining object spectral clustering.

[0257] (3) Compare the output vector with the manual result to ensure accuracy.

[0258] Please refer to Table 1 for the processing details of the multi-source image data following the above process.

[0259] Table 1

[0260]

[0261] 3. Construction of the Eco-Development Index (EDI) Model

[0262] (1) Constructing a scoring system:

[0263] Table 1 above contains ecological-development spatial vector data, which are automatically overlaid and reclassified. A scoring system of -10 to 10 is established. For a space or shoreline, the more it is positioned for development, the higher the score, and the closer it is to 10; the more it is positioned for protection, the lower the score, and the closer it is to -10.

[0264] Taking into account the sensitivity and protection level of ecological space and development space, scores are assigned according to the method in Table 2 (Score Calculation Table for Space and Shoreline).

[0265] Table 2

[0266]

[0267] (2) Constructing a spatial overlay algorithm:

[0268] Considering that different plans and nature reserves may overlap, the scores for the same space are summed according to their type and assignment to obtain the final score for that space or shoreline, as shown in the following formula.

[0269] ,in,

[0270] The planning legal effect weight for ecological spaces is set as follows (e.g., w=1.0 for national nature reserves and w=0.7 for local nature reserves), while non-ecological spaces are set to 1.0.

[0271] This serves as the base score for spatial partitioning;

[0272] If there are conflicting areas, the strictest score will be applied (e.g., ecological protection zone + industrial zone → EDI=-10).

[0273] The output of this formula is a spatially continuous EDI raster (with a resolution of 10 meters). In ArcGIS, the final scores of the space and shoreline are displayed to obtain the corresponding ecological-development index of the space and shoreline.

[0274] 4. Automatic planning and selection of transportation infrastructure

[0275] (1) Land / shoreline facilities (highways, railways, port land):

[0276] Screening principle: Prioritize continuous regions with EDI ≥ 5;

[0277] Avoidance rule: A 500m buffer is generated in the EDI≤-5 area.

[0278] Establish the optimization function: Where TF is the suitability index for the planned location of transportation infrastructure, and k is the number of planned transportation infrastructure facilities. Let EDI be the value of the k-th suitable region. Let k be the length or area of ​​the k-th suitable region. Let α be the construction cost of the k-th transportation infrastructure, and α be the cost coefficient for reconstruction and expansion when existing infrastructure is occupied, which is 1.0 for new construction.

[0279] The area with the highest TF value will be selected as the suitable location for planning transportation infrastructure.

[0280] (2) Water facilities (channels, anchorages):

[0281] By combining electronic nautical chart depth data and EDI layers, channels and anchorages are automatically selected and demarcated.

[0282] 1) Waterway (given the starting and ending points, a route needs to be selected, i.e., a specific path):

[0283] First, the basic navigation area is extracted. Based on the electronic nautical chart's water depth point cloud data, a high-precision water depth raster (resolution ≤10m) is generated through kriging interpolation; a minimum navigation water depth threshold is set (e.g., 15m for coastal channels and 4m for inland waterways), and continuous areas that meet the water depth requirements are extracted; a minimum contiguous scale is defined: broken waterways (channels) with a length of less than 5km and waterways with a width of less than 100m are removed.

[0284] Secondly, a topological network is constructed. Contiguous navigation areas are abstracted as network nodes (node ​​spacing ≤ 500m is considered connected); the reciprocal of the EDI exponent is used as the edge weight (the higher the EDI, the smaller the weight; paths with high development index are prioritized); Dijkstra's algorithm is used to calculate the lowest cumulative weight path from the pre-defined waterway start to the end point, ensuring a balance between physical connectivity and ecological economy.

[0285] Third, optimize channel generation. Generate an adaptive width corridor (base width = 100m, dynamically adjusted ±20% according to EDI) along the topology path; for segments crossing ecologically sensitive areas (EDI≤-5), the channel will oscillate within the adaptive width corridor, automatically generating alternative detour options.

[0286] 2) Anchorage

[0287] First, the basic navigation area is extracted. Based on electronic nautical chart depth point cloud data, a high-precision depth raster (resolution ≤10m) is generated through kriging interpolation; a minimum anchorage depth threshold is set (e.g., 15m for coastal anchorages, 4m for inland river anchorages), and continuous areas meeting the depth requirements are extracted; a minimum contiguous scale is defined: areas smaller than 1km² are excluded. 2 Broken waters.

[0288] Secondly, based on ecological coordination constraints, the preferred areas of anchorage are delineated from the candidate areas, considering EDI≥0. For all preferred areas, they are sorted in descending order according to EDI index values.

[0289] Please refer to the accompanying diagram. Figure 3 , Figure 3 An ecological collaborative planning device 110 for transportation infrastructure provided in an embodiment of the present invention includes:

[0290] The acquisition module 1101 is used to collect multi-source ecological planning image data and traffic remote sensing image data, and to perform GIS georegistration on the multi-source ecological planning image data to convert it into registered ecological planning image data with spatial reference; based on the image characteristics of the registered ecological planning data and the traffic remote sensing image data, the segmentation parameters are adaptively adjusted to perform multi-scale segmentation to obtain vector graphic data corresponding to multiple candidate areas.

[0291] The selection module 1102 is used to overlay multiple vector graphic data to calculate the ecological-development index value corresponding to each of the multiple candidate areas; based on the ecological-development index value, it performs traffic infrastructure site selection on the multiple candidate areas to obtain the traffic infrastructure selection results for the multiple candidate areas. It should be noted that the implementation principle of the aforementioned traffic infrastructure ecological collaborative selection device 110 can refer to the implementation principle of the aforementioned traffic infrastructure ecological collaborative selection method, and will not be repeated here. It should be understood that the division of the various modules in the above device is only a logical functional division; in actual implementation, they can be fully or partially integrated into a single physical entity, or they can be physically separated. Furthermore, these modules can all be implemented in software through processing element calls; they can all be implemented in hardware; or some modules can be implemented through processing element calls to software, and some modules can be implemented in hardware. For example, the traffic infrastructure ecological collaborative selection device 110 can be a separately established processing element, or it can be integrated into a chip in the above device. Alternatively, it can be stored in the memory of the above device as program code, and called and executed by a processing element of the above device. The implementation of other modules is similar. Furthermore, these modules can be integrated, either wholly or partially, or implemented independently. The processing element described here can be an integrated circuit with signal processing capabilities. During implementation, each step of the above method or each of the above modules can be completed through integrated logic circuits in the hardware of the processor element or through software instructions.

[0292] For example, these modules can be one or more integrated circuits configured to implement the above methods, such as one or more application-specific integrated circuits (ASICs), one or more digital signal processors (DSPs), or one or more field-programmable gate arrays (FPGAs). As another example, when a module is implemented using processing element scheduler code, the processing element can be a general-purpose processor, such as a central processing unit (CPU) or other processor capable of calling program code. Furthermore, these modules can be integrated together to implement a system-on-a-chip (SOC).

[0293] This invention provides a computer device 100, which includes a processor and a non-volatile memory storing computer instructions. When the computer instructions are executed by the processor, the computer device 100 executes the aforementioned transportation infrastructure ecological collaborative planning device 110. Figure 4 As shown, Figure 4 This is a structural block diagram of a computer device 100 provided in an embodiment of the present invention. The computer device 100 includes a transportation infrastructure ecological collaborative planning device 110, a memory 111, a processor 112, and a communication unit 113.

[0294] To enable data transmission or interaction, the memory 111, processor 112, and communication unit 113 are electrically connected to each other directly or indirectly. For example, these components can be electrically connected to each other through one or more communication buses or signal lines. The transportation infrastructure ecological collaborative planning device 110 includes at least one software function module that can be stored in the memory 111 or embedded in the operating system (OS) of the computer device 100 in the form of software or firmware. The processor 112 is used to execute the transportation infrastructure ecological collaborative planning device 110 stored in the memory 111, such as the software function modules and computer programs included in the transportation infrastructure ecological collaborative planning device 110.

[0295] This invention provides a readable storage medium, which includes a computer program. When the computer program runs, it controls the computer device where the readable storage medium is located to execute the aforementioned transportation infrastructure ecological collaborative planning device 110.

[0296] For illustrative purposes, the foregoing description has been made with reference to specific embodiments. However, the foregoing illustrative discussions are not intended to be exhaustive or to limit the present disclosure to the precise forms disclosed. Numerous modifications and variations are possible in accordance with the foregoing teachings. These embodiments were chosen and described in order to best illustrate the principles of the present disclosure and its practical application, thereby enabling those skilled in the art to best utilize the disclosure and to employ various embodiments with different modifications to suit a particular intended application.

Claims

1. A method for ecologically coordinated planning of transportation infrastructure, characterized in that, include: Collect multi-source ecological planning image data and traffic remote sensing image data, and perform GIS georegistration on the multi-source ecological planning image data to convert it into registered ecological planning image data with spatial reference; Based on the image characteristics of the registered ecological planning image data and the traffic remote sensing image data, the segmentation parameters are adaptively adjusted to perform multi-scale segmentation, and vector graphic data corresponding to each of the multiple candidate areas are obtained. By overlaying multiple vector graphic data, the ecological-development index values ​​corresponding to each of the multiple candidate areas are calculated. Based on the ecological-development index value, the site selection of transportation infrastructure is carried out for the multiple candidate areas, and the transportation infrastructure selection results of the multiple candidate areas are obtained. The process of selecting transportation infrastructure sites for the multiple candidate areas based on the ecological-development index value, resulting in transportation infrastructure site selection results for the multiple candidate areas, includes: Based on the ecological-development index value, site selection for land-based shoreline facilities and water-based facilities are carried out in the multiple candidate areas respectively, to obtain the site selection results for land-based shoreline facilities or water-based facilities. Based on the aforementioned ecological-development index value, land-based shoreline facility site selection is performed on the multiple candidate areas to obtain the land-based shoreline facility site selection results, including: Multiple consecutive candidate areas with an ecological-development index value greater than a first preset index threshold are designated as candidate areas, and candidate areas with an ecological-development index value less than a second preset index threshold are designated as avoidance areas; the avoidance areas are used to determine buffer avoidance areas within a preset range. By optimizing the function: Calculate the planning suitability index for each of the proposed areas, and take the proposed area corresponding to the largest planning suitability index as the site selection result for the land shoreline facilities. Where TF is the planning suitability index, and k is the number of planned transportation infrastructure facilities. Let EDI be the value of the k-th suitable region. Let k be the length or area of ​​the k-th suitable region. Let α be the construction cost of the kth transportation infrastructure, and let α be the cost coefficient for reconstruction and expansion when existing infrastructure is occupied.

2. The method according to claim 1, characterized in that, Based on the image characteristics of the registered ecological planning image data and the traffic remote sensing image data, the segmentation parameters are adaptively adjusted to perform multi-scale segmentation, obtaining vector graphic data corresponding to multiple candidate areas, including: Based on the planning map features of the registered ecological planning image data, vector graphic data corresponding to the registered ecological planning image data of the multiple candidate areas are obtained by optimizing color blocks and text filtering, and by integrating morphological filtering to eliminate legend label interference. Based on the remote sensing image features of the traffic remote sensing image data, vector graphic data corresponding to the traffic remote sensing image data of the multiple candidate areas are obtained through texture enhancement and edge detection.

3. The method according to claim 2, characterized in that, The step of overlaying multiple vector graphic data to calculate the ecological-development index values ​​corresponding to each of the multiple candidate areas includes: Through the formula: The ecological-development index value corresponding to each of the candidate areas is calculated; in, The ecological-development index value. Let be the weight of the planning legal effect corresponding to the i-th vector graphic data of the candidate area. This represents the base score corresponding to the i-th vector graphic data of the candidate region; If there are multiple overlapping vector graphics data in the area to be selected, the minimum base score among the multiple overlapping vector graphics data shall be taken as the base score of the area.

4. The method according to claim 1, characterized in that, The site selection and delineation results for waterway facilities include waterway delineation results. Based on the ecological-development index value, waterway facility sites are selected for the multiple candidate areas to obtain the waterway facility site selection and delineation results, including: Acquire electronic nautical chart depth point cloud data; Based on the electronic nautical chart water depth point cloud data, multiple high-precision water depth grids are generated by Kriging interpolation. Extract continuous regions that meet the minimum navigable water depth threshold from the multiple high-precision water depth grids, and remove broken waterways whose length or width does not meet the preset distance. The continuous region is abstracted as a network node. The node spacing is defined as connected if it is less than a preset spacing threshold. The inverse of the EDI exponent corresponding to the continuous region is used as the edge weight. The Dijkstra algorithm is used to calculate the minimum cumulative weight path from the preset starting point to the ending point. An adaptive width corridor is generated along the path with the lowest cumulative weight. For the route segment that crosses the ecologically sensitive area, the route is oscillated within the adaptive width corridor to determine the route selection result.

5. The method according to claim 1, characterized in that, The site selection and delineation results for the water facilities include anchorage selection results. Based on the ecological-development index value, water facility sites are selected for the multiple candidate areas to obtain the water facility site selection and delineation results, including: Acquire electronic nautical chart depth point cloud data; Based on the electronic nautical chart water depth point cloud data, multiple high-precision water depth grids are generated by Kriging interpolation. The continuous regions that meet the minimum anchoring depth threshold are extracted from the multiple high-precision water depth grids, and broken water areas with an area smaller than the preset area threshold are removed to obtain the basic navigation area. From the basic navigation area, areas with an ecological-development index greater than a third preset index threshold are selected as preferred anchorage areas, and the anchorage selection results are obtained by sorting them in descending order.

6. The method according to claim 1, characterized in that, The method further includes: A simulation model of the ecological impact of the selected planning scheme is constructed using generative adversarial networks. The vector graphic data corresponding to the selected planning results of the transportation infrastructure is used as the input of the generator. The generated simulated regional ecological landscape is compared with the historical ecological baseline data by a discriminator to calculate the similarity of ecological impact. If the similarity is lower than the preset threshold, the boundary range of the delineation scheme is adjusted based on the gradient information fed back by the discriminator until the ecological impact similarity reaches the preset threshold, and the final delineation result is obtained.

7. The method according to claim 1, characterized in that, The method further includes: By constructing an ecological-transportation collaborative sensing network, real-time access is provided to ecological sensor data and traffic flow prediction data of the candidate area. Based on the ecological sensor data and the traffic flow prediction data, areas where the short-term fluctuations of the associated indicators exceed a preset range are marked as temporary restricted areas. After the fluctuations of the indicators in the temporary restricted areas stabilize, the corresponding traffic infrastructure selection results are re-evaluated.

8. A device for ecological collaborative planning of transportation infrastructure, characterized in that, include: The acquisition module is used to collect multi-source ecological planning image data and traffic remote sensing image data, and to perform GIS georegistration on the multi-source ecological planning image data to convert it into registered ecological planning image data with spatial reference; based on the image characteristics of the registered ecological planning image data and the traffic remote sensing image data, the segmentation parameters are adaptively adjusted to perform multi-scale segmentation to obtain vector graphic data corresponding to multiple candidate areas. The delineation module is used to overlay multiple vector graphic data to calculate the ecological-development index value corresponding to each of the multiple candidate delineation areas; and to select the location of transportation infrastructure for the multiple candidate delineation areas based on the ecological-development index value to obtain the transportation infrastructure delineation results for the multiple candidate delineation areas. The selection module is specifically used for: Based on the ecological-development index value, site selection for land-based shoreline facilities and water-based facilities are carried out in the multiple candidate areas respectively, to obtain the site selection results for land-based shoreline facilities or water-based facilities. The selection module is further specifically used for: Multiple consecutive candidate areas with an ecological-development index value greater than a first preset index threshold are designated as candidate areas, and candidate areas with an ecological-development index value less than a second preset index threshold are designated as avoidance zones; the avoidance zones are used to determine buffer avoidance zones within a preset range; through an optimization function: Calculate the planning suitability index for each of the proposed areas, and take the proposed area with the highest planning suitability index as the site selection result for the land shoreline facilities; where TF is the planning suitability index, and k is the number of planned transportation infrastructures. Let EDI be the value of the k-th suitable region. Let k be the length or area of ​​the k-th suitable region. Let α be the construction cost of the kth transportation infrastructure, and let α be the cost coefficient for reconstruction and expansion when existing infrastructure is occupied.

Images