A Geochemical Rasterized Contour Recognition and Digitization Method Based on AI Agent Skills
By using AI Agent Skills for identification and correction, the problem of contour line occlusion and breakage in geochemical raster maps has been solved, enabling continuous digitization and vectorization output of contour lines and providing a stable basis for chemical value domain inversion and spatial coordinate re-injection.
Patent Information
- Application Number
- CN202611578737.0
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-10-13
- Publication Date
- 2026-11-17
AI Technical Summary
In processing geochemical raster maps, existing technologies struggle to reliably determine the original trend of contour lines after they are obscured or broken, resulting in insufficient continuous digital connection between rasterized contour lines and image recognition and value vector results.
AI Agent Skills are used to identify map structure, correct map tilt and edge deformation, limit the recognition range by contour net boundary, extract contour color layer and line pixels, generate contour candidate segment pool, and form hidden contour corridor by endpoint extension projection and direction envelope fitting, realize the generation of continuous candidate chain of contour, and finally generate valued contour vector chain.
It enables candidate line segments that are truncated by text to be reconnected to the same isoline chain position, so that the continuity relationship between the fracture endpoint, the text occlusion position and the hidden isoline corridor can be solidified with the chain position, providing a continuous line position basis for subsequent chemical value domain inversion and spatial coordinate refilling, and achieving rapid identification of geochemical rasterized isolines and value vectorization output.
Smart Images

Figure CN123120074A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of geochemical digitization technology, and in particular to a method for geochemical rasterized contour line identification and digitization based on AI Agent Skills. Background Technology
[0002] Geochemical survey results are typically stored as raster maps, scanned maps, or thematic maps. Related technologies mostly rely on geographic information platforms and image processing software to complete map registration, map frame cropping, color layering, edge detection, linear target extraction, and vector result processing. The processing usually involves first establishing map coordinate references based on the map frame or control points, then extracting contour color layers and linear pixels within the target map area, and finally generating contour results that can be loaded into the geographic information environment through manual verification or semi-automatic line vectorization.
[0003] In geochemical raster maps that include title text, element labels, sampling point text, and local scan deformation, existing processing methods typically rely on color thresholds, edge continuity, or manual line addition to cover obscured lines. However, since text edges and contour lines may have similar colors, spatial overlaps, and mutual interference in their trends, it is difficult to reliably determine the original trend of the broken endpoints left after the contour lines are obscured. Subsequent chemical content value reapplying and spatial coordinate re-injection are easily affected by the broken chain positions, resulting in insufficient continuous digital connection between rasterized contour lines from image recognition to value vector results. Summary of the Invention
[0004] In view of the aforementioned existing problems, the present invention is proposed.
[0005] Therefore, this invention provides a geochemical raster contour line identification and digitization method based on AI Agent Skills to solve the problems of contour line occlusion and breakage and the difficulty in continuous digitization of numerical backfill in geochemical raster maps.
[0006] To solve the above-mentioned technical problems, the present invention provides the following technical solution: This invention provides a method for geochemical rasterized contour line identification and digitization based on AI Agent Skills, comprising: Geochemical raster maps are acquired. Based on the map structure recognition skill in AI Agent Skills, the map frame boundary and contour line carrying area are located. Map tilt and edge deformation are corrected along the map frame boundary, and the contour line carrying area is solidified. The map coordinate order and contour line clear boundary are written into the map structure location spectrum. Based on the geochemical map structure location spectrum, the recognition range is limited by the contour line clear boundary, and the extension direction of linear pixels is constrained by the map coordinate order. Contour line color layers and linear pixels are extracted from the defined contour line clear area to generate a contour line candidate segment pool. Text occlusion locations and fracture endpoints are extracted from the contour line candidate segment pool. Endpoint extension projection is performed on the fracture endpoints on both sides of the text occlusion location. The shadow and orientation envelope are fitted to form a hidden isoline corridor. Based on the hidden isoline corridor, the candidate isoline segment pool is divided into corridors to generate a continuous candidate isoline chain. According to the line continuity relationship and sampling point content constraints in the continuous candidate isoline chain, the chemical value domain is inverted to generate a two-dimensional chemical numerical field. The two-dimensional chemical numerical field is then back-verified by isolines, and the isoline values are back-applied to the continuous candidate isoline chain to generate a zone-valued isoline vector chain. The zone-valued isoline vector chain is then back-injected with coordinates to form spatial coordinate nodes. The spatial coordinate nodes and the zone-valued isoline vector chain are then consolidated into vector results to generate a geochemical rasterized isoline digitization result.
[0007] As a preferred embodiment of the geochemical rasterized contour line identification and digitization method based on AI Agent Skills described in this invention, the step of correcting map tilt and edge deformation along the map frame boundary is specifically as follows: The geochemical raster map is acquired, and the geochemical raster map is classified into the map structure localization task according to the map structure recognition skill in AI Agent Skills. The continuous outer edge of the geochemical raster map is locked to form a map access localization table. The continuous edges of the drawing frame lines reserved in the positioning table are included in the drawing frame boundary correction range. The drawing frame boundary is located according to the continuous edges of the drawing frame lines, and the drawing frame boundary is traction-corrected according to the parallel extension relationship of the horizontal edges and the intersection and closure relationship of the vertical edges to form the drawing frame correction benchmark.
[0008] As a preferred embodiment of the geochemical rasterized contour line identification and digitization method based on AI Agent Skills described in this invention, the formation of the geochemical map structure localization spectrum is specifically as follows: Extract the corrected map frame boundary from the map frame correction datum, delineate the map surface range inside the map frame according to the corrected map frame boundary, locate the contour line bearing area, and perform net area removal on the edge areas of the contour line bearing area that cannot bear the contour line trend to form a contour line bearing net area. The correction boundary sequence in the map frame correction datum is converted into boundary sequence coordinates to form the map coordinate sequence. The outer edge of the net area borne by the contour lines is written into the net area boundary of the contour lines. The map coordinate sequence and the net area boundary of the contour lines are linked to the same map positioning position to form a geochemical map structure positioning spectrum.
[0009] As a preferred embodiment of the geochemical rasterized contour line identification and digitization method based on AI Agent Skills described in this invention, the identification range is defined by the net boundary of the contour lines, and the extension direction of the linear pixels is constrained by the map coordinate order, as follows: The identification range of the net contour boundary in the geochemical map structure positioning spectrum is limited. The image area within the net contour boundary is retained as the net contour identification window, and the image area outside the net contour boundary is moved out of the linear pixel extraction range to form the net contour identification window. The map coordinates are sequentially transferred into the contour net area recognition window, and the pixel extension direction of the linear pixels in the contour net area recognition window is calibrated and unified according to the map coordinates to form a pixel extension reference.
[0010] As a preferred embodiment of the geochemical rasterized contour line identification and digitization method based on AI Agent Skills described in this invention, the generation of the contour line candidate segment pool is specifically as follows: Based on the pixel extension benchmark, the contour line color layer is screened in the contour line net area recognition window, and the pixels that are continuously distributed and can correspond to the color features of geochemical contour lines are anchored to the same contour line color layer to form the contour line color layer. Based on the pixel extension benchmark, color pixels that are in the same line position range and extend adjacently are selected from the contour color layer, and pixel combinations that are adjacent in position and have continuous direction are connected to form a linear pixel queue. The linear pixel queue is extended and verified according to the coordinate order of the map to obtain a candidate queue in the same direction. The main direction angle and endpoint pixel distance of adjacent line positions are checked according to the extension end of the candidate queue in the same direction. Line positions whose main direction angle falls into the direction tolerance and whose endpoint pixel distance falls into the interval tolerance are connected to the same candidate line segment position. Line positions whose main direction angle exceeds the direction tolerance and whose endpoint pixel distance exceeds the interval tolerance are listed as candidate line segment positions separately to form a contour line candidate segment pool.
[0011] As a preferred embodiment of the geochemical rasterized contour line identification and digitization method based on AI Agent Skills described in this invention, the formation of concealed contour line corridors is specifically as follows: The candidate contour line segment pool is positioned by occluding the endpoints, locating the text occlusion position and the fracture endpoints extending toward the text occlusion position. At the same time, the fracture endpoints located on both sides of the same text occlusion position and with opposite endpoint orientations are written into the same occlusion endpoint pairing position to form an occlusion endpoint pairing table. Using the broken endpoint in the occlusion endpoint pairing table as the projection starting point, push outwards into the text occlusion position according to the end direction of the candidate line segment where the broken endpoint is located, forming the opposing projection trajectory within the occlusion, and perform opposing projection fitting to form the endpoint projection corridor. The continuous passage range between the two projection trajectories is gathered along the extension direction of the endpoint projection corridor, and the outward expansion area that deviates from the endpoint direction is eliminated to generate the corridor direction envelope area. The corridor direction envelope area that is continuous in direction and does not cross the adjacent projection corridor is solidified into a hidden contour corridor, thus forming a hidden contour corridor.
[0012] As a preferred embodiment of the geochemical rasterized contour line identification and digitization method based on AI Agent Skills described in this invention, the generation of a continuous candidate chain of contour lines is specifically as follows: The two ends of the hidden contour line corridor are back-attached to the corresponding break endpoints in the candidate contour line segment pool. Hidden contour line corridors that can continue the original trend are written into the reliable position of the corridor, and hidden contour line corridors that cannot continue the original trend are written into the position to be excluded from the corridor, thus forming a corridor back-proof list. Based on the reliable location of the corridor, the candidate contour line segment pool is divided into corridors for continuation. The candidate line segments that receive the evidence through the line position are connected to the same contour line chain position along the hidden contour line corridor and the continuous chain position is solidified. The connection relationship between the fracture endpoint, the text occlusion position and the hidden contour line corridor is solidified into the same chain position, generating a continuous candidate contour line chain.
[0013] As a preferred embodiment of the geochemical rasterized contour line identification and digitization method based on AI Agent Skills described in this invention, the generation of the two-dimensional chemical numerical field is specifically as follows: Candidate segments that are continuously connected within the same contour line continuous candidate chain are expanded into a continuous sequence of line positions according to the contour line trend. The break and reconnection relationship of the continuous sequence of line positions is marked, and the positions that pass through the hidden contour line corridor are written into the line position continuity relationship to form a line position continuity constraint spectrum. The sampling point neighborhood is attached to the continuous constraint spectrum of the line position. The sampling point content close to both sides of the continuous sequence of the line position is attached to the corresponding chain position neighborhood. The high and low value side positions of the attached sampling point content are then separated to form the sampling point content constraint spectrum. Based on the linear continuity constraint spectrum and the sampling point content constraint spectrum, the value range variation is expanded. The linear continuity relationship of the continuous candidate chain of contour lines is used as the value range boundary, and the sampling point content is used as the value range traction point. The chemical content variation relationship is expanded along the grid nodes between adjacent linear positions. The grid value range is filled on the expanded chemical content variation relationship to generate a two-dimensional chemical numerical field.
[0014] As a preferred embodiment of the geochemical rasterized contour line identification and digitization method based on AI Agent Skills described in this invention, the generation of the valued contour line vector chain is specifically as follows: For grid nodes with the same chemical content value and adjacent connectivity in a two-dimensional chemical numerical field, the same value adjacency tracking is performed. Continuous line positions are extracted based on the isovalue boundaries of adjacent grid nodes to generate the same value line positions. The chain position direction of the same value line positions is then verified to generate the same value line verification spectrum. Based on the fit between the position of the isoline and the continuous sequence of the line position, the reattached chain position is locked from the isoline backtesting spectrum, and the chain positions that pass through the hidden isoline corridor are merged into the same locked value range. At the same time, conflict stripping is performed on the numerical jump position, the isoline value is written along the retained chain position, and the strip line position is solidified according to the chain position extension order to generate the strip line vector chain.
[0015] As a preferred embodiment of the geochemical rasterized isoline identification and digitization method based on AI Agent Skills described in this invention, the generation of digitized geochemical rasterized isolines is specifically as follows: Extract chain nodes according to the extension order of the value contour vector chain, fill in the chain nodes that pass through the hidden contour corridor, attach the extracted chain nodes with values along the chain, and simultaneously retain the chain nodes with values along the same contour line to generate a value chain node sequence. According to the map coordinate order in the geochemical map structure positioning spectrum, the pixel chain positions in the value chain position node sequence are backfilled to the spatial positions corresponding to the map coordinate order. The backfilled node positions are checked for map frame boundary fitting, and abnormal nodes that deviate from the net boundary of the contour line are removed to form spatial coordinate nodes. Based on the chain extension order of the valued contour vector chain, spatial coordinate nodes are connected into spatial contour chains, and the spatial contour chains are bound with the same value attribute. The same value line value, chain order and hidden contour corridor continuation mark are bound to the corresponding spatial contour chain to form a valued spatial contour chain. The vector results of the spatial isoline chain with values are consolidated and solidified. The vector results are encapsulated according to the spatial coordinate nodes, spatial isoline chains, isoline values, and chain position continuity relationships. The encapsulated vector results are then organized and arranged according to the spatial coordinate fields, isoline value fields, and chain position continuity fields required by GIS, generating geochemical rasterized isoline digitization results.
[0016] The beneficial effects of this invention are as follows: by extracting the text-occluded position and the fracture endpoint from the candidate contour line segment pool, and solidifying the hidden contour line corridor with the endpoint extension projection, the endpoint projection corridor and the corridor direction envelope, the candidate line segments cut off by the text-occluded position are reconnected to the same contour line chain position. This allows the continuity relationship between the fracture endpoint, the text-occluded position and the hidden contour line corridor to be solidified with the chain position, providing a continuous line position basis for subsequent chemical value domain inversion, isopleth numerical reapplied and spatial coordinate refilling, achieving the effect of rapid identification of geochemical rasterized contour lines and vectorized output of values. Attached Figure Description
[0017] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the following description of the embodiments will be briefly introduced. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.
[0018] Figure 1 This is a flowchart of a geochemical rasterized contour line identification and digitization method based on AI Agent Skills.
[0019] Figure 2 A flowchart for generating structural localization spectra for geochemical maps.
[0020] Figure 3 This is a flowchart for the continuation of the concealed contour line corridor.
[0021] Figure 4 A flowchart for generating vector chains of contour lines with values and digital results. Detailed Implementation
[0022] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings.
[0023] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the spirit of the invention. Therefore, the invention is not limited to the specific embodiments disclosed below.
[0024] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in different places in this specification does not necessarily refer to the same embodiment, nor is it a single or selective embodiment that is mutually exclusive with other embodiments.
[0025] Reference Figures 1-4 This is one embodiment of the present invention, which provides a method for geochemical rasterized contour line identification and digitization based on AI AgentSkills, including the following steps: S1: Acquire geochemical raster maps, locate map frame boundaries and contour line bearing areas based on map structure recognition skills in AI Agent Skills, correct map tilt and edge deformation along map frame boundaries, perform net domain solidification on contour line bearing areas, write map coordinate order and contour line net domain boundaries, and form geochemical map structure positioning spectrum.
[0026] S1.1: Obtain geochemical raster maps, and simultaneously classify geochemical raster maps into map structure localization tasks based on map structure recognition skills in AI Agent Skills, and perform map edge locking on the continuous outer edges of geochemical raster maps to form a map access localization table.
[0027] Furthermore, when acquiring geochemical raster maps, the map structure recognition skill in AI Agent Skills first verifies whether the geochemical raster map contains a complete map outer edge and identifiable map frame lines. Geochemical raster maps capable of supporting map frame boundary positioning are then assigned to the map structure localization task. After the geochemical raster map is assigned to the map structure localization task, the continuous outer edge is traced segment by segment along the outer pixel area of the geochemical raster map, eliminating interference from map corner damage, scan blanks, and non-map frame decorative lines, so that the continuous outer edge is retained as the starting range for map edge locking.
[0028] Map edge locking revolves around the continuous outer edge of the geochemical raster map. The horizontal and vertical outer edges are located according to the pixel connection relationship of the continuous outer edges. The closed connection positions between the horizontal and vertical outer edges are checked. The outer edge lines that can enclose the main map surface of the geochemical raster map are kept within the map edge locking range. Scattered lines that cannot be connected into a closed edge are moved out of the map edge locking range. The retained outer edge lines, the outer edge connection positions, and the map coverage area are written into the same access position to form a map access positioning table.
[0029] It should be noted that the map structure recognition skill in AI Agent Skills plays a crucial role in the entire process, including map structure task incorporation, recognition range guidance, and subsequent coordinate transfer. The map structure recognition skill in AI Agent Skills incorporates geochemical raster maps into the map structure localization task, making the continuous outer edges of the geochemical raster map, the continuous edges of the map frame lines, and the contour-bearing areas the processing objects within the same map structure localization process. The map structure recognition skill further guides map edge locking, map frame boundary localization, map frame correction, and the solidification of the contour-bearing area's net domain, ensuring that the map coordinate order and the contour net domain boundary can be written into the geochemical map structure localization spectrum. The geochemical map structure localization spectrum further constrains contour color layer screening, linear pixel continuation, hidden contour corridor prediction, chemical value range inversion, and coordinate backfeeding, ensuring that the subsequent contour recognition and digitization processes always fall within the range defined by the contour net domain boundary and the map coordinate order.
[0030] S1.2: Connect the continuous edges of the drawing frame lines reserved in the drawing frame access positioning table to the drawing frame boundary correction range, locate the drawing frame boundary according to the continuous edges of the drawing frame lines, and perform drawing frame traction correction on the drawing frame boundary according to the parallel extension relationship of the horizontal edges and the cross closure relationship of the vertical edges to form the drawing frame correction benchmark.
[0031] The continuous edge of the map frame line is the map frame line position that can be continuously connected along the outer edge of the geochemical raster map and participate in the map frame boundary positioning when the map is included in the positioning table.
[0032] Furthermore, continuous edges of the map frame line are extracted from the map access positioning table, and these continuous edges are incorporated into the map frame boundary correction range. The continuity of the line positions is checked segment by segment along the continuous edges of the map frame line. The horizontal and vertical edges that can enclose the main surface of the geochemical raster map are extracted, and scattered line positions formed by scanning blanks, outer edge stains, and non-map frame decorative lines are excluded, so that the map frame boundary correction range retains only the continuous line positions that can participate in the map frame boundary positioning.
[0033] The frame boundary positioning is carried out around the continuous edges of the frame line within the frame boundary correction range. The upper and lower edges are aligned according to the parallel extension relationship of the horizontal edges, and the left and right edges are checked for end closure according to the cross closure relationship of the vertical edges. The connection positions between the horizontal and vertical edges that can form a closed frame fall into the same frame boundary. The frame traction correction pulls back the skewed lines according to the horizontal and vertical edges of the closed frame, so that the skewed edges are aligned with the parallel extension direction of the horizontal edges, and the curved edges are aligned with the closure direction of the vertical edges, forming the frame correction benchmark.
[0034] S1.3: Extract the corrected map frame boundary from the map frame correction datum, delineate the map surface range inside the map frame according to the corrected map frame boundary, locate the contour line bearing area, and perform net area removal on the edge areas of the contour line bearing area that cannot bear the contour line trend to form a contour line bearing net area.
[0035] Furthermore, the corrected map frame boundary is extracted from the map frame correction datum. The inner lines of the horizontal and vertical edges are located along the corrected map frame boundary, and the intersection of the inner lines of the horizontal and vertical edges is connected to form a closed map frame outline. The main image surface with continuous pixels is retained inward along the closed map frame outline, while the scan blank, text outside the map frame, and outer edge interference lines are moved outward to form the map surface range inside the map frame. The positions of the continuous distribution of linear graphics are located within the map surface range inside the map frame, and the main image surface that can support the contour trend is retained to form the contour carrying area.
[0036] Extract the corrected map frame boundary from the map frame correction datum. Locate the inner lines of the horizontal and vertical edges along the corrected map frame boundary. Connect the intersections of the inner lines of the horizontal and vertical edges to form a closed map frame outline. Retain the main map surface within the closed map frame outline as the map frame's internal map surface range. Locate the positions of continuously distributed linear graphics within the map frame's internal map surface range, retaining the main map surface capable of supporting the contour line trend, forming the contour line carrying area. Remove the title page content, explanatory content, edge blanks, and isolated lines along the outer edge of the contour line carrying area, retaining the edge lines capable of continuing the contour line trend. Connect the retained edge lines to form the outer edge of the net area, forming the contour line carrying net area.
[0037] It should be noted that the area within the map frame is enclosed by the corrected map frame boundary, retaining the main map surface of the geochemical raster map and removing the scanned blank space and outer edge interference lines outside the map frame.
[0038] S1.4: Convert the correction boundary sequence in the map frame correction datum into boundary sequence coordinates to form the map coordinate sequence, and write the outer edge of the net area carried by the contour lines into the net area boundary of the contour lines. Then, link the map coordinate sequence and the net area boundary of the contour lines to the same map positioning position to form a geochemical map structure positioning spectrum.
[0039] Furthermore, the correction boundary sequence is extracted from the correction datum of the drawing frame, and the connection positions of the horizontal and vertical edges are located segment by segment along the corrected drawing frame boundary. The horizontal coordinate progression relationship of the drawing is recorded according to the extension direction of the horizontal edge, and the vertical coordinate progression relationship of the drawing is recorded according to the closing direction of the vertical edge. The drawing positioning position is fixed at the intersection of the horizontal coordinate progression relationship and the vertical coordinate progression relationship.
[0040] The edge positions of the net area are extracted segment by segment along the outer edge of the net area borne by the contour lines. The edge positions of the net area that can enclose the net area borne by the contour lines are written into the net area boundary of the contour lines. The map coordinate sequence and the net area boundary of the contour lines are linked to the same map positioning position. The map positioning position simultaneously carries the correspondence between the map coordinate sequence and the net area boundary of the contour lines. The subsequent screening of the contour color layer and the constraint of the extension direction of the linear pixels fall within the same map range, forming a geochemical map structure positioning spectrum.
[0041] S2: Based on the geochemical map structure positioning spectrum, the identification range is defined by the net boundary of the contour lines, and the extension direction of the linear pixels is constrained by the map coordinate order. The contour line color layer and linear pixels are extracted in the same position for the defined contour line net area to generate a pool of contour line candidate segments.
[0042] S2.1: The identification range of the net contour boundary in the geochemical map structure positioning spectrum is limited. The image area within the net contour boundary is retained as the net contour identification window, and the image area outside the net contour boundary is moved out of the linear pixel extraction range to form the net contour identification window.
[0043] Furthermore, the closed-path tracing of the contour net area boundary in the geochemical map structure positioning spectrum is performed. The continuous connection status of the boundary line position is checked segment by segment along the contour net area boundary. The boundary line position that can enclose the contour-bearing net area is pressed into the same window edge position. The identification range is advanced inward along the window edge position to include the image area that can bear the contour color layer within the contour net area boundary into the same window bearing range, and the window outline is continued according to the turning position of the contour net area boundary.
[0044] According to the window outline after continuation, the linear pixel extraction range is locked. The label content, explanatory content and edge white space access path are truncated outward along the window outline, and the non-isoline positions are blocked outside the window outline. The image area where the isoline color layer is located is continued inward along the window outline. The image area within the isoline net area boundary falls into the same recognition window position, and the isoline net area boundary is written into the recognition window boundary position to form the isoline net area recognition window.
[0045] It should be noted that the access path for the map title content is the pixel extension path where the area containing the map title, scale, legend, and cartographic information in the geochemical raster map may have linear connectivity with the net area identification window of the contour lines.
[0046] The access path for the explanatory content refers to the text edge path between the area containing element names, hierarchical descriptions, annotations, and labels in the geochemical raster map and the net area recognition window of the contour lines, which may be mistakenly connected as linear pixels.
[0047] The edge blank access path is the outer edge connection path that may form interference lines between the blank area outside the map frame, the scan edge and the paper edge noise in geochemical raster maps and the net area recognition window of contour lines.
[0048] S2.2: Transfer the map coordinates sequentially into the contour net area recognition window, and perform pixel extension direction calibration and direction unification on the linear pixels in the contour net area recognition window according to the map coordinates to form a pixel extension reference.
[0049] Furthermore, when transferring the map coordinates sequentially into the contour net area recognition window, the horizontal and vertical coordinate progression relationships in the map coordinate sequence are retrieved. The horizontal coordinate progression relationship is attached to the horizontal window edge of the contour net area recognition window, and the vertical coordinate progression relationship is attached to the vertical window edge of the contour net area recognition window. The intersection of the horizontal and vertical window edges is written into the window positioning position of the contour net area recognition window.
[0050] Within the contour line net area recognition window, the line position direction is traced along the adjacent connection relationship of linear pixels. The horizontally extending linear pixels are aligned with the horizontal coordinate progression relationship, the vertically extending linear pixels are aligned with the vertical coordinate progression relationship, and the curved linear pixels are connected to the corresponding coordinate progression direction according to the direction change of the continuous pixels. Short line positions that deviate in direction are re-aligned, and isolated line positions that cannot be connected to the coordinate sequence of the map are temporarily not included in the direction unification range, thus forming a pixel extension benchmark.
[0051] S2.3: Based on the pixel extension benchmark, the contour color layer is screened in the contour net area recognition window, and the pixels that are continuously distributed and can correspond to the color characteristics of geochemical contour lines are anchored to the same contour color layer to form the contour color layer.
[0052] Furthermore, based on the pixel extension benchmark, the image area within the contour net area recognition window is scanned segment by segment along the unified direction of the pixel extension benchmark. Color pixels with the same extension direction are pressed into the same screening range, and color connected regions are divided according to the continuous connection relationship between color pixels. Color change boundaries are checked segment by segment along the color connected regions, and color change positions are separated from color continuous positions, so that color pixels that can be continuously extended are kept in the same color connected region.
[0053] The spatial distribution of color pixels is traced segment by segment along the color-connected region. Color pixels that are continuously distributed and can correspond to the color characteristics of geochemical contour lines are anchored to the same contour line color layer. Color pixels that have color jumps, broken distributions, or cannot continuously follow the contour line trend are removed from the contour line color layer. The correspondence between color pixels is continuously solidified along the contour line trend direction to form the contour line color layer.
[0054] It should be noted that the color-connected region is a pixel region within the contour net area recognition window that has similar color features, adjacent pixel positions, and can be continuously connected along the same contour trend.
[0055] S2.4: Based on the pixel extension benchmark, select color pixels that are in the same line position range and extend adjacently from the contour color layer, and perform line position continuation on pixel combinations that are adjacent in position and have continuous direction, and continuate pixel combinations that can follow the same contour trend into a linear pixel queue.
[0056] Furthermore, based on the pixel extension benchmark, color pixels are screened segment by segment along the color distribution direction of the contour color layer. Color pixels that fall within the same line position range and maintain an adjacent extension relationship are pressed into the same pixel combination position. The adjacency distance and extension direction between color pixels are checked segment by segment along the pixel combination position. Color pixels with continuous adjacency distance and consistent extension direction are kept in the same pixel combination range, while color pixels with broken adjacency distance and deviated extension direction are moved out of the pixel combination range, forming pixel combinations that are adjacent in position and continuous in direction.
[0057] Line positions are connected along adjacent and continuous pixel combinations. Adjacent pixel combinations are connected end to end to the same line position channel according to the contour trend. The connected pixel combinations are arranged into continuous lines along a unified direction. The connection relationship between the continuous lines and the contour trend is further verified. Pixel combinations that can continuously connect to the same contour trend are merged into the same linear pixel queue to form a linear pixel queue.
[0058] S2.5: Based on the coordinate order of the map, the linear pixel queue is extended in the same direction to obtain a candidate queue. The main direction angle and endpoint pixel distance of adjacent line positions are checked according to the extension end of the candidate queue. Line positions whose main direction angle falls into the direction tolerance and whose endpoint pixel distance falls into the interval tolerance are connected to the same candidate line segment position. Line positions whose main direction angle exceeds the direction tolerance and whose endpoint pixel distance exceeds the interval tolerance are listed as candidate line segments, forming a contour line candidate segment pool.
[0059] Furthermore, the line orientation is checked segment by segment along the beginning and end of the linear pixel queue. The extension direction of the linear pixel queue is aligned with the horizontal and vertical coordinate progression relationships in the image coordinate order. The changes in the orientation of the pixels before and after the alignment are checked. Linear pixel queues that can extend along the same coordinate progression direction are written into the same-direction candidate queue, while linear pixel queues that show backtracking, intersection, or deviation from the image coordinate order are removed from the same-direction candidate queue.
[0060] Based on the extension of the candidate queue in the same direction, the main direction angle and endpoint pixel distance of adjacent line positions are checked sequentially. Line positions whose main direction angle falls within the direction tolerance and whose endpoint pixel distance falls within the interval tolerance are connected to the same candidate line segment position. Line positions whose main direction angle exceeds the direction tolerance or whose endpoint pixel distance exceeds the interval tolerance are listed as candidate line segment positions separately. The continuous line positions in the candidate queue in the same direction are gathered according to the candidate line segment positions to form the contour line candidate segment pool.
[0061] It should be noted that the directional tolerance is determined based on the average directional fluctuation of the continuous linear pixel queue in the contour color layer, and is used to characterize the range of angles between adjacent line positions that can maintain the same direction of extension; the interval tolerance is determined based on the average line width and map resolution of the contour color layer, and is used to characterize the range of pixel distances between the extension ends of adjacent line positions that can be connected as the same candidate line segment.
[0062] The horizontal coordinate progression relationship is formed by continuously extracting the edge position sequence along the upper and lower horizontal edges of the frame correction datum, and then attaching the pixel positions inside the image according to the horizontal edge arrangement order. It is used to characterize the forward extension, foldback change and deviation state of linear pixels between horizontal reference edges.
[0063] The vertical coordinate progression relationship is formed by continuously extracting the edge position sequence along the left and right vertical edges of the frame correction reference, and then attaching the pixel positions inside the image according to the vertical edge arrangement order. It is used to characterize the forward extension, foldback change and deviation state of linear pixels between the vertical reference edges.
[0064] S3: Extract the text occlusion position and fracture endpoint from the contour candidate segment pool, perform endpoint extension projection and direction envelope fitting on the fracture endpoints on both sides of the text occlusion position to form a hidden contour corridor, and perform corridor continuation splitting on the contour candidate segment pool according to the hidden contour corridor to generate a continuous contour candidate chain.
[0065] S3.1: Position the occlusion endpoints of the contour candidate line segment pool, locate the text occlusion position and the fracture endpoints extending toward the text occlusion position, and write the fracture endpoints located on both sides of the same text occlusion position with opposite endpoint orientations into the same occlusion endpoint pairing position to form an occlusion endpoint pairing table.
[0066] Furthermore, candidate segments are retrieved segment by segment from the pool of candidate contour lines. The position of the line break is checked along the beginning and end of the candidate line segment. The occlusion area that overlaps with the text outline and covers the extension path of the contour line is found outside the position of the line break and the position of the text occlusion is located. The direction of the break is traced along the end of the candidate line segment. The end of the line segment that extends toward the position of the text occlusion is located as the break endpoint. The end of the line segment that deviates from the position of the text occlusion is excluded from the range of the occlusion endpoint.
[0067] Check the orientation of the broken endpoints on both sides of the same text occlusion position, and write the broken endpoints located on opposite sides of the text occlusion position and whose end directions can approach each other into the same occlusion endpoint pairing position; if there are multiple broken endpoints around the same text occlusion position, arrange the pairing order according to the degree of alignment of the endpoint orientation and the proximity of the line position, so that each occlusion endpoint pairing position corresponds to a set of broken endpoints that can enter the endpoint extension projection, forming an occlusion endpoint pairing table.
[0068] S3.2: Using the broken endpoint in the occlusion endpoint pairing table as the projection starting point, push outwards into the text occlusion position according to the end direction of the candidate line segment where the broken endpoint is located, forming the opposing projection trajectory within the occlusion, and perform opposing projection fitting to form the endpoint projection corridor.
[0069] The break endpoint is the end of the candidate line segment in the contour candidate line segment pool where the line segment is interrupted. That is, the end of the line segment left after the contour line position is truncated by text, annotation content or map noise. The break endpoint needs to retain the end direction and the position of the candidate line segment, which is used to push outward into the text occlusion position to form the opposing projection trajectory inside the occlusion.
[0070] Furthermore, using the broken endpoints in the occlusion endpoint pairing table as the projection starting point, the end direction is extracted along the end tangent direction of the candidate line segment where each broken endpoint is located, and this direction is extended point by point in the continuous pixel space of the image, so that the extension trajectory advances into the interior of the text occlusion position to form an initial extension path; the pixel passage relationship between the initial extension path and the text occlusion position coverage area is gradually aligned so that the extension path maintains consistency of direction and continuity of direction when entering the occlusion area.
[0071] After the opposing extending paths are formed, the extending paths from both sides of the same text occlusion position are joined segment by segment. The gaps, offsets or misalignments between the paths are continuously aligned so that the extending paths on both sides form a continuous path band within the occlusion area. The aligned path band is then solidified into an endpoint projection corridor.
[0072] S3.3: Along the extension direction of the endpoint projection corridor, the continuous passage range between the two projection trajectories is gathered, and the outward expansion area that deviates from the endpoint direction is eliminated to generate the corridor direction envelope area. The corridor direction envelope area that is continuous in direction and does not cross the adjacent projection corridor is solidified into a hidden contour corridor to form a hidden contour corridor.
[0073] Furthermore, the projection trajectories on both sides are traced segment by segment along the extension direction of the endpoint projection corridor. The pixel passage range that can be continuously connected between the two projection trajectories is extracted, and the excessively wide passage boundary is compressed according to the end direction of the broken endpoint. The outward expansion area that spreads to the outside of the text occlusion position, deviates from the end direction of the broken endpoint, or is close to the adjacent projection corridor in the passage range is removed, so that a stable extension corridor range is retained between the two projection trajectories, and the corridor direction envelope area is generated.
[0074] The continuity of the direction and the crossing status of adjacent projected corridors are checked segment by segment for the corridor direction envelope area. The direction envelope area of the corridor that extends continuously along the end of the fault point and does not cross the adjacent projected corridor is solidified as the range of connectable lines within the shading area. The direction envelope area of the corridor that changes abruptly, crosses into the adjacent projected corridor, or cannot connect the ends of the faults on both sides is excluded from the range of connectable lines, forming a hidden contour corridor.
[0075] It should be noted that the continuous passage range is the pixel range in the endpoint projection corridor located between the two projection trajectories, which can be continuously connected along the end of the fracture endpoint and allow the concealed contour line corridor to pass through.
[0076] S3.4: Back-attach the two ends of the hidden contour line corridor. Back-attach the two ends of the hidden contour line corridor to the corresponding break endpoints in the contour line candidate segment pool. Hidden contour line corridors that can continue the original trend are written into the reliable position of the corridor, and hidden contour line corridors that cannot continue the original trend are written into the position to be excluded from the corridor, thus forming a corridor back-proof list.
[0077] Furthermore, the first and last lines of the corridor are extracted from the concealed contour line corridor. The first line of the corridor is then back-attached to one side of the fracture endpoint in the contour line candidate segment pool, and the last line of the corridor is back-attached to the other side of the fracture endpoint in the contour line candidate segment pool. During the back-attachment process, the directional connection between the end line of the corridor and the end of the fracture endpoint is checked, and the positional fit between the end line of the corridor and the original line of the candidate segment is checked, thus forming the back-attachment relationship between the two ends of the corridor.
[0078] Based on the posting relationship at both ends of the corridor, the connection status between the hidden contour corridor and the original line trend is checked segment by segment. Hidden contour corridors whose extension direction at the end of the corridor is continuous with the direction of the corresponding fracture endpoint and whose internal line position does not deviate from the original line trend of the candidate line segment are written into the reliable position of the corridor. Hidden contour corridors whose end of the corridor is misaligned with the direction of the fracture endpoint and whose internal line position deviates from the original line trend of the candidate line segment are written into the position to be excluded from the corridor, forming a corridor verification list.
[0079] S3.5: Based on the reliable location of the corridor, the candidate contour line segment pool is divided into corridors for continuation. The candidate line segments that receive the evidence through the line position are connected to the same contour line chain position along the hidden contour line corridor and the continuous chain position is solidified. The connection relationship between the fracture endpoint, the text occlusion position and the hidden contour line corridor is solidified into the same chain position to generate a continuous candidate contour line chain.
[0080] Furthermore, the reliable location of the corridor is retrieved from the corridor return certificate list, and the corresponding hidden contour line corridor and the fracture endpoints on both sides are located according to the reliable location of the corridor. Candidate line segments that have already received return certificates through the line position are locked in the contour line candidate segment pool. The continuation order of the candidate line segments on both sides is arranged along the extension direction of the hidden contour line corridor, and the two ends of the corridor are respectively attached to the corresponding fracture endpoints. The candidate line segments that have not been written into the reliable location of the corridor are left in the unconnected range, so that the corridor connection diversion forms a reliable connection range and an unconnected range.
[0081] Connect the candidate line segments on both sides to the same contour line chain position along the credible continuation range, and continuously solidify the chain position order according to the original trend of the candidate line segment, the position of the fracture endpoint, and the position of the hidden contour line corridor; write the continuation relationship between the fracture endpoint, the text occlusion position, and the hidden contour line corridor in the same contour line chain position, so that the candidate line segment cut off by the text occlusion position is restored to a continuous chain position, generating a continuous candidate chain of contour lines.
[0082] It should be noted that the connection between the fracture endpoints, the text-obscured position, and the hidden contour line corridor refers to the chain connection relationship formed after the fracture endpoints located on both sides of the text-obscured position complete the line position reattachment with the two ends of the hidden contour line corridor, and the hidden contour line corridor passes through the interior of the text-obscured position and receives the trend of the candidate line segments on both sides.
[0083] S4: Based on the line continuity relationship and sampling point content constraints in the continuous candidate chain of contour lines, perform chemical value domain inversion on the continuous candidate chain of contour lines to generate a two-dimensional chemical numerical field. Then, perform isoline verification on the two-dimensional chemical numerical field and paste the isoline values back to the continuous candidate chain of contour lines to generate a valued contour line vector chain.
[0084] S4.1: Expand the candidate line segments that are continuously connected within the same contour line continuous candidate chain into a line position continuous sequence according to the contour line trend order, and mark the break and repair relationship of the line position continuous sequence. Write the position that passes through the hidden contour line corridor into the line position continuous relationship to form a line position continuous constraint spectrum.
[0085] Furthermore, candidate line segments that are already continuously connected within the same chain position are extracted from the continuous candidate chain of contour lines. The candidate line segments are arranged in order according to their successive positions in the contour trend. Adjacent line points are traced sequentially from the beginning to the end of the candidate line segment, and the extension direction between adjacent line points is recorded. The extension direction of the end of the previous candidate line segment is matched with the extension direction of the beginning of the next candidate line segment. Candidate line segments with continuous extension directions and whose end positions can be connected are connected to the same line position arrangement position to form a continuous line position sequence.
[0086] Based on the fixed fracture endpoints, text occlusion positions, and the continuity relationships between hidden contour corridors in the continuous candidate chain of contour lines, the fracture repair relationships of the continuous sequence of line positions are marked. The positions of candidate line segments that pass through the hidden contour corridors are marked as corridor repair positions, and the corridor repair positions are written into the line position continuity relationship. The continuous constraint content within the same chain position is collected according to the candidate line segment order, the line position continuity sequence, and the line position continuity relationship to form the line position continuity constraint spectrum.
[0087] S4.2: The sampling point neighborhood is attached to the continuous constraint spectrum of the line position. The sampling point content close to both sides of the continuous sequence of the line position is attached to the corresponding chain position neighborhood, and the high and low value side positions of the attached sampling point content are separated to form the sampling point content constraint spectrum.
[0088] Furthermore, a continuous linear sequence is extracted from the linear continuous constraint spectrum, and the chain position neighborhood is expanded along both sides of the continuous linear sequence. The sampling point map position and corresponding element content are extracted from the sampling point annotation content configured with the geochemical raster map or the content table of sampling points in the same batch. According to the nearest access position between the sampling point map position and the continuous linear sequence, the sampling point content close to both sides of the continuous linear sequence is attached to the corresponding chain position neighborhood, so that the sampling point content, the sampling point map position and the chain position neighborhood form a corresponding relationship.
[0089] When separating high and low value sides of the sampled point content after connection, the sampled point content values located on both sides of the continuous sequence of the line position within the same chain position neighborhood are checked. The sampled point content on the side with the dominant value is written into the high value side, and the sampled point content on the side with the non-dominant value is written into the low value side. When there are multiple sampled point contents within the same chain position neighborhood, the traction order is arranged according to the nearest access position between the sampled point map position and the continuous sequence of the line position, and the sampled point contents that can accept the changing direction of the chain position neighborhood are retained. The high value side, low value side, sampled point map position and corresponding chain position neighborhood are synchronously written into the continuous constraint spectrum of the line position to form the sampled point content constraint spectrum.
[0090] S4.3: Based on the linear continuity constraint spectrum and the sampling point content constraint spectrum, the value range variation is expanded. The linear continuity relationship of the continuous candidate chain of contour lines is used as the value range boundary, and the sampling point content is used as the value range traction point. The chemical content variation relationship is expanded along the grid nodes between adjacent linear positions. The grid value range is filled on the expanded chemical content variation relationship to generate a two-dimensional chemical numerical field.
[0091] Furthermore, the linear continuity relationship of the isoline continuous candidate chain is extracted from the linear continuity constraint spectrum. Grid nodes are arranged along the map area between adjacent linear positions, and adjacent linear positions are respectively used as the two sides of the value domain boundary of the grid node. According to the linear continuity relationship, the extension order between adjacent linear positions is checked. Grid nodes located between the same adjacent linear positions and not crossing other isoline continuous candidate chains are connected to the same value domain variation spread range. Grid nodes that cross other isoline continuous candidate chains or are separated from the adjacent linear position clamping relationship are excluded from the value domain variation spread range. The sampling point content, high-value lateral position, low-value lateral position and sampling point map position are extracted from the sampling point content constraint spectrum. The traction direction of the sampling point content is determined according to the nearest access position of the sampling point map position and the chain position neighborhood.
[0092] The expression for calculating the chemical content variation values of raster nodes is as follows: ; ; In the formula, This represents the node number of a raster node in a two-dimensional chemical numerical field, used to distinguish different raster nodes. Indicates the first Each grid node is used to write the chemical content variation value of the two-dimensional chemical numerical field; Indicates the first Content of high-value lateral sampling points corresponding to each grid node; Indicates the first Content of low-value lateral sampling points corresponding to each grid node; Indicates the first The content of sampling points corresponding to the most recently accessed location of each grid node According to the location of the sampling points on the map and the first The nearest access relationship between grid nodes is extracted from the sampling point content constraint spectrum; Indicates the first Each grid node to The distance between the corresponding sampling points on the map; Indicates the first Each grid node to The distance between the corresponding sampling points on the map; Indicates the first The changing index of each grid node; The sampling point traction adjustment coefficient (example range: 0.10 to 0.50) is set based on the sampling point density, content grading interval, and map scale relationship to adjust the traction intensity of the sampling point content corresponding to the nearest access position. The higher the value, the sparser the sampling point distribution, the larger the content grading interval, and the easier it is for the sampling point content corresponding to the nearest access position to change the trend of the grid node. It is used to adjust the influence intensity of the sampling point content corresponding to the nearest access position on the trend index. This indicates the stability of the content difference term, used to prevent the variability index from becoming unstable when the content difference between sampling points with high and low values is too small.
[0093] For empty raster nodes without chemical content variation relationships written in their adjacent order, raster value domain filling is performed to ensure that the empty raster nodes inherit the variation direction of adjacent raster nodes with written chemical content variation relationships. Raster nodes that cross other continuous candidate chains of contour lines, are separated from adjacent line positions, or cannot inherit the variation direction of high-value and low-value lateral positions are not written into the two-dimensional chemical numerical field. After raster value domain filling is completed, the raster nodes, chemical content variation values, sampling point content traction directions, and line position continuity relationships between adjacent lines are mapped to the same numerical field location, generating a two-dimensional chemical numerical field.
[0094] It should be noted that the expression is used to convert raster nodes between adjacent line positions into chemical content variation values in a two-dimensional chemical numerical field. The expression uses the content of low-value lateral sampling points as the starting point of the variation and the content of high-value lateral sampling points as the ending point of the variation. It determines which side the raster node is closer to by the map distance from the raster node to the high-value lateral and the low-value lateral, so that the raster node closer to the high-value lateral obtains a value closer to the high-value lateral sampling point content, and the raster node closer to the low-value lateral obtains a value closer to the low-value lateral sampling point content. At the same time, the variation index is adjusted by the content of the sampling point corresponding to the nearest access position, so that when multiple sampling points jointly affect the same raster node, the constraint of the nearest sampling point on the variation direction is given priority. This completes the continuous filling of blank raster areas between adjacent line positions and provides a numerical basis for subsequent extraction of equipotential line positions and generation of contour line vector chains.
[0095] S4.4: Perform same-value adjacency tracking on grid nodes with the same chemical content value and adjacent connectivity in the two-dimensional chemical numerical field. Extract continuous line positions based on the isovalue boundaries of adjacent grid nodes to generate same-value line positions. Then, perform chain position direction verification on the same-value line positions to generate same-value line verification spectrum.
[0096] The chemical content value is the geochemical element content value written into each grid node in the two-dimensional chemical numerical field. The chemical content value comes from the sampling point content constraint spectrum and is spread to the grid nodes between adjacent lines after being pulled by the line position continuity relationship in the line position continuity constraint spectrum.
[0097] A grid node is a map location node in a two-dimensional chemical numerical field, divided according to the area carried by contour lines. The grid node carries the corresponding map location, the connection relationship between adjacent nodes, and the chemical content value, which is used for subsequent same-value adjacency tracking and continuous line position extraction.
[0098] Furthermore, grid nodes are retrieved one by one from the two-dimensional chemical numerical field, and the chemical content values and connectivity between adjacent grid nodes are checked. Grid nodes with the same chemical content value and adjacent edges are connected to the same adjacent path. The next adjacent grid node is tracked along the adjacent path. When encountering inconsistent chemical content values, grid nodes that are disconnected, or grid nodes that cross the contour line carrying area, the connection is stopped, so that the grid nodes with the same value in the two-dimensional chemical numerical field form a continuous adjacent path with the same value.
[0099] Extract the isovalue boundaries between adjacent raster nodes along the continuous isovalue adjacency path, and connect the isovalue boundaries into continuous line positions according to the connection order of adjacent raster nodes to generate isovalue line positions; check the orientation fit relationship between the isovalue line positions and the continuous sequence of line positions segment by segment along the isovalue line positions, write the isovalue line positions that are orientation-connected, line positions adjacent and do not cross the continuous candidate chain of adjacent isovalue lines into the isovalue line backtest spectrum, and remove the isovalue line positions that are orientation-deviated, line positions misaligned or cross the continuous candidate chain of adjacent isovalue lines from the isovalue line backtest spectrum.
[0100] S4.5: Based on the fit between the position of the isoline and the continuous sequence of the line position, lock the reattached chain position from the isoline backtesting spectrum, and merge the chain positions that pass through the hidden isoline corridor into the same locked value range. At the same time, perform conflict stripping on the numerical jump position, write the isoline value along the retained chain position, and solidify the strip line position according to the chain position extension order to generate the strip line vector chain.
[0101] Furthermore, the positions and continuous sequences of isolines are extracted from the isoline backtesting spectrum. The fit between the isoline positions and the continuous sequences of the isolines is checked segment by segment along the isoline positions. The positions of continuous candidate chains that are adjacent to the isolines, have a continuous direction, and do not cross adjacent isolines are locked as reattached chain positions. For chain positions that pass through the hidden isoline corridor, they are merged into the same locking value range according to the connection relationship between the two ends of the corridor and the break endpoint, so that the chain positions on both sides of the text occlusion position and the hidden isoline corridor obtain the same isoline value connection position.
[0102] Continue to check the continuity of the same value line values around the locked value range. If there is a sudden change in the same value line value of adjacent positions in the same chain position, the same value line position deviates from the continuous sequence of the line position, or the value sequence of adjacent chain positions intersects, perform conflict stripping and move it out of the locked value range. Write the same value line values segment by segment along the reserved chain position, and continue the value line position according to the extension order of the reserved chain position. Solidify the same value line values, the continuous sequence of the line position, and the continuation mark of the hidden contour line corridor into the same vector chain position to generate the value contour line vector chain.
[0103] S5: Perform coordinate backfeeding on the vector chain of isolines with values to form spatial coordinate nodes, and solidify the vector results of the spatial coordinate nodes and the vector chain of isolines with values to generate digital results of geochemical rasterized isolines.
[0104] S5.1: Extract chain nodes according to the extension order of the value contour vector chain, fill in the chain positions that pass through the hidden contour corridor, attach the extracted chain nodes with values along the chain, and simultaneously retain the value of the same contour line along the chain nodes to generate a value chain node sequence.
[0105] Furthermore, the values of the contour lines, the values of the same contour lines, and the corridor continuation marks of the hidden contour lines are extracted from the contour line vector chain. According to the extension order of the contour line positions from the starting position to the ending position, the starting position, turning position, continuation position, and ending position of each contour line position are located in sequence. The positions that can represent the changes in the shape of the contour line and the continuation relationship of the continuation position are extracted as continuation nodes, and the continuation nodes are arranged according to the extension order of the contour line vector chain to form the initial arrangement of the continuation nodes.
[0106] First, locate the fracture endpoints at both ends of the concealed contour line corridor. Then, along the direction of the concealed contour line corridor, place supplementary points between the two fracture endpoints. The supplementary points do not change the contour line values; they are only used to fill in the missing chain nodes inside the text-covered areas, so that the chain nodes on both sides of the concealed contour line corridor can be arranged continuously in the same extension order.
[0107] The values of the same contour lines in the vector chain with value are written point by point along the arrangement order of the chain nodes, and the ordinary chain nodes, the positions of the hidden contour line corridor supplementary points and the subsequent chain nodes are attached to the same range of the same contour line values. The values of the same contour lines are retained synchronously with the chain nodes, indicating that the position order of each chain node in the vector chain with value and the corresponding value of the same contour line are retained together, generating a sequence of chain nodes with value.
[0108] It should be noted that the extension order is the order in which the value lines in the value contour vector chain are continuously connected from the starting chain position to the ending chain position according to the trend of the contour lines. The extension order is used to determine the order of extraction of chain position nodes, the supplementary point position of the hidden contour line corridor connecting chain position, and the position of the same value line value attached to the chain.
[0109] S5.2: According to the map coordinate order in the geochemical map structure positioning spectrum, the pixel chain positions in the value chain position node sequence are backfilled to the spatial positions corresponding to the map coordinate order. The backfilled node positions are checked for map frame boundary fitting, and abnormal nodes that deviate from the net boundary of the contour line are removed to form spatial coordinate nodes.
[0110] Furthermore, the map coordinate sequence, corrected map frame boundary, and contour net boundary are extracted from the geochemical map structure positioning spectrum. The horizontal and vertical coordinate progression relationships in the map coordinate sequence are used as references for recharge. Pixel chain positions are retrieved point by point along the arrangement order of the value chain position node sequence. The horizontal pixel position of the pixel chain position in the map is linked to the horizontal coordinate progression relationship, and the vertical pixel position of the pixel chain position in the map is linked to the vertical coordinate progression relationship. The horizontal and vertical linkage positions intersect to form the corresponding spatial positions.
[0111] The boundary alignment check of the map frame is carried out around the node positions after recharge. The alignment relationship between the node positions and the horizontal and vertical edges is checked along the corrected map frame boundary. The node positions that fall within the map frame area and maintain the chain position extension order are retained within the spatial position range. The relative relationship between the node positions and the outer edge of the net contour area is checked along the contour line net area boundary. Node positions that deviate from the contour line net area boundary, cross the map frame area, or disrupt the sequential order of the value chain position node sequence are removed.
[0112] The removed node positions are reconnected according to the original order of the value chain position node sequence, and the corresponding isoline values, chain position order and hidden isoline corridor connection marks are retained simultaneously; the node positions that have completed the recharge and boundary verification are written into the spatial coordinate bearing position to form spatial coordinate nodes.
[0113] S5.3: Based on the chain extension order of the valued contour vector chain, the spatial coordinate nodes are connected into a spatial contour chain, and the spatial contour chain is bound with the same value attribute. The same value line value, chain order and hidden contour corridor continuation mark are bound to the corresponding spatial contour chain to form a valued spatial contour chain.
[0114] Furthermore, the chain extension order is extracted from the isoline vector chain, spatial coordinate nodes are retrieved according to the chain extension order, and the chain order, node positions before and after, and isoline values corresponding to the spatial coordinate nodes are verified; spatial coordinate nodes belonging to the same isoline vector chain and whose sequential order is continuous are connected end to end, and spatial coordinate nodes that pass through the hidden isoline corridor and are occupied by supplementary points are incorporated into the connection position between the fracture endpoints to form a spatial isoline chain.
[0115] After the spatial contour lines are connected, the continuity of the values of the same contour lines is checked point by point along the spatial contour line chain. Spatial coordinate nodes that maintain the same contour line values within the same spatial contour line chain are included in the same value attribute binding range. The same contour line values and chain position order are written to the spatial coordinate nodes corresponding to ordinary chain positions. The hidden contour line corridor connection mark is synchronously written to the spatial coordinate nodes that are connected through the hidden contour line corridor.
[0116] After the isopleth values, chain position order, and hidden isopleth corridor continuation marks are bound to the corresponding spatial isopleth chain, the spatial continuation relationship is solidified along the starting node, ordinary node, supplementary node, and ending node of the spatial isopleth chain, and the spatial coordinate node, isopleth values, and chain position continuation status are retained in the same spatial chain position to form a spatial isopleth chain with values.
[0117] S5.4: Perform vector result consolidation on the spatial isoline chain with values, encapsulate the vector results according to spatial coordinate nodes, spatial isoline chains, isoline values and chain position continuity, and organize the encapsulated vector results according to the spatial coordinate fields, isoline value fields and chain position continuity fields required by GIS, and generate geochemical rasterized isoline digitization results.
[0118] Furthermore, spatial coordinate nodes, spatial isoline chains, isoline values, and chain position continuity relationships are extracted from the spatial isoline chain. The spatial coordinate nodes are arranged in the order of the starting node, continuous nodes, hidden isoline corridor supplementary point nodes, and ending node of the spatial isoline chain. Adjacent spatial coordinate nodes are connected into the same vector line segment along the spatial coordinate node sequence, and the vector line segments in the same spatial isoline chain are classified into the same vector result chain position.
[0119] Vector results encapsulation revolves around the same vector result chain position. Spatial coordinate nodes are written into the vector geometric position, spatial contour line chains are written into the main body of the vector chain position, and isopleth values are written into the chain position attribute position. The relationships of ordinary chain position continuation, hidden contour line corridor continuation, and fracture endpoint repair are written into the chain position continuation relationship. Content with broken chain positions, missing isopleth values, or disordered spatial coordinate node order is removed from the vector result chain position, while vector results that can fully express the spatial trend of contour lines and isopleth attributes are retained.
[0120] The vector results are organized and packaged according to the spatial coordinate fields, isopleth value fields, and chain position continuation fields required by GIS. The spatial coordinate fields are mapped to spatial coordinate nodes, the isopleth value fields are mapped to isopleth values, and the chain position continuation fields are mapped to chain position continuation relationships and hidden isopleth corridor continuation marks. The organized vector results are solidified one by one according to the spatial isopleth chain to generate geochemical rasterized isopleth digitization results.
[0121] It should be noted that GIS retrieval refers to the process of loading, displaying, and querying the attributes of vector results by the geographic information spatial data processing environment; the spatial coordinate field, isoline value field, and chain continuation field required for GIS retrieval are used to carry the sequential relationship of spatial coordinate nodes, isoline values, and spatial isoline chains, respectively.
[0122] In summary, this invention extracts text-occluded locations and fracture endpoints from the pool of candidate contour lines, and solidifies the hidden contour line corridor using endpoint extension projection, endpoint projection corridor, and corridor direction envelope. This enables candidate lines truncated by text-occluded locations to be reconnected to the same contour line chain position, allowing the continuity relationship between fracture endpoints, text-occluded locations, and hidden contour line corridors to be solidified along with the chain position. This provides a continuous line position basis for subsequent chemical value domain inversion, isopleths numerical reapplied, and spatial coordinate refilling, achieving the effect of rapid identification and value vectorization output of geochemical rasterized contour lines.
[0123] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.
Claims
1. An AI Agent Skills-based geochemical gridding contour recognition and digitizing method, characterized in that, include: The geochemical raster map is acquired, and the map frame boundary and contour line bearing area are located based on the map structure recognition skill in AI Agent Skills. The map tilt and edge deformation are corrected along the map frame boundary, and the net area of the contour line bearing area is solidified. The map coordinate sequence and the net area boundary of the contour line are written to form the geochemical map structure positioning spectrum. Based on the geochemical map structure positioning spectrum, the identification range is defined by the net boundary of the contour lines, and the extension direction of the linear pixels is constrained by the map coordinate order. The contour line color layer and linear pixels are extracted in the same position from the defined contour line net area to generate a pool of contour line candidate segments. Extract text occlusion locations and fracture endpoints from the contour candidate segment pool. Perform endpoint extension projection and direction envelope fitting on the fracture endpoints on both sides of the text occlusion location to form a hidden contour corridor. Based on the hidden contour corridor, perform corridor continuation splitting on the contour candidate segment pool to generate a continuous contour candidate chain. Based on the line continuity relationship and sampling point content constraints in the continuous candidate chain of contour lines, chemical value range inversion is performed on the continuous candidate chain of contour lines to generate a two-dimensional chemical numerical field. The two-dimensional chemical numerical field is then back-verified by isolines, and the isoline values are back-applied to the continuous candidate chain of contour lines to generate a valued contour line vector chain. Coordinates are backfilled into the vector chain of contour lines with values to form spatial coordinate nodes. The vector results of the spatial coordinate nodes and the vector chain of contour lines with values are then consolidated and solidified to generate digital results of geochemical rasterized contour lines.
2. The geochemical rasterized contour line identification and digitization method based on AI Agent Skills as described in claim 1, characterized in that, The specific steps for correcting the tilt and edge deformation of the drawing along the boundary of the drawing frame are as follows: The geochemical raster map is acquired, and the geochemical raster map is classified into the map structure localization task according to the map structure recognition skill in AI Agent Skills. The continuous outer edge of the geochemical raster map is locked to form a map access localization table. The continuous edges of the drawing frame lines reserved in the positioning table are included in the drawing frame boundary correction range. The drawing frame boundary is located according to the continuous edges of the drawing frame lines, and the drawing frame boundary is traction-corrected according to the parallel extension relationship of the horizontal edges and the intersection and closure relationship of the vertical edges to form the drawing frame correction benchmark.
3. The geochemical rasterized contour line identification and digitization method based on AI Agent Skills as described in claim 2, characterized in that, The formation of the geochemical map structure localization spectrum is as follows: Extract the corrected map frame boundary from the map frame correction datum, delineate the map surface range inside the map frame according to the corrected map frame boundary, locate the contour line bearing area, and perform net area removal on the edge areas of the contour line bearing area that cannot bear the contour line trend to form a contour line bearing net area. The correction boundary sequence in the map frame correction datum is converted into boundary sequence coordinates to form the map coordinate sequence. The outer edge of the net area borne by the contour lines is written into the net area boundary of the contour lines. The map coordinate sequence and the net area boundary of the contour lines are linked to the same map positioning position to form a geochemical map structure positioning spectrum.
4. The geochemical rasterized contour line identification and digitization method based on AI Agent Skills as described in claim 3, characterized in that, The identification range is defined by the net boundary of the contour lines, and the extension direction of the linear pixels is constrained by the coordinate order of the map elements, as follows: The identification range of the net contour boundary in the geochemical map structure positioning spectrum is limited. The image area within the net contour boundary is retained as the net contour identification window, and the image area outside the net contour boundary is moved out of the linear pixel extraction range to form the net contour identification window. The map coordinates are sequentially transferred into the contour net area recognition window, and the pixel extension direction of the linear pixels in the contour net area recognition window is calibrated and unified according to the map coordinates to form a pixel extension reference.
5. The geochemical rasterized contour line identification and digitization method based on AI Agent Skills as described in claim 4, characterized in that, The specific steps for generating the candidate contour line segment pool are as follows: Based on the pixel extension benchmark, the contour line color layer is screened in the contour line net area recognition window, and the pixels that are continuously distributed and can correspond to the color features of geochemical contour lines are anchored to the same contour line color layer to form the contour line color layer. Based on the pixel extension benchmark, color pixels that are in the same line position range and extend adjacently are selected from the contour color layer, and pixel combinations that are adjacent in position and have continuous direction are connected to form a linear pixel queue. The linear pixel queue is extended and verified according to the coordinate order of the map to obtain a candidate queue in the same direction. The main direction angle and endpoint pixel distance of adjacent line positions are checked according to the extension end of the candidate queue in the same direction. Line positions whose main direction angle falls into the direction tolerance and whose endpoint pixel distance falls into the interval tolerance are connected to the same candidate line segment position. Line positions whose main direction angle exceeds the direction tolerance and whose endpoint pixel distance exceeds the interval tolerance are listed as candidate line segment positions separately to form a contour line candidate segment pool.
6. The geochemical rasterized contour line identification and digitization method based on AI Agent Skills as described in claim 5, characterized in that, The formation of the concealed contour line corridor is as follows: The candidate contour line segment pool is positioned by occluding the endpoints, locating the text occlusion position and the fracture endpoints extending toward the text occlusion position. At the same time, the fracture endpoints located on both sides of the same text occlusion position and with opposite endpoint orientations are written into the same occlusion endpoint pairing position to form an occlusion endpoint pairing table. Using the broken endpoint in the occlusion endpoint pairing table as the projection starting point, push outwards into the text occlusion position according to the end direction of the candidate line segment where the broken endpoint is located, forming the opposing projection trajectory within the occlusion, and perform opposing projection fitting to form the endpoint projection corridor. The continuous passage range between the two projection trajectories is gathered along the extension direction of the endpoint projection corridor, and the outward expansion area that deviates from the endpoint direction is eliminated to generate the corridor direction envelope area. The corridor direction envelope area that is continuous in direction and does not cross the adjacent projection corridor is solidified into a hidden contour corridor, thus forming a hidden contour corridor.
7. The geochemical rasterized contour line identification and digitization method based on AI Agent Skills as described in claim 6, characterized in that, The generation of a continuous candidate chain of contour lines is specifically as follows: The two ends of the hidden contour line corridor are back-attached to the corresponding break endpoints in the candidate contour line segment pool. Hidden contour line corridors that can continue the original trend are written into the reliable position of the corridor, and hidden contour line corridors that cannot continue the original trend are written into the position to be excluded from the corridor, thus forming a corridor back-proof list. Based on the reliable location of the corridor, the candidate contour line segment pool is divided into corridors for continuation. The candidate line segments that receive the evidence through the line position are connected to the same contour line chain position along the hidden contour line corridor and the continuous chain position is solidified. The connection relationship between the fracture endpoint, the text occlusion position and the hidden contour line corridor is solidified into the same chain position, generating a continuous candidate contour line chain.
8. The geochemical rasterized contour line identification and digitization method based on AI Agent Skills as described in claim 1, characterized in that, The generation of the two-dimensional chemical numerical field is as follows: Candidate segments that are continuously connected within the same contour line continuous candidate chain are expanded into a continuous sequence of line positions according to the contour line trend. The break and reconnection relationship of the continuous sequence of line positions is marked, and the positions that pass through the hidden contour line corridor are written into the line position continuity relationship to form a line position continuity constraint spectrum. The sampling point neighborhood is attached to the continuous constraint spectrum of the line position. The sampling point content close to both sides of the continuous sequence of the line position is attached to the corresponding chain position neighborhood. The high and low value side positions of the attached sampling point content are then separated to form the sampling point content constraint spectrum. Based on the linear continuity constraint spectrum and the sampling point content constraint spectrum, the value range variation is expanded. The linear continuity relationship of the continuous candidate chain of contour lines is used as the value range boundary, and the sampling point content is used as the value range traction point. The chemical content variation relationship is expanded along the grid nodes between adjacent linear positions. The grid value range is filled on the expanded chemical content variation relationship to generate a two-dimensional chemical numerical field.
9. The geochemical rasterized contour line identification and digitization method based on AI Agent Skills as described in claim 8, characterized in that, The generation of the contour line vector chain with values is as follows: For grid nodes with the same chemical content value and adjacent connectivity in a two-dimensional chemical numerical field, the same value adjacency tracking is performed. Continuous line positions are extracted based on the isovalue boundaries of adjacent grid nodes to generate the same value line positions. The chain position direction of the same value line positions is then verified to generate the same value line verification spectrum. Based on the fit between the position of the isoline and the continuous sequence of the line position, the reattached chain position is locked from the isoline backtesting spectrum, and the chain positions that pass through the hidden isoline corridor are merged into the same locked value range. At the same time, conflict stripping is performed on the numerical jump position, the isoline value is written along the retained chain position, and the strip line position is solidified according to the chain position extension order to generate the strip line vector chain.
10. The geochemical rasterized contour line identification and digitization method based on AI Agent Skills as described in claim 1 or 9, characterized in that, The specific details of generating the digitalized geochemical rasterized contour lines are as follows: Extract chain nodes according to the extension order of the value contour vector chain, fill in the chain nodes that pass through the hidden contour corridor, attach the extracted chain nodes with values along the chain, and simultaneously retain the chain nodes with values along the same contour line to generate a value chain node sequence. According to the map coordinate order in the geochemical map structure positioning spectrum, the pixel chain positions in the value chain position node sequence are backfilled to the spatial positions corresponding to the map coordinate order. The backfilled node positions are checked for map frame boundary fitting, and abnormal nodes that deviate from the net boundary of the contour line are removed to form spatial coordinate nodes. Based on the chain extension order of the valued contour vector chain, spatial coordinate nodes are connected into spatial contour chains, and the spatial contour chains are bound with the same value attribute. The same value line value, chain order and hidden contour corridor continuation mark are bound to the corresponding spatial contour chain to form a valued spatial contour chain. The vector results of the spatial isoline chain with values are consolidated and solidified. The vector results are encapsulated according to the spatial coordinate nodes, spatial isoline chains, isoline values, and chain position continuity relationships. The encapsulated vector results are then organized and arranged according to the spatial coordinate fields, isoline value fields, and chain position continuity fields required by GIS, generating geochemical rasterized isoline digitization results.
