Surveying data adaptive fusion method based on cognitive computing

Abstract

The present application relates to the field of surveying and mapping fusion technology, in particular to a surveying and mapping data adaptive fusion method based on cognitive computing, which comprises the following steps: obtaining point cloud records, cutting the point cloud records into segments according to boundaries and rearranging to generate strip records, comparing differences to construct continuous area labels and fuse, extracting registration error directions to generate stable strike sections, summarizing the contents of overlapping segments to combine fusion entries, superimposing and compressing drawing lines to attach images to output fusion results. The present application cuts point clouds according to scanning order and constructs spatial strip records, forms aggregated information with sequential characteristics, constructs continuous area sequences through difference comparison, establishes the connection relationship between trajectories and segments in combination with the registration error direction and the spatial position, generates synchronous reference entries through the fusion and combination of point clouds and images, combines drawing line compression and image attachment operations, and integrates multiple groups of data contents to form a unified structure.

Description

Technical Field

[0001] This invention relates to the field of surveying and mapping fusion technology, and in particular to an adaptive fusion method for surveying and mapping data based on cognitive computing. Background Technology

[0002] The field of surveying and mapping fusion technology encompasses topographic surveying, photogrammetry, lidar mapping, and multi-source spatial data fusion. Its core function is to process heterogeneous surveying and mapping data, such as satellite imagery, aerial imagery, lidar point clouds, and geodetic data, through registration, resampling, correction, and fusion, to unify their representation within the same coordinate system or spatial reference system, generating digital elevation models, orthophotos, 3D terrain models, or geographic information layers. This field systematically includes data acquisition (e.g., satellite, UAV, ground-based lidar), data preprocessing (e.g., denoising, downsampling, geometric correction), data registration (e.g., image point cloud registration, inertial navigation data alignment), data fusion (e.g., multi-source fusion algorithms, fusion models), and the output of surveying and mapping products and services.

[0003] Among them, the adaptive fusion method for surveying and mapping data based on cognitive computing addresses the technical issues in the process of surveying and mapping data fusion, which involves fusing point cloud data, image data, and inertial navigation / positioning data from different sources. Specifically, it includes preprocessing the data from each source (e.g., denoising, downsampling, geometric correction, and error correction), then registering the data separately (e.g., converting point cloud data to geographic coordinates, transforming image data control points to geographic coordinates, calculating attitude and position information, and correcting inertial navigation data), and finally using a cognitive computing mechanism to adjust the weights or fusion strategies of each data source, so that the fusion model adaptively selects the data sources and fusion methods to participate in the fusion.

[0004] Existing technologies rely on unified registration strategies to process multi-source data. Under conditions of complex spatial orientation or path offset, it is difficult to effectively express changes in point cloud density distribution. It lacks an identification mechanism based on segmented aggregation differences, making it difficult to construct continuous spatial relationships. Registration records are only used for static alignment and do not combine error change trends to establish orientation stability judgments. This results in a lack of dynamic correspondence logic between trajectory segments and data segments. The fusion content is mostly executed at the overall level, lacking fine fusion methods for overlapping areas. Point clouds and images have fitting offsets in detailed parts, affecting the spatial resolution and structural integrity of the final fusion result. Summary of the Invention

[0005] To address the technical problems existing in the prior art, this invention provides an adaptive fusion method for surveying and mapping data based on cognitive computing. The technical solution is as follows: An adaptive fusion method for surveying and mapping data based on cognitive computing includes the following steps: S1: Acquire point cloud records generated during mobile mapping platform operations, divide them into continuous segments according to fixed spatial boundaries, extract the point aggregation positions within each segment and rearrange them in row and column order to form segment depiction results, and concatenate all segment depiction results in the acquisition order to generate a density depiction segment set. S2: Based on the density depiction fragment set, compare the aggregation position differences of adjacent fragments, select fragment groups whose differences exceed the control range, organize them into continuous regions according to the connection relationship in the land cover, mark the continuous regions with the dominant fusion mark, and generate the master control fusion fragment band. S3: Obtain the registration record of the surveying equipment and connect the trajectories in sequence according to time. Extract the trajectory segments with consistent continuous error changes and superimpose the corresponding spatial position markers. Integrate the spatial position of the trajectory segments with the spatial position of the main control fusion segment to generate a stable direction segment sequence. S4: Based on the stable directional segment sequence and the main control fusion segment band, extract the trajectory segments with overlapping spatial positions, summarize the corresponding point cloud and image depiction content one by one, combine the depiction content at the same position into fusion basic entries, and generate a fusion candidate segment set.

[0006] As a further embodiment of the present invention, the density depiction fragment set includes aggregation point coordinate distribution, fragment image sequence, and strip structure index; the master fusion fragment strip includes region continuity identifier, fusion priority marker, and spatial matching sequence; the stable orientation segment sequence includes registration direction information, continuous trajectory number, and spatial correspondence; and the fusion candidate fragment set includes fusion entry index, point cloud and image pairing results, and spatial overlapping fragment list.

[0007] As a further aspect of the present invention, the step of obtaining S1 is as follows: S101: Acquire point cloud records formed by the mobile mapping platform during field operations, and locate the three-dimensional coordinate data contained in each frame of point cloud according to the LiDAR scanning order. Based on the preset fixed spatial boundary configuration parameter set, clip the spatial position range of the point cloud data frame, split it into multiple point cloud segments according to the continuous scanning order, and number the spatial range index data of each segment to obtain the spatial boundary clipping sequence. S102: Based on the spatial index range of each point cloud segment in the spatial boundary clipping sequence, extract the position information of each three-dimensional coordinate point in the segment, identify the dense point group located in the local area, perform a sequence update operation by comparing the horizontal and vertical displacement relationship of each set of coordinates, and output the cluster point position order structure. S103: The coordinate sequences of each point cloud segment in the aggregation point positional structure are called in the order of their acquisition timestamps to perform a time-series connection operation, and the coordinate sequences corresponding to multiple segments are connected to the continuous spatial strip channel in chronological order to obtain a density depiction segment set.

[0008] As a further aspect of the present invention, the step of obtaining S2 is as follows: S201: Based on the adjacent segment data in the density depiction segment set, compare the corresponding aggregation point information in each segment in sequence, retrieve the displacement value difference in the horizontal and vertical directions, and perform item-by-item difference confirmation operation with the preset aggregation position comparison range. Screen out all segments whose differences exceed the spatial offset reference value and obtain the offset segment index column. S202: Call the segment number in the offset segment index column, extract the data frame of adjacent number according to the spatial scanning direction, identify all segments with consecutive numbers and located in the same spatial direction range, determine whether the number of consecutive occurrences in the sequence exceeds the regional grouping reference value, output the number set for segments that meet the sequence continuity rule, and obtain the segment chain set list. S203: Based on the numbered data in the fragment chain set list, perform a location matching operation on the area corresponding to each chain in the land cover data, and group the fragment chains with continuous positions according to the boundary connection relationship of the land features, and assign them master control attribute identification information to obtain the master control fusion fragment band.

[0009] As a further aspect of the present invention, the step of obtaining S3 is as follows: S301: Acquire the registration records formed by the mobile surveying equipment during its movement, and arrange all records in chronological order into a trajectory data frame sequence. Extract registration error information from each record in the sequence and perform direction judgment operation on the direction change of the error vector. By comparing the stability of the angle signs between adjacent error vectors, confirm the trajectory segments with consistent directions, and add corresponding spatial position identification codes to these segments to obtain the direction identification trajectory segment column. S302: Call the spatial position identifier code in the direction identifier trajectory segment column, sequentially retrieve the encoded data of each trajectory segment, extract the continuously numbered trajectory segments according to the spatial arrangement order of the trajectory segments, confirm the set of trajectory segments whose number change range is within the spatial continuous reference value range through the sequence continuity judgment operation, and output the spatial position sequence of these trajectory segments to obtain the continuous trajectory spatial column. S303: Based on the spatial position sequence in the continuous trajectory space column, the spatial position data of each segment in the main control fusion segment band are matched one-to-one with the coordinates of the two segments under the same spatial index system. The position segments of the continuous trajectory space sequence in the main control fusion segment band are extracted from the correspondence relationship, and the output sequence is arranged in spatial order to obtain the stable direction segment sequence.

[0010] As a further aspect of the present invention, the step of obtaining S4 is as follows: S401: Based on the stable orientation segment sequence and the main control fusion segment band, call the spatial position index information of each segment in the stable orientation segment, compare its position number with the position number in the main control fusion segment band under the spatial index system, perform a screening operation on the data frames with the same coordinate number, extract all segments with overlapping spatial positions, and obtain a list of spatially overlapping segments. S402: Based on the set of numbers in the list of spatially overlapping segments, extract the point cloud depiction content associated with the corresponding segment in the point cloud dataset and the corresponding image depiction content in the image dataset in sequence. Aggregate the two types of depiction content with the same spatial index in numerical order and compress them into an entry structure to obtain a point-image double-lined entry set. S403: Call the spatial numbering information in the dot shadow double-line entry set, arrange all entry structures according to the spatial orientation index order, organize the entry sequence according to the numbering change trend, output the entry dataset with continuous spatial numbering characteristics, and obtain the fusion candidate fragment set.

[0011] As a further aspect of the present invention, the method further includes: S5: Call the point cloud and image depiction content in the fusion candidate segment set, superimpose and compress multiple point cloud depictions into a single spatial depiction line at the same spatial position, attach the image depiction to the spatial depiction line at the corresponding position, rearrange all fusion segments in the direction order, and generate the mapping data fusion result. The mapping data fusion results include compressed point cloud line structure, image texture layer, and unified orientation fusion segments.

[0012] As a further aspect of the present invention, the step of obtaining S5 is as follows: S501: Call the point cloud depiction content in the fusion candidate fragment set, extract the three-dimensional coordinate values ​​of the point cloud line data of all entries in the same spatial position number according to the number order, and sequentially stack all the spatial paths of the same line segments along the direction to compress and form a linear unit structure to obtain the compressed spatial depiction line set. S502: Based on the path position index in the compressed spatial depiction line set, call the image depiction content under the corresponding number, locate the image area boundary coordinates on both sides of each path, and perform close-fitting processing on each image boundary and spatial path line, outputting a combination item structure composed of image boundaries and spatial lines to obtain a series of fitted image items. S503: Call the number sequence information in the image item column, extract all combined items according to the continuous spatial direction, construct a continuous direction structure according to the spatial arrangement of the numbers, integrate all combined items to form a continuous sequence, and obtain the mapping data fusion result.

[0013] The beneficial effects of the technical solutions provided in the embodiments of the present invention include at least the following: In this invention, by dividing point cloud data into continuous segments according to the scanning order and constructing spatial strip records, a set of aggregated information with sequential characteristics can be formed in the spatial structure. By comparing differences and combining continuous segment chains, a region sequence with associated logic is constructed. By matching the registration error direction with the spatial position, a stable connection relationship between the trajectory path and the spatial segment is established. By combining point cloud and image data at the fusion location, a fusion entry with a synchronous spatial reference is generated. By spatial line compression and image pasting operations, multiple sets of data content are linked together to form a unified structure, enhancing the matching completeness and spatial consistency of point cloud and image during the fusion process. Attached Figure Description

[0014] Figure 1 This is a flowchart of the method of the present invention; Figure 2 This is a flowchart illustrating the process of obtaining the density-description fragment set according to the present invention. Figure 3 This is a flowchart illustrating the acquisition process of the main control fusion segment band in this invention; Figure 4 This is a flowchart illustrating the process of obtaining the stable directional segment sequence in this invention. Figure 5 This is a flowchart illustrating the process of obtaining the candidate fragment set in this invention. Figure 6 This is a flowchart illustrating the process of obtaining the data fusion results of this invention. Detailed Implementation

[0015] The technical solution of the present invention will now be described with reference to the accompanying drawings.

[0016] In embodiments of the present invention, words such as "exemplarily," "for example," etc., are used to indicate that something is an example, illustration, or description. Any embodiment or design described as "exemplary" in the present invention should not be construed as being more preferred or advantageous than other embodiments or designs. Specifically, the use of the word "exemplary" is intended to present the concept in a concrete manner. Furthermore, in embodiments of the present invention, the meaning expressed by "and / or" can be both, or either one.

[0017] In the embodiments of this invention, the terms "image" and "picture" may sometimes be used interchangeably. It should be noted that, without emphasizing the distinction between them, they convey the same meaning. Similarly, the terms "of," "corresponding (relevant)," and "corresponding" may sometimes be used interchangeably. It should be noted that, without emphasizing the distinction between them, they convey the same meaning.

[0018] In this embodiment of the invention, sometimes a subscript such as W1 may be written in a non-subscript form such as W1. When the difference is not emphasized, the meaning they express is the same.

[0019] To make the technical problems, technical solutions and advantages of the present invention clearer, a detailed description will be given below in conjunction with the accompanying drawings and specific embodiments.

[0020] Please see Figure 1 This invention provides a technical solution: an adaptive fusion method for surveying and mapping data based on cognitive computing, comprising the following steps: S1: Acquire point cloud records formed by the mobile mapping platform during field operations, divide them into continuous segments according to the LiDAR scanning sequence and fixed spatial boundaries, read the three-dimensional coordinates point by point in each segment and record the point aggregation position, rearrange the aggregation positions in each segment according to the row and column order to form the segment depiction result, and string all the segment depiction results together into a spatial strip record according to the acquisition time order to generate a density depiction segment set. S2: Based on the density, the fragment set is described, and the differences in the aggregation position are compared one by one in the adjacent fragments. Fragments with differences exceeding the control range are recorded as candidate fragments. In the candidate fragments, the continuous fragment chains are searched in spatial order. These fragment chains are organized into continuous regions according to their connection relationship in the land cover. The continuous regions are marked with the dominant fusion mark to generate the master-controlled fusion fragment band. S3: Acquire the registration records formed by the mobile surveying equipment during its movement, arrange the records into a continuous trajectory in chronological order, read the registration error information of each time in the trajectory and record the direction of change, overlay the corresponding spatial position marker on the trajectory segments with consistent direction of change, and organize the spatial positions of these trajectory segments to correspond one-to-one with the spatial positions of the main control fusion segment to generate a stable trajectory segment sequence. S4: Based on the stable orientation segment sequence and the main control fusion segment band, select segments in the stable orientation segment whose spatial location overlaps with the main control fusion segment band, summarize the corresponding point cloud depiction content and image depiction content for each of these segments, combine the point cloud depiction and image depiction at the same location into fusion basic items, and then arrange all fusion basic items in order according to spatial orientation to generate a fusion candidate segment set. S5: Call the point cloud depiction and image depiction content in the fusion candidate fragment set, superimpose and compress multiple point cloud depictions into a single spatial depiction line at the same spatial location, attach the image depiction to the spatial depiction line at the corresponding location, rearrange all fusion fragments according to the direction on the entire spatial depiction line, and generate the mapping data fusion result.

[0021] The density depiction fragment set includes the distribution of aggregation point coordinates, fragment image sequence, and band structure index. The master fusion fragment band includes region continuity identifier, fusion priority marker, and spatial matching sequence. The stable orientation segment sequence includes registration direction information, continuous trajectory number, and spatial correspondence. The fusion candidate fragment set includes fusion entry index, point cloud and image pairing results, and spatial overlapping fragment list. The mapping data fusion results include compressed point cloud line structure, image texture layer, and unified orientation fusion fragment.

[0022] Please see Figure 2 The steps to obtain S1 are as follows: S101: Acquire point cloud records formed by the mobile mapping platform during field operations, and locate the three-dimensional coordinate data contained in each frame of point cloud according to the LiDAR scanning order. Based on the preset fixed spatial boundary configuration parameter set, clip the spatial position range of the point cloud data frame, split it into multiple point cloud segments according to the continuous scanning order, and number the spatial range index data of each segment to obtain the spatial boundary clipping sequence. To acquire point cloud records generated by a mobile mapping platform during field operations, each frame of point cloud data needs to be acquired via an external LiDAR sensor. The scanning sequence and 3D coordinate information must be recorded in real time. The scanning sequence can be configured by setting frame numbers starting from 1 and incrementing sequentially according to the trigger frequency. During the acquisition process, a spatial boundary configuration parameter set needs to be set. This parameter set requires preset 3D boundary values, where the ranges in the x, y, and z directions should be set in conjunction with the mapping area. For example, the following settings could be used: Based on this spatial boundary, the three-dimensional coordinates contained in each frame of the point cloud are... Perform point-by-point judgment, where for any point Conditions met: If a frame is selected, it is retained; otherwise, it is discarded. The cropped point cloud frames will be saved as the original fragment sequence in frame number order. To ensure continuity between frames, a maximum inter-frame distance threshold can be set. For example, if the Euclidean distance between any two points in the current frame and the previous frame does not exceed 1.5 meters, then... If the frames are consecutive, they are considered as a single segment; otherwise, they are split into new segments, each assigned a segment number. Simultaneously, it collects the spatial range index information corresponding to the segments, such as recording the minimum and maximum values ​​in each segment. Coordinate values ​​are used as index boundaries, such as in fragments. middle: This range is used as the boundary information of the cropping result. The spatial boundary index of each segment is summarized and saved as a spatial boundary cropping sequence: ,in Corresponding fragments The three-dimensional space cutting range.

[0023] S102: Based on the spatial index range of each point cloud segment in the spatial boundary clipping sequence, extract the position information of each three-dimensional coordinate point in the segment, identify the dense point group located in the local area, perform sequence update operation by comparing the horizontal and vertical displacement relationship of each set of coordinates, and output the cluster point position structure. Based on spatial boundary cutting sequence Each Corresponding fragment The spatial index range is used to read the set of three-dimensional coordinate points it contains, and the three-dimensional coordinates are extracted point by point. And construct the coordinate matrix A sliding window method is used to extract point groups within a local area. For example, the sliding window size is set to 1 meter cubic meter, and the number of points contained in the window is... If statistics are performed, If it is a dense cluster of points, then click on all points in the dense cluster. Calculate the coordinate difference for any two points. and Calculate its lateral displacement separately: ,like and Points are considered as adjacent clusters. By comparing the horizontal and vertical displacements of the current point with those of its neighbors and establishing a connection sequence, a sequence update operation is performed. In practice, points can be renumbered according to the connectivity order and a sequence linked list can be formed, such as a point group. Sequential connection is Repeat the above operations until all dense point groups are clustered. The final output cluster point order structure contains multiple connected point sets, each point set representing a spatially continuous cluster sequence within a local region.

[0024] S103: The coordinate sequences of each point cloud segment in the aggregation point positional structure are called in the order of their acquisition timestamps to perform a time-series connection operation, and the coordinate sequences corresponding to multiple segments are connected to the continuous spatial strip channel in chronological order to obtain a density depiction segment set. For the coordinate sequence of each segment in the aggregation point positional structure, it is determined according to the order of its corresponding acquisition timestamps. To perform a time-series join operation, first extract the timestamp attribute of each point, then sort the point sequence in ascending order based on the time values, and set a threshold for judging time continuity. seconds, if the time difference between two adjacent points is less than or equal to If the time is continuous, then it is determined to be continuous in time. Pairs of points with adjacent times are connected to form continuous time segments, thus constructing a sequence from the fragment. The extracted continuous coordinate sequences are then concatenated in chronological order to form a continuous strip-shaped coordinate set, such as the sequence... , , Sort by time and then access the strip channel: During the operation, it is necessary to maintain the spatial and temporal continuity of the points within each segment, and at the same time record the start and end indices of each segment and its average density parameters: ,in For points, For the volume of space, if If the density is defined as a segment, the final output density segment set consists of multiple segments with spatial density and temporal continuity.

[0025] Please see Figure 3 The steps to obtain S2 are as follows: S201: Based on the adjacent segment data in the density depiction segment set, compare the corresponding aggregation point information in each segment in sequence, retrieve the displacement value difference in the horizontal and vertical directions, and perform item-by-item difference confirmation operation with the preset aggregation position comparison range. Screen out all segments whose differences exceed the spatial offset reference value and obtain the offset segment index column. Based on the density-based segment data of adjacent segments in the segment set, the clustering point information within each segment needs to be extracted sequentially in ascending order of segment number. For each group of adjacent segments, for example... and Extract the three-dimensional coordinates of all points in its aggregation point set and construct a point array. Calculate the lateral displacement difference for each corresponding point pair. Difference with longitudinal displacement Where the lateral displacement difference is the absolute value of the difference in the x-coordinates of adjacent points, and the longitudinal displacement difference is the absolute value of the difference in the y-coordinates, that is: Then, the maximum and average values ​​of all differences in this group of points are calculated and compared one-to-one with the preset boundary parameters of the cluster location control range. For example, if the preset control range is a maximum lateral deviation of no more than 100 mm, the maximum deviation is calculated based on the maximum difference between the two points. The maximum longitudinal deviation shall not exceed Then, for each pair of adjacent points, determine its... or The current segment is recorded as a difference segment. All difference determination results are then statistically summarized. Segments exceeding the spatial offset baseline value are filtered out. The spatial offset baseline value is set according to the scene; for example, it is set to [value missing] for an urban scene. Open space setting When the proportion of point pairs judged to be out of bounds in a certain segment exceeds When this happens, the entire segment is determined to be an offset segment. For example, if a segment contains 1800 point pairs, and 580 of these point pairs exceed the limits for horizontal and vertical displacement, the percentage exceeding the limits is calculated as follows: If the value is greater than 0.3, then the segment number is filtered and entered into the offset segment index column.

[0026] S202: Call the segment number in the offset segment index column, extract the data frame of adjacent number according to the spatial scanning direction, identify all segments with consecutive numbers and located in the same spatial direction range, determine whether the number of consecutive occurrences in the sequence exceeds the regional grouping benchmark value, output the number set for segments that meet the sequence continuity rule, and obtain the segment chain set list. The fragment number in the offset fragment index column is retrieved, and the data frame corresponding to each number is scanned sequentially according to the original numbering. The spatial location index value of each fragment is extracted and compared, for example, the fragment number. and Corresponding spatial index areas: and If both segments are within the same spatial direction, then they are considered as a group of segments. If multiple segments consecutively meet the conditions of consecutive numbering and consistent spatial direction, they are grouped together. The regional grouping baseline value is set to 3, meaning the number of consecutive combinations must satisfy the following: When consecutive number combinations, for example When the quantity is 3, the output is a chain of fragments; if it is only 3, the output is a chain of fragments. The quantity is 2, which does not meet the condition and is removed. This operation is applied to the entire offset fragment index column, so that each group of combinations must maintain continuity in number and the spatial indexes must be nearly consistent in the main direction. In this way, fragment combinations that meet the continuity rule are obtained, each group of combinations forms a fragment chain, and the output is saved in the form of a numbered list to form a fragment chain set list.

[0027] S203: Based on the numbered data in the fragment chain set list, perform location matching operation on the area corresponding to each chain in the land cover data, and group the fragment chains with continuous positions according to the boundary connection relationship of the land cover, and assign the main control attribute identification information to obtain the main control fusion fragment band. Based on the numbered data in the fragment chain set list, the geographic coordinate area corresponding to each number is retrieved one by one. Land cover data is then retrieved and overlaid with the spatial index range of each fragment in the chain for matching. The matching method involves finding the corresponding land cover patch area in the land cover layer to obtain its boundary information, and then determining whether adjacent fragments in the chain are located within the same land cover boundary area. If the difference in horizontal or vertical coordinates at the boundary connection between adjacent fragments is less than [a certain value], the matching is considered successful. If the boundaries are continuous, they are grouped into a set of chains. Chains with boundary connections are sequentially grouped into a unified segment group. The unified segment group is assigned a master control attribute identifier, and the master control attribute field value is set as follows: This is used to identify the chain belonging to the main control fusion category, and finally generate the main control fusion segment band. The main control fusion segment band is composed of multiple chains with spatial connection, continuous boundaries and originating from the offset segment chain set. Each band has continuous numbering, stable spatial index connection and consistent covered ground feature types.

[0028] Please see Figure 4 The steps to obtain S3 are as follows: S301: Acquire the registration records formed by the mobile surveying equipment during its movement, and arrange all records in chronological order into a trajectory data frame sequence. Extract registration error information from each record in the sequence and perform direction judgment operation on the direction change of the error vector. By comparing the stability of the angle signs between adjacent error vectors, confirm the trajectory segments with consistent directions, and add corresponding spatial position identification codes to these segments to obtain the direction identification trajectory segment column. To acquire registration records generated by mobile mapping equipment during its movement, it is necessary to extract each registration data point in chronological order and construct a trajectory data frame sequence. The 3D position and corresponding error vector information from each registration record should be organized into an array. For each data record, its error vector components should be extracted. For each of the two adjacent error vectors, calculate the direction of the included angle. The basis for determining the change in direction is to compare the sign of the cosine of the angle between the adjacent vectors. If the angle between the two vectors... satisfy If the two vectors are in the same direction, it is considered that the direction of the current trajectory segment is stable. If the directions of multiple consecutive pairs of adjacent error vectors are all positive and consistent, the direction of the current trajectory segment is considered stable. The start index and end index are extracted as trajectory segment numbers for such consecutive segments. Spatial position identification codes are generated based on the coordinates of the midpoint of the trajectory segment. The coding format can be defined as block number + segment number. For example, the 5th segment of the third block is recorded as "03-05". When performing the operation, for example, if the error vectors from R7 to R11 are all in the same direction and the starting point coordinates are (150.0, 130.0), then the block number 1500-1300 to which the point belongs is used as a prefix to finally form the direction identification trajectory segment column.

[0029] S302: Call the spatial position identifier code in the direction identifier trajectory segment column, sequentially retrieve the encoded data of each trajectory segment, extract the continuously numbered trajectory segments according to the spatial arrangement order of the trajectory segments, confirm the set of trajectory segments whose number change range is within the spatial continuous reference value range through the sequence continuity judgment operation, and output the spatial position sequence of these trajectory segments to obtain the continuous trajectory spatial column. The spatial location identifier code in the trajectory segment column is called, and the trajectory segment data is retrieved item by item in numerical order. Each data segment is parsed to obtain its spatial location coordinate block and its position number in the sequence through its number. For trajectory segments with adjacent numbers, such as segments 03-05 and 03-06, their spatial location coordinates are extracted and it is determined whether they belong to a continuous block range in the spatial index coordinate system. The spatial continuity reference value can be set to a position index change of no more than 2 units. The determination method is that if trajectory segments Pi and Pi+1 are in the coordinate index range... and If the number of continuous trajectory segments reaches the minimum continuous sequence standard value, set to 3, then the combination is considered a valid spatial segment sequence. During operation, a sliding window method can be used to check all identified segments. In each window, it is checked whether the continuous trajectory number and the position number are consistent. For example, if the continuous segment number is 03-04, 03-05, 03-06, and the corresponding coordinate positions are (151, 132), (152, 132), and (153, 132) respectively, and it is continuous in the x-direction and does not jump in the y-direction, then it is included in the continuous trajectory spatial column and its spatial position sequence is output and saved in the form of both number and coordinate information.

[0030] S303: Based on the spatial position sequence in the continuous trajectory spatial column, the spatial position data of each segment in the main control fusion segment band are matched one by one. The coordinates of the two under the same spatial index system are compared one by one. The position segment of the continuous trajectory spatial sequence in the main control fusion segment band is extracted by the correspondence relationship, and the output sequence is arranged in spatial order to obtain the stable direction segment sequence. Based on the spatial position sequence in the continuous trajectory spatial column, and the corresponding spatial position coordinates of each segment within the master control fusion segment band, the start and end coordinate values ​​of each trajectory segment in a unified spatial index coordinate system are extracted. Then, a point-by-point comparison operation is performed on this coordinate segment within the master control fusion segment band, comparing the coordinate values ​​of each point in the trajectory segment. The system checks if a segment containing the given point exists within the main control segment. If it does, the segment number is extracted and added to the matching sequence. This process is repeated to form a corresponding list from the trajectory sequence to the fusion segment. When multiple trajectory segments correspond consecutively to the same fusion segment or adjacent fusion segments, their numbers are sorted according to the trajectory direction, and their segment numbers are extracted to form a stable segment sequence. In the example, if the trajectory segment coordinates (150, 130), (151, 130), and (152, 130) fall into the main control fusion segments L22, L23, and L24 respectively, then this segment constitutes a stable directional segment sequence, and the output sequence is L22→L23→L24.

[0031] Please see Figure 5 The steps to obtain S4 are as follows: S401: Based on the stable orientation segment sequence and the master control fusion segment band, call the spatial position index information of each segment in the stable orientation segment, compare its position number with the position number in the master control fusion segment band under the spatial index system, perform screening operation on data frames with the same coordinate number, extract all segments with overlapping spatial positions, and obtain a list of spatially overlapping segments. Based on the stable orientation segment sequence and the master control fusion segment band, it is necessary to extract the spatial location index parameter of each segment in the stable orientation segment. This parameter consists of the two-dimensional coordinate position of the starting point of each segment and the segment number, and is recorded in the form of a structure. Meanwhile, each segment in the master fusion segment also records a corresponding spatial location number. For each corresponding With all corresponding Perform a comparison operation, checking item by item whether there are completely equal indices in the coordinate space. If so, and This involves determining whether the two segments overlap spatially, screening all segments that meet this condition, extracting their complete records, and saving their corresponding numbers. and If the numbers are different but the coordinates are the same, the numbers are recorded in the order of the stable directional segments to construct a list of spatially overlapping segments. During the execution process, if the stable directional sequence contains numbers A103 and A104 with position coordinates of (251,370) and (252,370), and the coordinates of numbers B032 and B033 in the master control fusion segment are (251,370) and (253,370) respectively, then A103 and B032 form a spatially overlapping relationship, while A104 and B033 do not form a matching relationship. Only the overlapping pair formed by A103 and B032 is retained in the screening results. Finally, after all the matching is completed, the list of spatially overlapping segments is obtained.

[0032] S402: Based on the set of numbers in the list of spatially overlapping segments, extract the point cloud depiction content associated with the corresponding segment in the point cloud dataset and the corresponding image depiction content in the image dataset in sequence. Aggregate the two types of depiction content with the same spatial index in numerical order and compress them into an entry structure to obtain the point cloud and image double depiction entry set. Based on the set of numbers in the list of spatially overlapping segments, for each pair of matching numbers, the corresponding point cloud depiction content in the point cloud dataset is first extracted. This content consists of a set of three-dimensional coordinates and can be represented as a set of... The data array simultaneously retrieves image block information with the same number from the image dataset. The image information is represented by a matrix of pixel coordinates and grayscale values. For segments with the same location coordinates in two data sources, such as segment A103, which corresponds to 10,500 point records in the point cloud set and a 2048×1024 resolution tile in the image set, these two types of data need to be merged into a unified entry structure. During aggregation, the data is sorted according to the number order to ensure that the number order is from smallest to largest. The merged entry structure includes point cloud fields, image fields, and spatial number fields. Data compression is performed on the merged structure. During the compression process, the point cloud sampling rate can be set to 0.5, that is, one point is taken for every two points. The image resolution is scaled to 70% of the original tile size, that is, the compressed image size is 1433×716. The corresponding entry uses the number as the primary key and is recorded as an entry item: segment A103 → point cloud array compressed to 5,250 points → image thumbnail 1433×716. Each group of entry structures is stored according to this standard. The above process is executed cyclically to generate point-image double-line entry sets in sequence.

[0033] S403: Call the spatial numbering information in the dot shadow double-line entry set, arrange all entry structures according to the spatial orientation index order, organize the entry sequence according to the numbering change trend, output the entry dataset with continuous spatial numbering features, and obtain the fusion candidate fragment set; The spatial numbering information in the dot-matrix entry set is retrieved. The number of each entry is extracted and sorted according to its spatial orientation index. The sorting is based on the trend of the x and y coordinates implied in the number. To determine the trend, the coordinate information in each number needs to be analyzed. By comparing the coordinate differences between adjacent numbers, if the position corresponding to number B105 is (300, 400) and B106 is (301, 400), then it is determined that they are increasing positively in the x-direction. If consecutive numbers satisfy this condition... and If the sequence is confirmed to be spatially continuous, an ordered entry dataset is constructed based on this continuous numbering sequence. The entry sets that satisfy the above continuous numbering characteristics are combined into the same set of fusion candidate segments. If a certain group of entry numbers is D210, D211, D212, D213, and its coordinate indices are (120, 200), (121, 200), (122, 200), (123, 200) respectively, it is determined to be a set of entry sequences with continuous spatial numbers, and it is output as a set of fusion candidate segments.

[0034] Please see Figure 6 The steps to obtain S5 are as follows: S501: Call the point cloud depiction content in the fusion candidate fragment set, extract the three-dimensional coordinate values ​​of the point cloud line data of all entries in the same spatial position number according to the number order, and stack all the spatial paths of the same line segments in sequence along the direction, compress them to form a linear unit structure, and obtain the compressed spatial depiction line set. The point cloud depiction content from the candidate fusion fragment set is retrieved. For each entry's recorded spatial location number, the 3D coordinate sequence is extracted sequentially from its point cloud line data. The data structure is as follows: The system reads entries in ascending order of spatial number. During the reading process, point cloud entries with the same number are stacked line by line according to the spatial direction. The stacking process involves extracting the first and last points of the line segments of each point under the current number, calculating their projection length in the main direction, and overlapping all line segment projections to form equally spaced segment line groups. Then, the points are aligned according to the linear direction of the point position. Corresponding line segments refer to point sets with the same number and the same direction. For example, two point cloud lines numbered D101 and D102 represent the center line segment and the two side edge line segments of the road, respectively. Their spatial direction is the main x-axis direction. Then, the line segments along the x-direction of adjacent points in each line segment are taken and compressed into a unified linear unit by repeated stacking. This linear unit retains the overlapping parts of all line segments and represents them in the form of a simplified point set. For example, three line segments contain 80, 75, and 70 points respectively. The 45 consecutive overlapping points in the common section are retained as the compressed output. Finally, all line segments with corresponding numbers are compressed and constructed into a compressed spatial drawing line set according to their numbers.

[0035] S502: Based on the path location index in the compressed spatial depiction line set, call the image depiction content under the corresponding number, locate the image area boundary coordinates on both sides of each path, and perform close-fitting processing on each image boundary and spatial path line, outputting a combination item structure composed of image boundaries and spatial lines to obtain a list of fitted image items. Based on the path location index in the compressed spatial depiction line set, the image depiction content is called for the number of each line structure, and the image data block corresponding to the number is extracted. The boundary coordinate region adjacent to the path in the image region is identified. During the extraction process, the range on the left and right sides of each path line is extended by 0.5 meters. After determining the extended area, the boundary pixel coordinate points located in this area are searched in the image data, and the outermost boundary line segment is extracted to form the left and right image boundaries. Then, the image boundaries and spatial path lines are fitted. The fitting operation is to compare the coordinates of the image boundary point array and the path point array. The image boundary points are projected at equal intervals on the left and right sides of each path point. The boundary point in the image with a distance of less than 0.2 meters from the path is selected as the effective fitting point using the minimum distance principle. If the fitting condition is met at 5 consecutive points on the path line segment, the boundary line segment and the path are retained to form a combined structure. The structure content includes the path number, path point sequence, left boundary image point set, and right boundary image point set. Each combined structure is recorded as a fitted image entry. All numbered structures are saved in numerical order to form a fitted image entry column.

[0036] S503: Call the number sequence information in the image item column, extract all combined items according to the continuous spatial direction, construct a continuous direction structure according to the spatial arrangement of the numbers, integrate all combined items to form a continuous sequence, and obtain the mapping data fusion result. The system retrieves the sequence number information from the image item column, extracts all item numbers one by one, and arranges them in ascending order. During the sorting process, the spatial coordinates implicit in the number are used as the primary sorting reference. For example, if the coordinates of number F120 are (150, 360), F121 are (151, 360), and F122 are (152, 360), then the primary sorting direction is the positive x-axis. The system checks whether the number arrangement is continuous. If the difference between the numbers is 1 and there is no jump in the corresponding spatial coordinates, then the coordinate change satisfies the condition. and If the path is continuous, it is determined to be a continuous spatial orientation. All combination items that meet this feature are extracted and arranged into a structured sequence in order. Each item in the sequence contains a path number, corresponding image fitting information and spatial coordinate position. Finally, all combination items corresponding to consecutive numbers are aggregated into a complete spatial structure. The integrated output of this continuous orientation structure is the mapping data fusion result.

[0037] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A cognitive computing-based adaptive fusion method for surveying and mapping data, characterized in that, Includes the following steps: S1: Acquire point cloud records generated during mobile mapping platform operations, divide them into continuous segments according to fixed spatial boundaries, extract the point aggregation positions within each segment and rearrange them in row and column order to form segment depiction results, and concatenate all segment depiction results in the acquisition order to generate a density depiction segment set. S2: Based on the density depiction fragment set, compare the aggregation position differences of adjacent fragments, select fragment groups whose differences exceed the control range, organize them into continuous regions according to the connection relationship in the land cover, mark the continuous regions with the dominant fusion mark, and generate the master control fusion fragment band. S3: Obtain the registration record of the surveying equipment and connect the trajectories in sequence according to time. Extract the trajectory segments with consistent continuous error changes and superimpose the corresponding spatial position markers. Integrate the spatial position of the trajectory segments with the spatial position of the main control fusion segment to generate a stable direction segment sequence. S4: Based on the stable directional segment sequence and the main control fusion segment band, extract the trajectory segments with overlapping spatial positions, summarize the corresponding point cloud and image depiction content one by one, combine the depiction content at the same position into fusion basic entries, and generate a fusion candidate segment set.

2. The adaptive fusion method for mapping data based on cognitive computing according to claim 1, characterized in that: The density depiction fragment set includes the distribution of aggregation point coordinates, fragment image sequence, and band structure index. The master fusion fragment band includes region continuity identifier, fusion priority marker, and spatial matching sequence. The stable orientation segment sequence includes registration direction information, continuous trajectory number, and spatial correspondence. The fusion candidate fragment set includes fusion entry index, point cloud and image pairing results, and spatial overlapping fragment list.

3. The adaptive fusion method for mapping data based on cognitive computing according to claim 1, characterized in that, The steps for obtaining S1 are as follows: S101: Acquire point cloud records formed by the mobile mapping platform during field operations, and locate the three-dimensional coordinate data contained in each frame of point cloud according to the LiDAR scanning order. Based on the preset fixed spatial boundary configuration parameter set, clip the spatial position range of the point cloud data frame, split it into multiple point cloud segments according to the continuous scanning order, and number the spatial range index data of each segment to obtain the spatial boundary clipping sequence. S102: Based on the spatial index range of each point cloud segment in the spatial boundary clipping sequence, extract the position information of each three-dimensional coordinate point in the segment, identify the dense point group located in the local area, perform a sequence update operation by comparing the horizontal and vertical displacement relationship of each set of coordinates, and output the cluster point position order structure. S103: The coordinate sequences of each point cloud segment in the aggregation point positional structure are called in the order of their acquisition timestamps to perform a time-series connection operation, and the coordinate sequences corresponding to multiple segments are connected to the continuous spatial strip channel in chronological order to obtain a density depiction segment set.

4. The adaptive fusion method for surveying and mapping data based on cognitive computing according to claim 1, characterized in that, The steps for obtaining S2 are as follows: S201: Based on the adjacent segment data in the density depiction segment set, compare the corresponding aggregation point information in each segment in sequence, retrieve the displacement value difference in the horizontal and vertical directions, and perform item-by-item difference confirmation operation with the preset aggregation position comparison range. Screen out all segments whose differences exceed the spatial offset reference value and obtain the offset segment index column. S202: Call the segment number in the offset segment index column, extract the data frame of adjacent number according to the spatial scanning direction, identify all segments with consecutive numbers and located in the same spatial direction range, determine whether the number of consecutive occurrences in the sequence exceeds the regional grouping reference value, output the number set for segments that meet the sequence continuity rule, and obtain the segment chain set list. S203: Based on the numbered data in the fragment chain set list, perform a location matching operation on the area corresponding to each chain in the land cover data, and group the fragment chains with continuous positions according to the boundary connection relationship of the land features, and assign them master control attribute identification information to obtain the master control fusion fragment band.

5. The adaptive fusion method for mapping data based on cognitive computing according to claim 1, characterized in that, The steps for obtaining S3 are as follows: S301: Acquire the registration records formed by the mobile surveying equipment during its movement, and arrange all records in chronological order into a trajectory data frame sequence. Extract registration error information from each record in the sequence and perform direction judgment operation on the direction change of the error vector. By comparing the stability of the angle signs between adjacent error vectors, confirm the trajectory segments with consistent directions, and add corresponding spatial position identification codes to these segments to obtain the direction identification trajectory segment column. S302: Call the spatial position identifier code in the direction identifier trajectory segment column, sequentially retrieve the encoded data of each trajectory segment, extract the continuously numbered trajectory segments according to the spatial arrangement order of the trajectory segments, confirm the set of trajectory segments whose number change range is within the spatial continuous reference value range through the sequence continuity judgment operation, and output the spatial position sequence of these trajectory segments to obtain the continuous trajectory spatial column. S303: Based on the spatial position sequence in the continuous trajectory space column, the spatial position data of each segment in the main control fusion segment band are matched one-to-one with the coordinates of the two segments under the same spatial index system. The position segments of the continuous trajectory space sequence in the main control fusion segment band are extracted from the correspondence relationship, and the output sequence is arranged in spatial order to obtain the stable direction segment sequence.

6. The adaptive fusion method for surveying and mapping data based on cognitive computing according to claim 1, characterized in that, The steps for obtaining S4 are as follows: S401: Based on the stable orientation segment sequence and the main control fusion segment band, call the spatial position index information of each segment in the stable orientation segment, compare its position number with the position number in the main control fusion segment band under the spatial index system, perform a screening operation on the data frames with the same coordinate number, extract all segments with overlapping spatial positions, and obtain a list of spatially overlapping segments. S402: Based on the set of numbers in the list of spatially overlapping segments, extract the point cloud depiction content associated with the corresponding segment in the point cloud dataset and the corresponding image depiction content in the image dataset in sequence. Aggregate the two types of depiction content with the same spatial index in numerical order and compress them into an entry structure to obtain a point-image double-lined entry set. S403: Call the spatial numbering information in the dot shadow double-line entry set, arrange all entry structures according to the spatial orientation index order, organize the entry sequence according to the numbering change trend, output the entry dataset with continuous spatial numbering characteristics, and obtain the fusion candidate fragment set.

7. The adaptive fusion method for mapping data based on cognitive computing according to claim 1, characterized in that, The method further includes: S5: Call the point cloud and image depiction content in the fusion candidate segment set, superimpose and compress multiple point cloud depictions into a single spatial depiction line at the same spatial position, attach the image depiction to the spatial depiction line at the corresponding position, rearrange all fusion segments in the direction order, and generate the mapping data fusion result. The mapping data fusion results include compressed point cloud line structure, image texture layer, and unified orientation fusion segments.

8. The adaptive fusion method for surveying and mapping data based on cognitive computing according to claim 1, characterized in that, The steps for obtaining S5 are as follows: S501: Call the point cloud depiction content in the fusion candidate fragment set, extract the three-dimensional coordinate values ​​of the point cloud line data of all entries in the same spatial position number according to the number order, and sequentially stack all the spatial paths of the same line segments along the direction to compress and form a linear unit structure to obtain the compressed spatial depiction line set. S502: Based on the path position index in the compressed spatial depiction line set, call the image depiction content under the corresponding number, locate the image area boundary coordinates on both sides of each path, and perform close-fitting processing on each image boundary and spatial path line, outputting a combination item structure composed of image boundaries and spatial lines to obtain a series of fitted image items. S503: Call the number sequence information in the image item column, extract all combined items according to the continuous spatial direction, construct a continuous direction structure according to the spatial arrangement of the numbers, integrate all combined items to form a continuous sequence, and obtain the mapping data fusion result.

Images