Dual-uav cooperative terahertz imaging detection system and control method

Abstract

The present application relates to the technical field of terahertz, specifically to a dual unmanned aerial vehicle cooperative terahertz imaging detection system and control method, the dual unmanned aerial vehicles fly around the component, the transmitting end continuously projects the terahertz beam to build a three-dimensional energy flow diagram, the receiving end samples and point illuminates the corresponding intersection area, extracts the energy flow mutation edge along the trajectory, removes the angle mutation to retain the same path component boundary, circumscribes the stable irradiation area, and embeds the flight control output imaging instruction according to the point position change. The present application sequentially solidifies the irradiation point and the flight position along the trajectory, constructs a three-dimensional continuous energy flow relationship, extracts the mutation boundary according to the adjacent difference, and screens the stable trend according to the consistency of the irradiation direction; under the path constraint, the consistent irradiation area is circumscribed, the point position change in the area is mapped as the cooperative control of attitude and displacement, the irradiation direction is adjusted synchronously with the structure boundary, the profile recognition, irradiation stability and task consistency are improved, and the complex surface fluctuation interference is weakened.

Description

Technical Field

[0001] This invention relates to the field of terahertz technology, and in particular to a dual-UAV collaborative terahertz imaging detection system and control method. Background Technology

[0002] The field of terahertz technology encompasses the generation, transmission, control, and detection of terahertz waves. This includes generating terahertz pulses using optoelectronic devices, transmitting terahertz beams via free-space paths, acquiring the absorption and scattering characteristics of the target object to the terahertz waves through reflection or transmission, and recording signals to form images using imaging detection devices. In imaging detection, this field typically involves outputting a beam from a terahertz source, setting a fixed detection distance and scanning path, and recording the received signals to form a spatial array, constituting the basic imaging process. Among these, the traditional dual-UAV collaborative terahertz imaging detection system and control method utilizes two UAVs, one carrying a terahertz transmitting unit and the other a terahertz receiving unit, and completes regional imaging detection by controlling the UAVs' flight trajectories. Traditional methods typically involve a transmitting drone outputting a terahertz beam and a receiving drone recording the signal reflected or transmitted by the object under test. The positional relationship between the two drones is determined through ranging calibration or time synchronization to maintain a fixed observation angle and baseline. Flight control involves setting waypoint sequences and flight altitudes to allow the drones to cover the area under test along the path. Data acquisition involves sequentially triggering the terahertz source and synchronously recording the received waveforms, thereby forming a spatial signal distribution based on the time difference between transmission and reception. Image generation involves establishing spatial coordinates using a scanning trajectory and collecting point array data to form the imaging process.

[0003] Existing technologies rely on preset waypoints and fixed observation relationships to organize the point array. The correlation between spatial point locations and flight attitude is weak. Airflow and attitude fluctuations can easily cause coordinate drift. Reflection changes in structural intersections or occlusion areas are amplified into pseudo-features by uniform scanning. Ranging or time synchronization is not stable enough in multipath environments. Baseline offset causes spatial distribution distortion. The acquisition process lacks endogenous discrimination of the continuity of energy flow abrupt changes and directional consistency. Boundary and non-boundary data are mixed, often requiring supplementary flights or post-processing corrections, resulting in decreased consistency of imaging results and extended mission cycles. Summary of the Invention

[0004] To address the technical problems existing in the prior art, embodiments of the present invention provide a dual-UAV cooperative terahertz imaging detection system and control method. The technical solution is as follows: A dual-UAV collaborative terahertz imaging detection system and control method, comprising the following steps: S1: Obtain the flight status of the two UAVs around the structural components. The transmitter projects terahertz beams onto the surface of the structure, and the receiver flies synchronously close to the intersection area. It samples the spatial points formed by the illumination, binds the sampling action to the flight position, and stores the corresponding spatial points in sequence. The points correspond to the illumination data, and outputs a three-dimensional energy flow illumination layer. S2: Based on the relative energy flow relationship between the irradiated points in the three-dimensional energy flow irradiation layer, extract the energy flow change of continuous points along the flight trajectory direction, observe whether a continuous structure is formed in the energy flow abrupt change region, focus on processing the linear segments formed by continuous abrupt changes, and extract the edge path band that constitutes the energy flow jump. S3: Based on the change of the irradiation direction of the line segment in the energy flow jump edge path zone, the direction difference of the continuous irradiation points is compared, and the irradiation trajectory is continued to be tracked in the area with consistent path direction. The segments with abrupt angle changes are eliminated, and the path segments with consistent direction are retained and gathered into a group of intersection and boundary lines with unified direction. S4: Call the irradiation path segments in the unified intersection boundary line group, track the path extension distance during its spatial extension, process the connection status between paths, and in the remaining irradiation segment space after stripping the discontinuous boundary area, delineate the irradiation area with the same extension direction and extract it as a terahertz irradiation stable block. S5: Call up all the illumination point information in the terahertz illumination stable block, compare the position change status of the illumination points in the flight mission, adjust the flight control action according to the position change, embed the illumination pointing and displacement switching control action in the flight control template, and output the terahertz imaging control command.

[0005] As a further aspect of the present invention, the three-dimensional energy flow illumination layer includes a set of spatial point coordinates, energy flow distribution characteristics, illumination coverage density identifiers, and trajectory mapping indexes. The energy flow jump edge path zone specifically includes energy flow change amplitude characteristics, continuous edge segments, path direction characteristics, and jump aggregation area identifiers. The directionally unified intersection boundary line group includes direction consistency characteristics, effective line segment set, intersection boundary description, and direction stability identifiers. The terahertz illumination stable block specifically includes stable coverage range, regional consistency characteristics, spatial connectivity area identifiers, and illumination uniformity parameters. The terahertz imaging control commands include flight attitude control parameters, illumination pointing configuration parameters, displacement switching control parameters, and imaging task constraint parameters.

[0006] As a further aspect of the present invention, the step of obtaining the three-dimensional energy flow illumination layer is as follows: S101: Obtain the flight path information of the two UAVs around the structural component. The transmitting UAV continuously emits terahertz beams to the surface of the structure according to the path. The receiving UAV follows the path at the boundary of the irradiation area according to the synchronization command and records the spatial position in the intersection area to generate a set of irradiation trajectory points. S102: Based on the set of illumination trajectory points, call the receiving device of the receiving drone to obtain the reflected signal intensity, frequency and arrival time of each point, bind the corresponding illumination parameters with the flight position, and obtain the point illumination signal parameter group after sorting them in time order. S103: Based on the point illumination signal parameter group, extract the spatial coordinates and illumination parameter information, generate the spatial distribution relationship of the points in the structural surface model, construct the mapping matrix between signal intensity and coordinates, and obtain the three-dimensional energy flow illumination layer.

[0007] As a further aspect of the present invention, the step of obtaining the energy flux transition edge path band is as follows: S201: Based on the relative energy flow relationship between each pair of irradiation points in the three-dimensional energy flow irradiation layer, call the direction vector, intensity and spatial coordinates of the irradiation points, extract continuous points along the flight trajectory direction, calculate the energy flow change between adjacent points, and obtain the energy flow change sequence of the irradiation points. S202: Based on the energy flow change sequence of the irradiation point, screen the mutation location and extract continuous points. Combine their spatial distribution direction and point clustering, identify the related continuous regions and obtain the spatial mutation extension fragment. S203: Call the points in the spatial mutation extension segment, analyze the continuity characteristics between spatial direction vectors, screen the continuous point series with similar directions, connect them sequentially to form a linear region, and obtain the energy flow jump edge path zone.

[0008] As a further aspect of the present invention, the step of obtaining the direction-unified intersection boundary line group is as follows: S301: Based on the set of all line segments extracted from the energy flow jump edge path zone, call the start and end point position information of each line segment and its order relationship in the sequence, calculate the change value of the illumination direction between adjacent line segments, filter the continuous line segment segments with consistent illumination direction, and generate a set of line segment segments with consistent direction. S302: Based on each continuous segment in the set of line segments with consistent direction, call the direction change value of adjacent line segments, determine the degree of deviation from the direction of the segment, remove the line segments with abrupt changes in direction, and retain the path line segments with stable direction to obtain a group of path direction stable segments. S303: Call all line segments in the path direction stable segment group, merge them into path groups in the same direction according to the direction value, identify the spatial intersection points in each path group, integrate the path boundaries in the intersection structure, and obtain a direction-unified intersection boundary line group.

[0009] As a further aspect of the present invention, the step of obtaining the terahertz irradiation stable block is as follows: S401: Call all the illumination path segments in the unified direction intersection boundary line group, extract the extension trajectory and relative coordinate set of the path in space, and judge the extension direction classification between the path segments according to the starting direction parameter and direction consistency standard of the illumination path, and obtain the direction difference range of the illumination path. S402: Based on the directional difference range of the irradiation path, track the relative displacement of the irradiation path in the spatial extension, identify the path unit with interrupted directional continuity according to the coordinate set of the midpoint of the adjacent path and the spatial distance relationship, and call its corresponding coordinate index to perform the stripping operation to obtain the coordinate set of the boundary discontinuous region. S403: Based on the set of coordinates of the discontinuous boundary region, filter the path units with the same extension direction from the remaining path set, perform sequence matching based on the direction vector group and the path number, extract the spatial index set within the closed path region, and obtain the terahertz irradiation stable block.

[0010] As a further aspect of the present invention, the step of obtaining the terahertz imaging control command is as follows: S501: Call up all the irradiation point information involved in the terahertz irradiation stability block, extract the position coordinates of each point under different time sequences based on the flight mission time axis, and compare the position changes of each point during the mission process according to the horizontal and vertical displacement directions to obtain the irradiation point position offset sequence. S502: Based on the irradiation point position offset sequence, filter out the point group that frequently shifts in the task time sequence, call the offset direction corresponding to its continuous position change, identify the sudden change stage in combination with the time interval, and classify and analyze each stage to obtain offset dynamic prediction reference data. S503: Call the point number and illumination direction information involved in the offset dynamic prediction reference data, adjust the pointing parameters of frequently changing points and the displacement switching content in the flight control template, make corresponding modifications to the illumination timing and displacement triggering nodes, and obtain terahertz imaging control commands.

[0011] The beneficial effects of the technical solutions provided in the embodiments of the present invention include at least the following: In this invention, by solidifying the irradiation point and flight position along the trajectory sequence, the energy flow forms a continuous and comparable relationship in three-dimensional space. The abrupt boundary is extracted based on the energy flow difference between adjacent points, and a stable trajectory is selected by combining the consistency of the irradiation direction. Under the constraint of path space extension, the irradiation area with consistent direction is delineated. Then, the changes in the point position within the area are transformed into attitude and displacement coordinated control, so that the irradiation direction and flight trajectory are adjusted synchronously with the changes in the structural boundary. This enhances the ability to identify the contour of the intersection area, the controllability of the stable irradiation range, and the consistency of results from multiple missions, and reduces the interference of complex surface undulations on the imaging process. Attached Figure Description

[0012] 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 three-dimensional energy flow illumination layer according to the present invention. Figure 3This is a flowchart illustrating the process of obtaining the edge path band of the energy flow transition in this invention. Figure 4 This is a flowchart illustrating the process of obtaining the directional unified intersection boundary line group of the present invention. Figure 5 This is a flowchart illustrating the process of obtaining the terahertz irradiation stable block in this invention. Figure 6 This is a flowchart illustrating the acquisition of terahertz imaging control commands according to the present invention. Detailed Implementation

[0013] The present invention will now be described in further detail with reference to the accompanying drawings. These drawings are simplified schematic diagrams, illustrating only the basic structure of the invention, and therefore only show the components relevant to the invention.

[0014] refer to Figures 1 to 6 A dual-UAV collaborative terahertz imaging detection system and control method, comprising the following steps: S1: During the flight of the two drones around the structural component, the transmitting drone continuously projects terahertz beams onto the structural surface, and the receiving drone flies close to the intersection area and samples the spatial points generated by each irradiation. The sampling action is bound to the current flight position, and the corresponding spatial points are stored sequentially according to the flight trajectory. The points are then output as a three-dimensional energy flow irradiation layer, corresponding to the irradiation data. S2: Based on the relative energy flow relationship between each pair of irradiation points in the three-dimensional energy flow irradiation layer, extract the energy flow change process of all continuous points along the flight trajectory direction, stop in the energy flow abrupt change area and observe whether a continuous structure is formed, perform centralized processing on the linear segments formed by continuous abrupt changes, and extract the energy flow jump edge path band based on this. S3: Extract the irradiation direction change of each line segment based on the edge path zone of energy flow jump, compare the direction difference of continuous irradiation points, continue to track the irradiation trajectory in the path with consistent direction, remove segments with sudden angle changes during the process, retain all path line segments with constant direction, and gather them into a unified direction intersection boundary line group. S4: Call all irradiation path segments in the unified direction intersection boundary line group, track the extension distance between paths during the spatial extension of the irradiation path, process the connection shape between path lines, peel off the discontinuous boundary area from the layer, delineate the irradiation area with consistent extension direction in the collection space of the remaining irradiation segments, and extract it as a terahertz irradiation stable block. S5: Call up all the illumination point information involved in the terahertz illumination stability block, compare the position change status of the illumination point in the flight mission, arrange the flight control adjustment actions according to these position changes, directly embed the control actions of illumination pointing and displacement switching in the flight control template, and output terahertz imaging control commands.

[0015] The three-dimensional energy flow illumination layer includes a set of spatial point coordinates, energy flow distribution characteristics, illumination coverage density identifiers, and trajectory mapping indexes. The energy flow jump edge path zone specifically includes energy flow change amplitude characteristics, continuous edge segments, path direction characteristics, and jump aggregation area identifiers. The direction-unified intersection boundary line group includes direction consistency characteristics, effective line segment set, intersection boundary description, and direction stability identifiers. The terahertz illumination stable block specifically includes stable coverage range, regional consistency characteristics, spatial connectivity area identifiers, and illumination uniformity parameters. The terahertz imaging control commands include flight attitude control parameters, illumination pointing configuration parameters, displacement switching control parameters, and imaging task constraint parameters.

[0016] Please see Figure 2 The steps for obtaining the 3D energy flow illumination layer are as follows: S101: Obtain the flight path information of the two UAVs around the structural component. The transmitting UAV continuously emits terahertz beams to the surface of the structure according to the path. The receiving UAV follows the path at the boundary of the irradiation area according to the synchronization command and records the spatial position in the intersection area to generate a set of irradiation trajectory points. Starting with acquiring flight path information of two UAVs around a structural component, the outer contour of the component to be inspected is first disassembled. During this process, the length, width, and height dimensions of the structural component in the design drawings or modeling files are read. The structural surface is then divided into segments at preset intervals. For example, for a beam surface with a length of 30 meters, a flight control node is divided every 0.5 meters, forming a continuous node sequence. This node sequence is then converted into a set of spatial coordinates recognizable by the UAVs and loaded into the flight control unit of the transmitting UAV. During flight, the transmitting UAV sequentially calls each coordinate node as the current flight target point, and the flight direction is determined by comparing the difference between the current real-time positioning coordinates and the target coordinates. Upon reaching the target point, the terahertz beamline emission is triggered. During the emission, a stable beamline is continuously output and a fixed attitude angle is maintained. Simultaneously, the receiving UAV receives synchronization command information from the transmitter. The synchronization command includes the current illumination node number and time identifier. The receiving UAV calls the pre-stored boundary following trajectory according to the number and adjusts its own flight path to always be located at the boundary position of the emission beamline coverage area. During the following flight, it continuously reads its own real-time spatial coordinates and records the spatial position at multiple consecutive moments in the beamline intersection area. For example, it records several sets of coordinate data within one illumination node dwell period. Finally, all the recorded spatial coordinates are summarized according to the node number and time order to form a complete illumination trajectory point set.

[0017] S102: Based on the set of illumination trajectory points, call the receiving device of the receiving drone to obtain the reflected signal strength, frequency and arrival time of each point, bind the corresponding illumination parameters with the flight position, and obtain the point illumination signal parameter group after sorting them in time order. Using the set of illumination trajectory points as input, the receiving device of the receiving UAV is first invoked sequentially according to the order of the points in the set to perform signal acquisition. During the process, the receiving device reads the intensity value, main frequency information, and arrival time information of the reflected signal within the time window corresponding to each point. The intensity value is directly output by the power detection unit inside the receiving device, the frequency information is extracted by the spectrum detection unit from the main peak value within the current signal range, and the arrival time is determined by the difference between the receiving time marker and the transmitting time marker. Subsequently, the acquired signal parameters are compared with the spatial coordinate information corresponding to the point. The binding process involves using the point number as an index to store the spatial coordinates, intensity values, frequency values, and arrival time into the same data recording unit, thus completing the single-point data encapsulation. This acquisition and binding operation is repeated for all points to form a continuous data recording sequence. Subsequently, this sequence is sorted according to time stamps, arranged solely based on chronological order without introducing other filtering conditions. During the sorting process, the time difference between adjacent data records is compared. When the time difference exceeds the preset continuous acquisition interval, the data is divided into different time periods, ultimately resulting in a set of point illumination signal parameters arranged in chronological order.

[0018] S103: Based on the point illumination signal parameter set, extract the spatial coordinates and illumination parameter information, generate the spatial distribution relationship of the points in the structural surface model, construct the mapping matrix between signal intensity and coordinates, and obtain the three-dimensional energy flow illumination layer.

[0019] Based on the aforementioned point illumination signal parameter set as input, the spatial coordinate information is first read one by one, and the read coordinates are matched with the surface model of the structural component. During the matching process, the validity of the coordinates is determined by whether they fall within the spatial range defined by the structural surface. Only valid coordinates are processed further. Then, the signal intensity information of the point is extracted from the corresponding data record, and the point is assigned to the corresponding surface area unit according to the structural surface division rules. Each area unit corresponds to a fixed area range on the structural surface. When there are multiple points in the same area unit, the intensity values ​​of these points are read sequentially and processed centrally, and used as the representative intensity data source of the area unit. Area units without point records are directly marked as null. After completing the traversal of the entire structural surface, all area units are arranged according to their distribution relationship in spatial coordinates to form a two-dimensional distribution layer with spatial position as index and signal intensity as associated data. Finally, multiple two-dimensional distribution layers are sequentially superimposed according to the height direction or surface hierarchy relationship to generate a three-dimensional energy flow illumination layer consistent with the spatial morphology of the structural surface.

[0020] Please see Figure 3 The steps for obtaining the edge path band of the energy flow transition are as follows: S201: Based on the relative energy flow relationship between each pair of irradiation points in the three-dimensional energy flow irradiation layer, the direction vector, intensity and spatial coordinates of the irradiation points are called to extract continuous points along the flight trajectory direction, and the energy flow change between adjacent points is calculated to obtain the energy flow change sequence of the irradiation points. Based on the relative energy flow relationships between illumination points in a 3D energy flow illumination layer, the layer is first read layer by layer. Irradiation points arranged according to the flight trajectory are extracted from each layer. During execution, for a single flight trajectory, the spatial coordinates, direction vector, and energy flow intensity of each illumination point are sequentially retrieved. The direction vector is determined by the difference between the coordinates of the point and its adjacent points on the trajectory. Specifically, the direction of coordinate change between the current point and the previous point is used as the flight direction identifier. Then, adjacent point pairs are continuously selected along the direction sequence on the same trajectory. Energy flow change judgment is performed on each pair of points, and the energy flow intensity values ​​of the two points are directly read during the judgment process. The values ​​are then used to perform numerical difference calculations. For example, if the energy flux intensity of the previous irradiation point is recorded as 2.8 units and that of the next irradiation point is recorded as 3.6 units, the difference between the two is recorded as 0.8 units, which is used as the energy flux change for that pair of adjacent points. This numerical reading and difference recording operation is repeated for all adjacent points in the trajectory to form a sequence of energy flux change records arranged in the order of the trajectory. No smoothing or fitting operations are introduced during the execution process; only the original change results are retained. At the same time, each change record is associated with its corresponding spatial coordinate position. For example, the change value of 0.8 is bound to the coordinate position of the next irradiation point. Finally, within a complete flight trajectory, an irradiation point energy flux change sequence consisting of multiple continuous change values ​​is obtained.

[0021] S202: Based on the energy flow change sequence of the irradiation point, screen the mutation location and extract continuous points. Combine their spatial distribution direction and point clustering to identify related continuous regions and obtain spatial mutation extension fragments. Based on the aforementioned energy flux change sequence at the irradiation point as input, each energy flux change record is first read sequentially according to the sequence. An interval judgment is then performed on the change values. Records with a change amplitude less than 0.3 units are marked as a gradual change interval, records with a change amplitude between 0.3 and 0.7 units are marked as a transitional change interval, and records with a change amplitude greater than 0.7 units are marked as an abrupt change interval. Subsequent processing is then performed only on records marked as abrupt change intervals. During this process, the coordinates of the irradiation point corresponding to the change record are extracted, and the interval is extended forward and backward by one point each. The change marking status of adjacent points obtained from the extension is read again. When adjacent points are also in abrupt change intervals or transitional change intervals, these points are... The data is then incorporated into the same continuous group. For example, if three consecutive change values ​​on a trajectory are 0.8, 0.9, and 0.6 units respectively, the corresponding three points are recorded as a continuous group. After extracting a single group, the process continues to move forward in the sequence until all change records have been traversed. Based on this, a spatial distribution direction judgment is performed on each continuous group. During the judgment process, the spatial coordinates of the first and last points in the group are read, and the main extension direction of the group is determined by comparing the magnitude of their changes in the length or height direction. At the same time, the number of points in the group is counted. When the number of points is not less than 3 and the spatial coordinate changes show a continuous distribution in a single direction, the continuous group is identified as a continuous region with a correlation, and the whole is recorded as a spatially abrupt extension segment.

[0022] S203: Call the points in the spatial mutation extension segment, analyze the continuity characteristics between spatial direction vectors, screen the continuous point series with similar directions, connect them sequentially to form a linear region, and obtain the energy flow jump edge path zone.

[0023] Starting with spatially abruptly extended segments as input, the system first reads the set of points contained in each segment and reorders them according to the original flight sequence. During execution, for each sorted point sequence, the direction vector values ​​corresponding to adjacent points are called one by one. The differences between the direction vectors are judged, using the component changes of the direction vector along each coordinate axis as the comparison criterion. When the component changes of two adjacent points are both less than the numerical range corresponding to a preset angle range, they are judged to be close in direction. For example, the component changes in the length direction are less than 0.1, the component changes in the width direction are less than 0.1, and the component changes in the height direction are less than 0.0. In case 5, the two points are determined to be in the same direction. Then, adjacent points that meet the condition are sequentially connected to form a continuous point sequence. When a point with a change in direction exceeding the above interval is encountered, the current point sequence ends and a new point sequence is started. The direction judgment and connection operation is repeated for all points within the same spatial abrupt extension segment. Finally, several linear point sequences are obtained by sequentially connecting points with similar directions. Each linear point sequence is then recorded as an independent unit, maintaining its original spatial arrangement. After all segments are processed, all obtained linear point sequences are summarized and output to form the corresponding energy flow jump edge path zone.

[0024] Please see Figure 4 The steps to obtain the unified direction intersection boundary line group are as follows: S301: Based on the set of all line segments extracted from the edge path zone of energy flow jump, call the start and end point position information of each line segment and its order relationship in the sequence, calculate the change value of the illumination direction between adjacent line segments, filter the continuous line segment segments with consistent illumination direction, and generate a set of line segment segments with consistent direction. Based on the extracted line segment set in the edge path zone of the energy flux transition, the process first involves sequentially decomposing all line segments within the path zone. During this process, the spatial coordinates of the starting and ending points of each line segment are read one by one, while simultaneously reading the segment's position in the original path sequence. The illumination direction value corresponding to the line segment is determined by the difference between the starting and ending point coordinates. This direction value is recorded in the form of three-axis direction components. For example, if a line segment has starting coordinates (12.0, 1.5, 0.8) and ending coordinates (12.6, 1.5, 0.8), then the length component of this line segment is 0.6, and the other direction components are 0. Next, adjacent line segments are grouped according to their numbering order for processing. The direction component data of each group of adjacent line segments is read, and the direction change value is calculated. During the calculation, the values ​​of each direction component are compared between the preceding and following line segments. The difference is used to obtain the change value. For example, if the length component of the previous line segment is 0.6 and the next line segment is 0.58, the change value is recorded as 0.02. At the same time, the same value comparison is performed on the width and height directions. After completing a set of direction change records, it is bound and stored with the number of the next line segment. This reading and calculation action is repeated for all adjacent line segments in the path band to form a continuous sequence of direction change records. On this basis, a consistency filtering operation is performed. During the filtering, line segments with a direction change value less than 0.05 are marked as direction consistent intervals, line segments with a change value between 0.05 and 0.15 are marked as gradual change intervals, and line segments with a change value greater than 0.15 are marked as abrupt change intervals. Only line segments corresponding to the direction consistent intervals and gradual change intervals are continuously spliced. When multiple consecutive line segments fall into the above two types of intervals, they are merged and recorded as a continuous line segment fragment, finally forming a set of direction consistent line segment segments composed of multiple sets of direction consistent line segments.

[0025] S302: Based on each continuous segment in the set of line segments with consistent direction, call the direction change value of adjacent line segments, determine the degree of deviation from the direction of the segment, remove the line segments with abrupt changes in direction, and retain the path line segments with stable direction to obtain the path direction stable segment group. Based on the input set of line segments with consistent direction, the process first reads each continuous line segment fragment in the set one by one. The line segments within each fragment are numbered and organized according to their original order. During execution, the direction value of the first line segment of the fragment is used as the fragment direction reference value. This reference value is directly determined by reading the three-axis direction components of the first line segment. Then, starting from the second line segment, the direction components are read one by one, and an offset judgment is performed between them and the fragment direction reference value. During the judgment, the change magnitude of the current line segment's components in the length, width, and height directions is calculated. When all three change magnitudes are less than 0.1, the line segment direction is considered consistent with the fragment direction. When any change magnitude is between 0.1 and 0.25, it is considered a slight offset. When any change magnitude is greater than 0.2... At time 5, a direction change is determined. For example, if the reference direction length component of a segment is 0.6 and the current segment length component is 0.85, the change range is 0.25, corresponding to the change interval. During the execution process, the segment determined to be a direction change is directly removed from the current segment, while maintaining the original order of the segments before and after it. For segments determined to be consistent or slightly offset, they are retained, and the continuity record within the segment is updated synchronously during the retention process. After all segments within a single segment are judged, if the number of remaining segments is not less than 2, they are recorded as a whole as a valid path segment. The above reference reading, offset judgment and removal operations are repeated for all continuous line segment segments in the segment set. Finally, a path direction stable segment group composed of multiple groups of direction stable line segments is obtained.

[0026] S303: Call all line segments in the path direction stable segment group, merge them into path groups in the same direction according to the direction value, identify the spatial intersection points in each path group, integrate the path boundaries in the intersection structure, and obtain the direction unified intersection boundary line group.

[0027] Based on the input of stable path direction segment groups, all line segments within a segment group are first read centrally, and their direction component data are extracted one by one. During the process, merging is performed according to the dominant direction of the direction component. When a line segment has a length direction component greater than 0.5 and all other direction components are less than 0.2, it is classified into the length-oriented path group; when a line segment has a height direction component greater than 0.4 and all other direction components are less than 0.2, it is classified into the height-oriented path group. After completing the direction classification of all line segments, spatial intersection judgment is performed for each same-direction path group. During the judgment process, the start and end coordinates of the line segments within the group are read one by one. The coordinates of any two line segments are compared. When the end coordinates of one line segment and the start coordinates of another line segment are within the same direction, the line segment is classified into the length-oriented path group. When the difference in all three coordinate directions is less than 0.05, the location is recorded as a spatial intersection point. For example, if two line segments meet the above conditions at coordinates (15.0, 1.5, 0.8), then the point is recorded as an intersection node. After the intersection point is extracted, the set of line segments associated with each intersection node is traced outward, the start and end coordinate ranges of these line segments are read, and their boundary positions are marked in space. During the marking process, the coordinate value corresponding to the outermost line segment is recorded as the path boundary. After the boundary integration operation is completed for all line segments in the same intersection structure, the intersection structure as a whole is recorded as a boundary line with a unified direction. Finally, the intersection identification and boundary integration are repeated for all path groups in the same direction, and a unified direction intersection boundary line group is formed.

[0028] Please see Figure 5 The steps for obtaining stable blocks under terahertz irradiation are as follows: S401: Call all illumination path segments in the unified direction intersection boundary line group, extract the extension trajectory and relative coordinate set of the path in space, and judge the extension direction classification between path segments according to the starting direction parameter and direction consistency standard of the illumination path, and obtain the direction difference range of the illumination path. Based on the fundamental concept of all illumination path segments within a unified direction convergence boundary line group, the process begins by reading each illumination path segment individually. During this process, the starting and ending coordinates of each segment in space, along with the resulting extended trajectory information, are extracted sequentially. Simultaneously, the segment's numbering order in the original path set is retrieved as an association index. Building upon this, the spatial extension trajectory of a single path segment is decomposed into several consecutive coordinate points along the extension direction. For example, a 1.2-meter-long path segment is divided into six coordinate points spaced 0.2 meters apart. These coordinate points are then arranged into a relative coordinate set according to their generation order. Next, the starting direction parameter corresponding to this path segment is read. This direction parameter is determined by the direction component of the first segment of the path and serves as the direction reference value for the path. Finally, adjacent path segments are paired according to their path numbering order. The values ​​of each directional component are extracted and compared. Directional consistency is determined by reading the component differences in length, width, and height directions item by item. When the difference between adjacent path segments in the three directional components is less than the preset consistency standard interval, they are determined to be extended in the same direction. When the difference between any directional component is in the middle interval, it is recorded as directional offset. When the difference exceeds the maximum interval threshold, it is recorded as a significant change in direction. For example, when the difference in length component is less than 0.1, it is recorded as consistent; between 0.1 and 0.3, it is recorded as offset; and greater than 0.3, it is recorded as change. By repeating the directional component reading and interval judgment operation on all adjacent segments in the entire path, a sequence of directional difference records of the path in the overall extension process is formed. All recorded difference values ​​are classified and organized according to the interval range, and finally, the complete directional difference interval of the illumination path is obtained.

[0029] S402: Based on the difference range of the illumination path direction, track the relative displacement of the illumination path in the spatial extension, identify the path unit with interrupted directional continuity according to the coordinate set of the midpoint of the adjacent path and the spatial distance relationship, and call its corresponding coordinate index to perform the stripping operation to obtain the coordinate set of the boundary discontinuous region. Based on the input of the difference range of the illumination path direction, the extension trajectory of each illumination path is first re-traversed. During the execution, the midpoint coordinates of each continuous line segment in the path are read one by one. The midpoint coordinates are determined by reading the spatial positions of the start and end points of the line segment and taking the median of their spatial positions. Then, the midpoint coordinates of adjacent line segments are combined into midpoint coordinate pairs, and spatial distance judgment is performed on the coordinate pairs. The judgment is made by comparing the coordinate differences between the two midpoints in the length, width, and height directions to determine their relative displacement. When the displacement in all three directions is within the continuous variation range, it is recorded as normal extension. When the displacement in any direction exceeds the preset continuous range, it is recorded as an extension anomaly, such as when the displacement in the length direction exceeds 0. A line segment is marked as interrupted when its displacement exceeds 0.5 meters or 0.3 meters in the vertical direction. Combined with the classification results of the line segment within the direction difference interval, when a line segment is marked as having both a significant change in direction and an abnormal midpoint displacement, it is determined as a path unit with a discontinuous direction. Then, the path number and coordinate index information corresponding to this path unit are retrieved, and the line segment and its associated midpoint coordinates are stripped. During the stripping process, it is removed from the original path extension sequence and recorded separately. For the remaining line segments in the same path, the midpoint displacement judgment and direction interval comparison are continued until the entire path is traversed. Finally, the spatial coordinates corresponding to all the stripped path units are summarized to form a set of coordinates for the boundary discontinuous region.

[0030] S403: Based on the set of coordinates of the discontinuous boundary region, filter the path units with the same extension direction from the remaining path set, perform sequence matching based on the direction vector group and the path number, extract the spatial index set within the closed path region, and obtain the terahertz irradiation stable block.

[0031] Given the known coordinate set of the discontinuous boundary region, the remaining unstripped path segments are first centrally read and rearranged according to their original path numbering order. During the process, the direction component data of each path segment is extracted one by one, and the direction component values ​​are used as the matching criteria. When the component values ​​of multiple path segments in the length or height direction are in the same direction interval, they are grouped into the same candidate path unit. Subsequently, within each candidate path unit, sequence matching is performed according to the path numbering order. During the matching process, the continuity of the numbering and the difference of the direction components of adjacent path segments are compared one by one. Only when the numbers are continuous and the difference of the direction components remains within a consistent interval are they included in the same contiguous path unit. For each continuous path unit, a spatial closure check is performed. By comparing the endpoint coordinates of the first and last line segments of the path unit, if the spatial distance is less than the preset closure judgment range, the spatial range corresponding to the path unit is marked as a closed region, and the spatial index numbers corresponding to all line segments in the region are extracted. For example, if the difference between the first and last coordinates of a group of path line segments numbered 12 to 18 is less than 0.1 meters, then all spatial indices corresponding to numbers 12 to 18 are extracted as a group of region indices. The direction matching, sequence comparison and closure judgment operations are repeated for all path units. Finally, the set of spatial indices located in the closed path region is obtained, and the terahertz irradiation stable block is obtained accordingly.

[0032] Please see Figure 6 The steps for obtaining terahertz imaging control commands are as follows: S501: Call all the irradiation point information involved in the terahertz irradiation stability block, extract the position coordinates of each point under different time sequences based on the flight mission time axis, compare the position changes of each point during the mission according to the horizontal and vertical displacement directions, and obtain the irradiation point position offset sequence. Based on the fundamental information of all irradiation points involved in the terahertz irradiation stable block, each irradiation point within the stable block is first numbered and read. During execution, the time stamp information of each point throughout the entire flight mission is retrieved, and sorted according to the flight mission timeline. The spatial coordinates of the same point at different time points are extracted one by one. The spatial coordinates are recorded in the form of numerical values ​​in the length, lateral, and height directions. For example, the coordinates of an irradiation point at 10 seconds are (8.0, 1.2, 0.9), and at 12 seconds are (8.1, 1.25, 0.9). Subsequently, the coordinate data at adjacent time points are compared. During the comparison, the difference in lateral and vertical coordinates is read separately. The horizontal direction is defined as the width direction of the structure, and the vertical direction is defined as the length direction of the structure. When the absolute value of the difference in the horizontal coordinates is greater than 0.05 meters, it is recorded as a horizontal offset; when the absolute value of the difference in the vertical coordinates is greater than 0.05 meters, it is recorded as a vertical offset. If neither of them exceeds the range, it is recorded as a stable state. For example, if the horizontal change of the above point is 0.05 meters and the vertical change is 0.1 meters, then this temporal change is recorded as a vertical offset. The coordinate reading and direction judgment operation is repeated for the same irradiation point throughout the entire task time sequence. The offset direction and the corresponding offset amount at each moment are connected in time sequence to form an offset record chain. Finally, the same processing is performed on all irradiation points in the stable block, and the resulting sequence is an irradiation point position offset sequence composed of multiple point offset record chains.

[0033] S502: Based on the irradiation point position offset sequence, filter out the point group that frequently offsets in the task time sequence, call the offset direction corresponding to its continuous position change, identify the sudden change stage in combination with the time interval, and classify and analyze each stage to obtain offset dynamic prediction reference data. Based on the input condition of the irradiation point position offset sequence, the offset record chain corresponding to each irradiation point in the sequence is first read one by one. During the execution, the number of times the point is recorded as having an offset in the entire task time sequence is counted. When the proportion of the offset count to the total number of records for that point exceeds 30%, the point is marked as a frequently offset point. For example, if a point is recorded as having a horizontal or vertical offset 8 times in 20 time nodes, its offset proportion is 40%, which meets the frequent offset condition. Then, all points marked as frequently offset are gathered into a point group. For each point in the point group, its continuous offset records are read, and the offset direction is sorted according to the time sequence. The offset records with the same continuous direction are divided into... For the same stage, when the direction of two adjacent offsets changes or the time interval exceeds a preset threshold, it is divided into a new stage. For example, if a point continuously offsets vertically between 10 and 14 seconds, and then shifts to lateral offset between 15 and 16 seconds, two stages are formed. During the stage division process, the time interval between adjacent offset records is read simultaneously. When the time interval is less than 2 seconds and the offset suddenly increases by more than 0.1 meters, the stage is marked as a mutation stage. After completing the stage identification of a single point, the offset direction sorting, stage division and mutation marking operations are repeated for all points in the point group. Finally, the stage number, offset direction, duration and corresponding time interval of each point are summarized to form offset dynamic prediction reference data.

[0034] S503: Call the point number and illumination direction information involved in the offset dynamic prediction reference data, adjust the pointing parameters of frequently changing points and the displacement switching content in the flight control template, make corresponding modifications to the illumination timing and displacement trigger nodes, and obtain terahertz imaging control commands.

[0035] The input, dynamic offset prediction reference data, is used to first read the illumination point number information and corresponding offset direction records one by one. During execution, the point number is used as an index to call the corresponding pointing parameters and displacement switching configuration items in the flight control template for that point. Then, based on the time interval of the abrupt change phase recorded in the prediction reference data, the original displacement trigger time node in the flight control template is compared. If the predicted abrupt change phase is earlier than the original trigger node, the trigger node is advanced by the corresponding time difference; if the predicted abrupt change phase is later than the original trigger node, the trigger node is delayed by the corresponding time difference. For example, if the original displacement trigger time is set to 15 seconds, but the prediction data indicates that the point experiences a directional change at 13 seconds, then... The trigger node is adjusted to the 13th second. At the same time, the direction correction operation is performed on the irradiation direction parameters corresponding to the frequently offset points. During the correction process, the dominant offset direction in the prediction stage is directly read and written into the control template as a new pointing parameter. For example, if the original pointing parameter is longitudinal, but the offset direction in the prediction stage is concentrated in the lateral direction, the pointing parameter is adjusted to lateral. The original template parameters of the other points that are not marked as frequently offset remain unchanged. After the parameter adjustment of a single point is completed, the next point is processed until the template of all points in the prediction reference data is updated. Finally, the updated pointing parameter set, displacement trigger node sequence and irradiation timing configuration are exported in a unified manner to form the corresponding terahertz imaging control command.

[0036] 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 dual-UAV cooperative terahertz imaging detection system and control method, characterized in that, Includes the following steps: S1: Obtain the flight status of the two UAVs around the structural components. The transmitter projects terahertz beams onto the surface of the structure, and the receiver flies synchronously close to the intersection area. It samples the spatial points formed by the illumination, binds the sampling action to the flight position, and stores the corresponding spatial points in sequence. The points correspond to the illumination data, and outputs a three-dimensional energy flow illumination layer. S2: Based on the relative energy flow relationship between the irradiated points in the three-dimensional energy flow irradiation layer, extract the energy flow change of continuous points along the flight trajectory direction, observe whether a continuous structure is formed in the energy flow abrupt change region, focus on processing the linear segments formed by continuous abrupt changes, and extract the edge path band that constitutes the energy flow jump. S3: Based on the change of the irradiation direction of the line segment in the energy flow jump edge path zone, the direction difference of the continuous irradiation points is compared, and the irradiation trajectory is continued to be tracked in the area with consistent path direction. The segments with abrupt angle changes are eliminated, and the path segments with consistent direction are retained and gathered into a group of intersection and boundary lines with unified direction. S4: Call the irradiation path segments in the unified direction intersection boundary line group, track the path extension distance during its spatial extension, process the connection status between paths, and in the remaining irradiation segment space after stripping the discontinuous boundary region, delineate the irradiation area with the same extension direction and extract it as a terahertz irradiation stable block.

2. The dual-UAV cooperative terahertz imaging detection system and control method according to claim 1, characterized in that: The three-dimensional energy flow irradiation layer includes a set of spatial point coordinates, energy flow distribution characteristics, irradiation coverage density identifiers, and trajectory mapping indexes. The energy flow jump edge path zone specifically includes energy flow change amplitude characteristics, continuous edge segments, path direction characteristics, and jump aggregation area identifiers. The directional unified intersection boundary line group includes directional consistency characteristics, effective line segment set, intersection boundary description, and directional stability identifiers. The terahertz irradiation stable block specifically includes stable coverage range, regional consistency characteristics, spatial connectivity area identifiers, and irradiation uniformity parameters.

3. The dual-UAV cooperative terahertz imaging detection system and control method according to claim 1, characterized in that: The steps for obtaining the three-dimensional energy flow illumination layer are as follows: S101: Obtain the flight path information of the two UAVs around the structural component. The transmitting UAV continuously emits terahertz beams to the surface of the structure according to the path. The receiving UAV follows the path at the boundary of the irradiation area according to the synchronization command and records the spatial position in the intersection area to generate a set of irradiation trajectory points. S102: Based on the set of illumination trajectory points, call the receiving device of the receiving drone to obtain the reflected signal intensity, frequency and arrival time of each point, bind the corresponding illumination parameters with the flight position, and obtain the point illumination signal parameter group after sorting them in time order. S103: Based on the point illumination signal parameter group, extract the spatial coordinates and illumination parameter information, generate the spatial distribution relationship of the points in the structural surface model, construct the mapping matrix between signal intensity and coordinates, and obtain the three-dimensional energy flow illumination layer.

4. The dual-UAV cooperative terahertz imaging detection system and control method according to claim 1, characterized in that: The steps for obtaining the energy flux jump edge path band are as follows: S201: Based on the relative energy flow relationship between each pair of irradiation points in the three-dimensional energy flow irradiation layer, call the direction vector, intensity and spatial coordinates of the irradiation points, extract continuous points along the flight trajectory direction, calculate the energy flow change between adjacent points, and obtain the energy flow change sequence of the irradiation points. S202: Based on the energy flow change sequence of the irradiation point, screen the mutation location and extract continuous points. Combine their spatial distribution direction and point clustering, identify the related continuous regions and obtain the spatial mutation extension fragment. S203: Call the points in the spatial mutation extension segment, analyze the continuity characteristics between spatial direction vectors, screen the continuous point series with similar directions, connect them sequentially to form a linear region, and obtain the energy flow jump edge path zone.

5. The dual-UAV cooperative terahertz imaging detection system and control method according to claim 1, characterized in that: The steps for obtaining the unified directional intersection boundary line group are as follows: S301: Based on the set of all line segments extracted from the energy flow jump edge path zone, call the start and end point position information of each line segment and its order relationship in the sequence, calculate the change value of the illumination direction between adjacent line segments, filter the continuous line segment segments with consistent illumination direction, and generate a set of line segment segments with consistent direction. S302: Based on each continuous segment in the set of line segments with consistent direction, call the direction change value of adjacent line segments, determine the degree of deviation from the direction of the segment, remove the line segments with abrupt changes in direction, and retain the path line segments with stable direction to obtain a group of path direction stable segments. S303: Call all line segments in the path direction stable segment group, merge them into path groups in the same direction according to the direction value, identify the spatial intersection points in each path group, integrate the path boundaries in the intersection structure, and obtain a direction-unified intersection boundary line group.

6. The dual-UAV cooperative terahertz imaging detection system and control method according to claim 1, characterized in that: The steps for obtaining the terahertz irradiation stable block are as follows: S401: Call all the illumination path segments in the unified direction intersection boundary line group, extract the extension trajectory and relative coordinate set of the path in space, and judge the extension direction classification between the path segments according to the starting direction parameter and direction consistency standard of the illumination path, and obtain the direction difference range of the illumination path. S402: Based on the directional difference range of the irradiation path, track the relative displacement of the irradiation path in the spatial extension, identify the path unit with interrupted directional continuity according to the coordinate set of the midpoint of the adjacent path and the spatial distance relationship, and call its corresponding coordinate index to perform the stripping operation to obtain the coordinate set of the boundary discontinuous region. S403: Based on the set of coordinates of the discontinuous boundary region, filter the path units with the same extension direction from the remaining path set, perform sequence matching based on the direction vector group and the path number, extract the spatial index set within the closed path region, and obtain the terahertz irradiation stable block.

7. The dual-UAV cooperative terahertz imaging detection system and control method according to claim 1, characterized in that: Also includes: S5: Call up all illumination point information in the terahertz illumination stabilization block, compare the position changes of the illumination points during the flight mission, adjust the flight control actions according to the position changes, embed illumination pointing and displacement switching control actions in the flight control template, and output terahertz imaging control commands; The terahertz imaging control commands include flight attitude control parameters, illumination pointing configuration parameters, displacement switching control parameters, and imaging task constraint parameters.

8. The dual-UAV cooperative terahertz imaging detection system and control method according to claim 7, characterized in that: The steps for obtaining the terahertz imaging control command are as follows: S501: Call up all the irradiation point information involved in the terahertz irradiation stability block, extract the position coordinates of each point under different time sequences based on the flight mission time axis, and compare the position changes of each point during the mission process according to the horizontal and vertical displacement directions to obtain the irradiation point position offset sequence. S502: Based on the irradiation point position offset sequence, filter out the point group that frequently shifts in the task time sequence, call the offset direction corresponding to its continuous position change, identify the sudden change stage in combination with the time interval, and classify and analyze each stage to obtain offset dynamic prediction reference data. S503: Call the point number and illumination direction information involved in the offset dynamic prediction reference data, adjust the pointing parameters of frequently changing points and the displacement switching content in the flight control template, make corresponding modifications to the illumination timing and displacement triggering nodes, and obtain terahertz imaging control commands.

Images