A radar manual plotting automatic evaluation method and system based on structured geometric drawing chain constraints

Abstract

The application provides a radar manual plotting automatic evaluation method and system based on structured geometric plotting chain constraint, and belongs to the field of marine radar collision avoidance technology. The method sequentially performs field reading, plotting area positioning, ink extraction, standard answer calculation, standard geometric plotting chain generation, mask matching and index calculation, item evaluation and visual feedback on the electronic plotting graph. The electronic plotting graph is represented as a structured geometric answer composed of observation points, relative motion lines, perpendicular lines, construction lines, action points and post-avoidance relative motion lines and other geometric elements. The coverage index is used to measure the missing and incomplete degree of the geometric elements. The geometric accuracy index is used to punish the ink of multiple drawing, disorderly drawing and deviation from the standard geometric area. The key composition step coverage is introduced to identify the similar post-avoidance relative motion lines but the error in the intermediate drawing step. Finally, the evaluation is completed by the structured geometric constraint and the interpretable index.

Description

Technical Field

[0001] This invention relates to the field of collision avoidance technology for marine radar, and more particularly to an automatic evaluation method and system for radar manual plotting based on structured geometric drawing chain constraints. Background Technology

[0002] Manual radar plotting is a crucial component of maritime radar observation and collision avoidance decision-making training. Trainees are required to plot the relative position of the target vessel on a radar chart based on multiple radar observations, connect the lines to obtain the original relative motion line, construct a relative motion triangle to determine the target vessel's true heading and speed, calculate the nearest encounter distance and time to reach the nearest encounter point from the relative motion line, assess the encounter situation, formulate and plot an avoidance plan, and then verify the avoidance effect. The above plotting process requires not only correct final results but also correct plotting steps.

[0003] For the evaluation of radar manual plotting responses, existing technologies mainly fall into two categories: The first is the traditional evaluation method based on manual interpretation by teachers. This requires evaluators to check each item on the radar manual plotting evaluation form to verify the trainee's plotting process, calculation results, and avoidance conclusions. While this method can incorporate teacher experience to judge the response process, the evaluation process relies on manual observation and subjective judgment, resulting in a large workload, long feedback cycles, and difficulty in ensuring consistency among different evaluators. It is ill-suited for the automated evaluation needs of large-scale training responses. Even if trainees complete the plotting using radar manual plotting training software, the electronic plotted graph usually needs to be submitted or saved separately for manual interpretation by evaluators, failing to achieve independent, rapid, and repeatable automated evaluation. The second category is the automated evaluation method developed in recent years based on image processing or black-box models. For example, it uses whole-image pixel similarity, binary ink overlap comparison, or deep learning models to identify the electronic plotted graph and provide evaluation results.

[0004] However, existing automatic evaluation methods for radar manual plotting based on image processing or black box models still have the following significant defects: (1) They lack a structured geometric understanding of the answer ink marks and cannot clearly determine whether a key geometric element has been drawn completely; (2) They are difficult to distinguish between different types of errors such as omissions, under-drawing, over-drawing, random drawing, and line segments deviating from the standard geometric area; (3) They do not explicitly constrain the drawing order and geometric structure relationship of radar manual plotting, and are prone to giving higher scores to answers with "similar results but incorrect steps"; (4) The evaluation results lack interpretable basis and it is difficult to output the deduction reasons corresponding to specific evaluation items.

[0005] Therefore, there is an urgent need to propose an automatic evaluation method and system for radar manual plotting that is independent of radar manual plotting training software, oriented towards electronic plotting, capable of performing structured geometric understanding of the answer ink marks, capable of checking key plotting steps, capable of distinguishing error types such as omissions and extra plots, and capable of outputting interpretable sub-scores and visual feedback results, in order to overcome the above-mentioned deficiencies. Summary of the Invention

[0006] To address the technical problems of existing radar manual plotting evaluation methods, such as high labor costs, inability to check plotting steps by only comparing final results, insensitivity to overdrawing or missing plots, and lack of interpretable sub-scores, this invention provides an automatic evaluation method and system for radar manual plotting based on structured geometric plotting chain constraints. This invention uses an independent desktop evaluation program as the application platform, representing the electronic plotted graph as a structured geometric plotting chain composed of geometric elements in the plotting sequence. Interpretable indicators such as coverage, geometric accuracy, and coverage of key plotting steps are used to constrain and evaluate the responses, thereby improving the objectivity, interpretability, and robustness of the evaluation.

[0007] The technical means employed in this invention are as follows: An automatic evaluation method for radar manual plotting based on structured geometric drawing chain constraints includes: S1. Obtain the radar manual plotting electronic drawing diagram, perform optical character recognition on the question table fields and answer fields in the electronic drawing diagram, obtain the pre-filled results of question data and answer data, and establish a data input chain for automatic evaluation of radar manual plotting. S2. Based on the electronic plotting diagram in the data input chain, combined with the radar image green grid positioning results and the auxiliary detection results of the target detection model, determine the radar plotting area range, and establish the conversion relationship between radar plane coordinates and pixel plane coordinates based on the plotting area boundary, the ship's origin position and radar range. S3. Extract the student's black ink marks from the drawing area and generate a mask of the answer ink marks. And preserve the lines, points, and compositional traces; S4. Based on the three observation data in the question, the ship's course, the ship's speed, and the rules of manual navigation plotting, calculate the target ship's motion parameters, the nearest encounter distance, the time to reach the nearest encounter point, and the standard answer for avoidance. S5. Generate a standard geometric construction chain based on the standard answer for obstacle avoidance, including three observation points, the original relative motion line, the nearest encounter distance perpendicular line, the relative motion triangle construction line, the action point, the relative motion line after avoidance, and the construction line in the direction change or speed change scenario, and generate corresponding standard element masks for each standard geometric element. The union of all standard element masks is denoted as the standard geometric region mask. ; S6. Mask the ink of the answer Masks with standard elements Perform matching and calculate element coverage. and geometric precision And calculate the coverage of key plotting steps; S7. Calculate the numerical error based on the answer data and the standard answer, and integrate the numerical error and element coverage. Geometric accuracy And the coverage of key mapping steps, generate sub-scores according to the ten evaluation items of radar manual plotting; S8. Correct the total score according to the geometric consistency upper limit rule, and output the total score, the scores of each item, the reasons for the deduction of each item, and a visual feedback chart that overlays the student's ink marks with the standard geometric construction chain.

[0008] Further, step S1 includes: S11. Receive electronic plotting diagrams from radar manual plotting training software; S12. In the independent desktop evaluation interface, perform optical character recognition on the question table fields of the electronic drawing diagram to obtain the ship's heading, ship's speed and the target ship's three observed bearings and distances, as the question data pre-filling results. S13. In the independent desktop evaluation interface, perform optical character recognition on the answer fields of the electronic drawing diagram to obtain the target ship's true course, true speed, nearest encounter distance, arrival time at the nearest encounter point, and avoidance maneuver parameters, as the answer data pre-fill result; S14. Display the pre-filled results of the question data and the pre-filled results of the answer data in the independent desktop assessment interface, and provide a manual verification and correction interface for assessors to verify and correct the fields. S15. Use the corrected question data and the corrected answer data as the data input chain.

[0009] Further, step S2 includes: S21. Based on the concentric circles, azimuth scale, grid line color and geometric distribution characteristics of the radar map in the electronic drawing, detect green grids and determine candidate areas of the drawing area; S22. Use an object detection model to detect the electronic drawing and output the candidate bounding box of the drawing area and the candidate bounding box of the preset geometric elements. S23. The green grid positioning results are fused with the candidate bounding boxes output by the target detection model to determine the boundary of the radar plotting area, the origin position of the ship, and the effective plotting radius. S24. Based on the boundary of the drawing area, the location of the ship's origin, and the radar range read from the data in the question, establish the conversion relationship between the radar plane coordinates and the pixel plane coordinates with the ship's origin as the coordinate origin.

[0010] Further, step S3 includes: S31. Extract the binary foreground of the black ink marks in the drawing area based on the color threshold, and distinguish the dark handwriting drawn by the students from the light background, green grid and question table. S32. Dilate the binary foreground to connect the tiny gaps caused by screenshot compression, thin handwriting, or broken lines. S33. Within the candidate bounding box of geometric elements output by the target detection model, perform a closing operation on the dilated binary foreground to enhance the continuity of points, line segments, and local composition traces. S34. Use Hough transform to detect line segments from the binary foreground, and merge the line segment detection results with the binary processing results to obtain the answer ink mask. .

[0011] Further, step S4 includes: S41. Based on the three observation azimuths and distances in the question data, and combined with the established conversion relationship between radar plane coordinates and pixel plane coordinates, determine the positions of the target ship's first observation point A1, second observation point A2, and last observation point A3 in the pixel plane. S42. Determine the original relative motion line by connecting the first observation point A1 and the last observation point A3, and use the intermediate observation point A2 for the evaluation of the integrity and coverage of the observation points. S43. Use the second observation point A2 as an intermediate observation point for evaluation of the completeness and coverage of subsequent observation points; S44. Draw a perpendicular line from the origin of this ship to the original relative motion line, and obtain the nearest meeting distance DCPA and the time to reach the nearest meeting point TCPA based on the position of the foot of the perpendicular. S45. Construct a relative motion triangle based on the ship's motion vector and the original relative motion line, and calculate the target ship's true heading and true speed. S46. Determine the action point A4 based on the avoidance maneuver parameters, construct the relative motion line after avoidance, and construct a new motion triangle based on the relative motion line after avoidance and the motion vector of the ship itself. S47. Based on the new motion triangle, calculate the new course or speed of the ship after the avoidance maneuver.

[0012] Further, step S5 includes: S51. Based on the calculated nearest encounter distance, time to reach the nearest encounter point, target ship's true course, target ship's true speed, action point, relative motion line after avoidance, and the ship's new course or speed after avoidance, generate a standard geometric drawing chain composed of several standard geometric elements in the order of drawing. Among them, the standard geometric elements include the first observation point A1, the second observation point A2, the last observation point A3, the original relative motion line determined by the first observation point A1 and the last observation point A3, the DCPA perpendicular line drawn from the ship's origin to the original relative motion line, the A1-M construction line and the M-A3 construction line in the relative motion triangle, action point A4, the relative motion line after avoidance starting from action point A4, and the Q construction line in the course change scenario or the R construction line in the speed change scenario. S52. Generate a corresponding standard element mask for each standard geometric element. The observation point generates a circular mask area with a preset radius, the line segment generates a strip mask area with a preset line width, the perpendicular line generates a strip mask area with a preset line width, and the construction line generates a strip mask area with a preset line width. S53, Mask all standard elements Perform a union operation to obtain a standard geometric region mask. .

[0013] Further, step S6 includes: S61. Mask the ink marks of student answers. Masks with standard elements Perform pixel-level matching element by element, and calculate the element coverage for each standard geometric element. :

[0014] In the above formula, This indicates the number of foreground pixels in the mask. Indicates the student's ink mask and the first The intersection of standard element masks; element coverage This is used to measure whether the standard geometric element has been completely drawn by the student. A low coverage rate indicates that the corresponding point, line segment or construction step has been omitted, missing or incomplete. S62. Calculate the ink mask used for answering questions. Relative to standard geometric region mask geometric precision :

[0015] Geometric accuracy Used to measure the proportion of a student's ink marks that fall within the standard geometric area; when a student draws too many, draws randomly, or the line segments deviate significantly from the standard geometric area, the area falling within the standard geometric area is masked. In addition to increased ink marks, geometric precision It then decreased; Further, step S7 includes: S71. Extract the answer values ​​for each numerical item from the answer data, extract the standard values ​​for the corresponding numerical items from the standard answers, and calculate the error for each numerical item. ; S72. Set full tolerance for each numerical item. and zero threshold ,when When the error score for this numerical item is... ;when When the error score for that numerical item is 0, the score is 0. When this value is calculated linearly, the scoring function is: ; S73. Integrate the error scores of each numerical item with the element coverage, geometric accuracy and key mapping step coverage. According to the ten evaluation items of radar manual plotting, match each evaluation item with the corresponding geometric elements and numerical quantities to generate sub-item scores.

[0016] Further, step S8 includes: S81. Calculate the geometric consistency score based on the obtained element coverage, geometric accuracy, and key composition step coverage. S82. Set the total score correction upper limit according to the different levels of the geometric consistency score, where each level corresponds to a different total score upper limit value; S83. Determine the current level of geometric consistency score. If key geometric elements are missing, geometric accuracy is too low, or key composition steps are not covered enough, then the sum of the sub-scores generated in S7 is corrected according to the upper limit of the total score corresponding to this level to obtain the corrected total score. S84. Overlay the generated answer ink mask with the generated standard geometric drawing chain to generate a visual feedback diagram, in which the student ink and standard geometric elements are displayed in different visual styles. S85. Output the corrected total score, the score for each sub-item, the reason for deduction for each sub-item, and a visual feedback chart.

[0017] This invention also provides an automatic evaluation system for radar manual plotting based on structured geometric construction chain constraints, implemented using the aforementioned automatic evaluation method for radar manual plotting based on structured geometric construction chain constraints. The system includes: a data input module, a plotting area positioning and coordinate conversion module, an ink extraction module, a standard answer calculation module, a standard geometric construction chain generation module, a mask matching and index calculation module, a sub-item evaluation module, and a total score correction and visualization feedback module, wherein: The data input module acquires the radar manual plotting electronic drawing, performs optical character recognition on the question table fields and answer fields in the electronic drawing, obtains the pre-filled results of the question data and answer data, and establishes a data input chain for automatic evaluation of radar manual plotting. The mapping area positioning and coordinate conversion module is used to determine the radar mapping area range based on the electronic mapping diagram in the data input chain, combined with the radar image green grid positioning result and the auxiliary detection result of the target detection model, and to establish the conversion relationship between radar plane coordinates and pixel plane coordinates based on the mapping area boundary, the ship's origin position and radar range. The ink extraction module is used to extract the student's black ink marks within the drawing area and generate an answer ink mark mask. And preserve the lines, points, and compositional traces; The standard answer calculation module is used to calculate the target ship's motion parameters, nearest encounter distance, arrival time at the nearest encounter point, and avoidance standard answer based on the three observation data in the question data, the ship's course, the ship's speed, and the rules of nautical manual plotting; The standard geometric construction chain generation module is used to generate a standard geometric construction chain based on the avoidance standard answer, which includes three observation points, the original relative motion line, the nearest encounter distance perpendicular line, the relative motion triangle construction line, the action point, the relative motion line after avoidance, and the construction line in the direction change or speed change scenario, and to generate a corresponding standard element mask for each standard geometric element. The union of all standard element masks is denoted as the standard geometric region mask. ; The mask matching and index calculation module is used to mask the answer ink marks. Masks with standard elements Perform matching and calculate element coverage. and geometric precision And calculate the coverage of key plotting steps; The sub-evaluation module is used to calculate the numerical error based on the answer data and the standard answer, and to integrate the numerical error and element coverage. Geometric accuracy And the coverage of key mapping steps, generate sub-scores according to the ten evaluation items of radar manual plotting; The total score correction and visualization feedback module corrects the total score according to the geometric consistency upper limit rule, and outputs the total score, the scores of each item, the reasons for deductions for each item, and a visualization feedback chart that overlays the student's ink marks with the standard geometric construction chain.

[0018] Compared with the prior art, the present invention has the following advantages: 1. The automatic evaluation method for radar manual plotting based on structured geometric drawing chain constraints provided by this invention establishes a data input chain by facing the electronic drawing diagram output by radar manual plotting training software. Optical character recognition is used only as the pre-filled result of the question table field and the answer field. After manual verification, it enters the subsequent evaluation. This decouples the evaluation process from the running state of the training software and avoids the compatibility and reliability problems caused by over-reliance on the internal data interface of the software or the embedded drawing control.

[0019] 2. The automatic evaluation method for radar manual plotting based on structured geometric drawing chain constraints provided by this invention determines the radar plotting area range by combining the green grid positioning results with the target detection model auxiliary detection results, and establishes the conversion relationship between radar plane coordinates and pixel plane coordinates. This realizes the redundancy verification of plotting area positioning and the accurate mapping of coordinate transformation, thereby improving the positioning reliability of subsequent ink mark extraction and geometric matching.

[0020] 3. The automatic evaluation method for radar manual plotting based on structured geometric drawing chain constraints provided by this invention extracts the student's black plotting ink marks in the drawing area and forms a mask of the answer ink marks, retaining evaluation clues such as lines, points and composition traces, and realizes the transformation of the student's answer from ordinary screenshots into the image foundation construction of a set of computable geometric elements, thus providing data support for structured geometric evaluation.

[0021] 4. The radar manual plotting automatic evaluation method based on structured geometric drawing chain constraints provided by this invention represents the standard answer as a structured geometric drawing chain containing three observation points, the original relative motion line, the nearest encounter distance perpendicular line, the construction line of the relative motion triangle, the action point, the relative motion line after avoidance, and the construction line in the scenario of changing direction or speed, rather than an ordinary pixel image. This realizes the semantic improvement of the evaluation object from an unstructured image to an ordered set of geometric elements, making the evaluation have clear geometric meaning, divisibility, and interpretability.

[0022] 5. The automatic evaluation method for radar manual plotting based on structured geometric drawing chain constraints provided by this invention uses element coverage rate to measure the degree of omission and incompleteness of each standard geometric element, and uses geometric precision to penalize over-drawing, random drawing, and ink marks that deviate from the standard geometric area. This achieves the distinction and identification of two types of error: omission and over-drawing, and overcomes the shortcomings of existing whole-image pixel similarity or binary ink mark overlap comparison methods that cannot distinguish error types.

[0023] 6. The automatic evaluation method for radar manual plotting based on structured geometric plotting chain constraints provided by this invention introduces a coverage index for key plotting steps, and independently checks intermediate plotting steps such as construction lines of relative motion triangles, construction lines in change-of-course or change-of-velocity scenarios, and translational construction lines of new course. This achieves accurate identification of the response "the relative motion lines are similar after avoidance but the intermediate plotting steps are incorrect", avoiding the limitation of existing black-box models that only compare the final results and cannot check the plotting steps.

[0024] 7. The automatic evaluation method for radar manual plotting based on structured geometric construction chain constraints provided by this invention corrects the total score according to the geometric consistency upper limit rule and outputs the total score, sub-item scores, reasons for deduction, and a visual feedback diagram that overlays the student's ink marks with the standard geometric construction chain. This achieves the objectivity, interpretability, and convenience of manual review of the evaluation results, and overcomes the shortcomings of existing automatic evaluation methods that lack interpretable basis and are difficult to output reasons for deduction corresponding to specific evaluation items.

[0025] In summary, by applying the technical solution of this invention, the existing radar manual plotting evaluation methods suffer from high labor costs, the inability to check plotting steps by only comparing the final results, insensitivity to overdrawing or missing plots, and a lack of interpretable sub-item scores. These shortcomings are systematically addressed through structured geometric plotting chain constraints, dual-index measurement of coverage and geometric accuracy, coverage checks for key plotting steps, and a geometric consistency upper limit correction mechanism. Therefore, the technical solution of this invention solves the problems of existing radar manual plotting evaluation methods, such as reliance on large amounts of manual interpretation, lack of structured geometric understanding in automatic evaluation, inability to distinguish between overdrawing and missing plot errors, lack of explicit constraints on plotting order and geometric construction relationships, and lack of interpretable evaluation results.

[0026] Based on the above reasons, this invention can be widely promoted in fields such as marine radar observation and collision avoidance decision training, ship driver skills assessment, automated teaching evaluation in maritime colleges and universities, and intelligent maritime education equipment. Attached Figure Description

[0027] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0028] Figure 1 This is the overall flowchart of the automatic evaluation method for radar manual plotting based on structured geometric drawing chain constraints according to the present invention.

[0029] Figure 2This is a schematic diagram of the drawing area positioning and coordinate conversion provided in an embodiment of the present invention.

[0030] Figure 3 This is a flowchart for extracting student-drawn ink marks, provided as an embodiment of the present invention.

[0031] Figure 4 This is a schematic diagram of the generation of a standard geometric construction chain provided in an embodiment of the present invention.

[0032] Figure 5 This is a schematic diagram illustrating the standard element mask generation and coverage calculation provided in an embodiment of the present invention.

[0033] Figure 6 This is a schematic diagram illustrating the calculation of geometric accuracy and coverage of key mapping steps provided in an embodiment of the present invention.

[0034] Figure 7 A diagram of the desktop evaluation software interface provided in an embodiment of the present invention.

[0035] Figure 8 The diagram shows the evaluation effect of the response provided in the embodiment of the present invention.

[0036] Figure 9 The image shows the ink extraction effect provided in an embodiment of the present invention.

[0037] Figure 10 This is a diagram illustrating the superposition effect of a standard geometric drawing chain provided in an embodiment of the present invention.

[0038] Figure 11 This is a functional architecture diagram of the radar manual plotting automatic evaluation system of the present invention. Detailed Implementation

[0039] To enable those skilled in the art to better understand the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of the present invention.

[0040] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this invention are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of the invention described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover a non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0041] like Figure 1 As shown, this invention provides an automatic evaluation method for radar manual plotting based on structured geometric drawing chain constraints, comprising: S1. Obtain the radar manual plotting electronic drawing diagram, perform optical character recognition on the question table fields and answer fields in the electronic drawing diagram, obtain the pre-filled results of question data and answer data, and establish a data input chain for automatic evaluation of radar manual plotting. S2. Based on the electronic plotting diagram in the data input chain, combined with the radar image green grid positioning results and the auxiliary detection results of the target detection model, determine the radar plotting area range, and establish the conversion relationship between radar plane coordinates and pixel plane coordinates based on the plotting area boundary, the ship's origin position and radar range. S3. Extract the student's black ink marks from the drawing area and generate a mask of the answer ink marks. And preserve the lines, points, and compositional traces; S4. Based on the three observation data in the question, the ship's course, the ship's speed, and the rules of manual navigation plotting, calculate the target ship's motion parameters, the nearest encounter distance, the time to reach the nearest encounter point, and the standard answer for avoidance. S5. Generate a standard geometric construction chain based on the standard answer for obstacle avoidance, including three observation points, the original relative motion line, the nearest encounter distance perpendicular line, the relative motion triangle construction line, the action point, the relative motion line after avoidance, and the construction line in the direction change or speed change scenario, and generate corresponding standard element masks for each standard geometric element. The union of all standard element masks is denoted as the standard geometric region mask. ; S6. Mask the ink of the answer Masks with standard elements Perform matching and calculate element coverage. and geometric precision And calculate the coverage of key plotting steps; S7. Calculate the numerical error based on the answer data and the standard answer, and integrate the numerical error and element coverage. Geometric accuracy And the coverage of key mapping steps, generate sub-scores according to the ten evaluation items of radar manual plotting; S8. Correct the total score according to the geometric consistency upper limit rule, and output the total score, the scores of each item, the reasons for the deduction of each item, and a visual feedback chart that overlays the student's ink marks with the standard geometric construction chain.

[0042] In a specific implementation, as a preferred embodiment of the present invention, step S1 includes: S11. Receive electronic plotting diagrams from radar manual plotting training software; S12. In the independent desktop evaluation interface, perform optical character recognition on the question table fields of the electronic drawing diagram to obtain the ship's heading, ship's speed and the target ship's three observed bearings and distances, as the question data pre-filling results. S13. In the independent desktop evaluation interface, perform optical character recognition on the answer fields of the electronic drawing diagram to obtain the target ship's true course, true speed, nearest encounter distance, arrival time at the nearest encounter point, and avoidance maneuver parameters, as the answer data pre-fill result; S14. Display the pre-filled results of the question data and the pre-filled results of the answer data in the independent desktop assessment interface, and provide a manual verification and correction interface for assessors to verify and correct the fields. S15. Use the corrected question data and the corrected answer data as the data input chain.

[0043] In specific implementation, as a preferred embodiment of the present invention, such as Figure 2 As shown, the system combines the radar map's green grid positioning results with the target detection model's assisted detection results to determine the radar mapping area, i.e., step S2, includes: S21. Based on the concentric circles, azimuth scale, grid line color and geometric distribution characteristics of the radar map in the electronic drawing, detect green grids and determine candidate areas of the drawing area; S22. Use an object detection model to detect the electronic drawing and output the candidate bounding box of the drawing area and the candidate bounding box of the preset geometric elements. S23. The green grid positioning results are fused with the candidate bounding boxes output by the target detection model to determine the boundary of the radar plotting area, the origin position of the ship, and the effective plotting radius. S24. Based on the boundary of the drawing area, the location of the ship's origin, and the radar range read from the data in the question, establish the conversion relationship between the radar plane coordinates and the pixel plane coordinates with the ship's origin as the coordinate origin.

[0044] In specific implementation, as a preferred embodiment of the present invention, such as Figure 3 As shown, step S3 includes: S31. Extract the binary foreground of the black ink marks in the drawing area based on the color threshold, and distinguish the dark handwriting drawn by the students from the light background, green grid and question table. S32. Dilate the binary foreground to connect the tiny gaps caused by screenshot compression, thin handwriting, or broken lines. S33. Within the candidate bounding box of geometric elements output by the target detection model, perform a closing operation on the dilated binary foreground to enhance the continuity of points, line segments, and local composition traces. S34. Use Hough transform to detect line segments from the binary foreground, and merge the line segment detection results with the binary processing results to obtain the answer ink mask. .

[0045] In this embodiment, as Figure 9 As shown in the image, the ink extraction results demonstrate the student-drawn lines, points, and auxiliary composition traces separated from the electronic drawing by the system. This result provides an image basis for subsequent calculations of coverage, geometric accuracy, and coverage in key composition steps.

[0046] In a specific implementation, as a preferred embodiment of the present invention, step S4 includes: S41. Based on the three observation azimuths and distances in the question data, and combined with the established conversion relationship between radar plane coordinates and pixel plane coordinates, determine the positions of the target ship's first observation point A1, second observation point A2, and last observation point A3 in the pixel plane. S42. Determine the original relative motion line by connecting the first observation point A1 and the last observation point A3, and use the intermediate observation point A2 for the evaluation of the integrity and coverage of the observation points. S43. Use the second observation point A2 as an intermediate observation point for evaluation of the completeness and coverage of subsequent observation points; S44. Draw a perpendicular line from the origin of this ship to the original relative motion line, and obtain the nearest meeting distance DCPA and the time to reach the nearest meeting point TCPA based on the position of the foot of the perpendicular. S45. Construct a relative motion triangle based on the ship's motion vector and the original relative motion line, and calculate the target ship's true heading and true speed. S46. Determine the action point A4 based on the avoidance maneuver parameters, construct the relative motion line after avoidance, and construct a new motion triangle based on the relative motion line after avoidance and the motion vector of the ship itself. S47. Based on the new motion triangle, calculate the new course or speed of the ship after the avoidance maneuver.

[0047] In specific implementation, as a preferred embodiment of the present invention, such as Figure 4As shown, step S5 includes: S51. Based on the calculated nearest encounter distance, time to reach the nearest encounter point, target ship's true course, target ship's true speed, action point, relative motion line after avoidance, and the ship's new course or speed after avoidance, generate a standard geometric construction chain composed of several standard geometric elements in the order of construction. Among them, the standard geometric elements include the first observation point A1, the second observation point A2, the last observation point A3, the original relative motion line determined by the first observation point A1 and the last observation point A3, the DCPA perpendicular line drawn from the ship's origin to the original relative motion line, the A1-M construction line and the M-A3 construction line in the relative motion triangle, action point A4, the relative motion line after avoidance starting from action point A4, and the Q construction line in the course change scenario or the R construction line in the speed change scenario. In this embodiment, the standard geometric construction chain not only describes the final answer line segment, but also describes the key intermediate construction steps required to obtain the answer.

[0048] S52. Generate a corresponding standard element mask for each standard geometric element. The observation point generates a circular mask area with a preset radius, the line segment generates a strip mask area with a preset line width, the perpendicular line generates a strip mask area with a preset line width, and the construction line generates a strip mask area with a preset line width. S53, Mask all standard elements Perform a union operation to obtain a standard geometric region mask. .

[0049] In this embodiment, as Figure 10 As shown, the system can overlay standard geometric construction chains onto student electronic drawing diagrams or student ink traces to display the positional relationships between standard observation points, standard relative motion lines, standard construction lines, relative motion lines after avoidance, and student ink traces. This overlay result is used to interpret evaluation conclusions and provides a basis for manual review by evaluators.

[0050] In specific implementation, as a preferred embodiment of the present invention, such as Figure 5 As shown, step S6 includes: S61. Mask the ink marks of student answers. Masks with standard elements Perform pixel-level matching element by element, and calculate the element coverage for each standard geometric element. :

[0051] In the above formula, This indicates the number of foreground pixels in the mask. Indicates the student's ink mask and the first The intersection of standard element masks; element coverage This is used to measure whether the standard geometric element has been completely drawn by the student. A low coverage rate indicates that the corresponding point, line segment, or construction step has been missed, omitted, or incomplete. The output coverage rate of each standard element is shown in the table below:

[0052] S62. Calculate the ink mask used for answering questions. Relative to standard geometric region mask geometric precision :

[0053] Geometric accuracy Used to measure the proportion of a student's ink marks that fall within the standard geometric area; when a student draws too many, draws randomly, or the line segments deviate significantly from the standard geometric area, the area falling within the standard geometric area is masked. In addition to increased ink marks, geometric precision It then declined.

[0054] Therefore, element coverage is used to evaluate whether the drawing is complete, while geometric precision is used to evaluate whether the drawing is reasonable. For example... Figure 6 As shown, the system further calculates the coverage rate of key mapping steps such as the A1-M tectonic line, M-A3 tectonic line, Q tectonic line, R tectonic line, and new course translational tectonic line, and uses this as the coverage rate of key mapping steps. Using this indicator, even if the student's final drawn relative motion line after avoidance is relatively close to the standard line segment, if any intermediate mapping steps are missing or obviously incorrect, the system can still deduct points for the corresponding evaluation item or set a maximum total score.

[0055] In a specific implementation, as a preferred embodiment of the present invention, step S7 includes: S71. Extract the answer values ​​for each numerical item from the answer data, extract the standard values ​​for the corresponding numerical items from the standard answers, and calculate the error for each numerical item. ; S72. Set full tolerance for each numerical item. and zero threshold ,when When the error score for this numerical item is... ;when When the error score for that numerical item is 0, the score is 0. When this value is calculated linearly, the scoring function is: ; S73. Integrate the error scores of each numerical item with the element coverage, geometric accuracy and key mapping step coverage. According to the ten evaluation items of radar manual plotting, match each evaluation item with the corresponding geometric elements and numerical quantities to generate sub-item scores.

[0056] In one implementation, the full-score tolerance and zero-score threshold for each numerical item are shown in the table below:

[0057] In a specific implementation, as a preferred embodiment of the present invention, step S8 includes: S81. Calculate the geometric consistency score based on the obtained element coverage, geometric accuracy, and key composition step coverage. S82. Set the total score correction upper limit according to the different levels of the geometric consistency score, where each level corresponds to a different total score upper limit value; S83. Determine the current level of geometric consistency score. If key geometric elements are missing, geometric accuracy is too low, or key composition steps are not covered enough, then the sum of the sub-scores generated in S7 is corrected according to the upper limit of the total score corresponding to this level to obtain the corrected total score. S84. Overlay the generated answer ink mask with the generated standard geometric drawing chain to generate a visual feedback diagram, in which the student ink and standard geometric elements are displayed in different visual styles. S85. Output the corrected total score, the score for each sub-item, the reason for deduction for each sub-item, and a visual feedback chart.

[0058] In one implementation, the upper limit for geometric consistency is set to 94, 88, 82, and 78 at different levels. When key geometric elements are missing, geometric accuracy is too low, or the coverage of key construction steps is insufficient, the system still sets a corresponding upper limit for the total score, even if some numerical results are close to the standard answer.

[0059] The numerical error score and the geometric consistency score are combined according to the evaluation items to obtain the sub-scores; then the total score is adjusted according to the geometric consistency upper limit rule. In one implementation, there are ten sub-scores, each corresponding to the calculated geometric element and numerical quantity, as shown in the table below:

[0060] It should be noted that the tenth item, "Analysis of the causes of errors," is not automatically evaluated by the system in this embodiment and is not included in the total score of automatic evaluation by default. Instead, it is processed manually or by the subsequent text analysis module to ensure the reliability of the evaluation results.

[0061] like Figure 7 As shown, the desktop assessment software interface is used to import electronic graphs, display pre-filled OCR fields, receive manual verification results, initiate single or batch assessments, and output total scores, sub-scores, reasons for deductions, and visual feedback. Assessors can view field recognition results, drawing area positioning results, ink extraction results, and standard overlay results on this interface.

[0062] like Figure 8 As shown in the diagram, the evaluation results illustrate the output of the system after automatically evaluating a student's answer, including the standard answer, the student's answer, sub-item evaluations, total score correction, and reasons for deductions. This diagram serves to explain how the present invention can transform numerical errors, geometric coverage, and the completion status of the mapping steps into interpretable evaluation results.

[0063] like Figure 11 As shown, this embodiment of the invention also provides an automatic evaluation system for radar manual plotting based on structured geometric construction chain constraints, implemented using the aforementioned automatic evaluation method for radar manual plotting based on structured geometric construction chain constraints. The system includes: a data input module, a plotting area positioning and coordinate conversion module, an ink extraction module, a standard answer calculation module, a standard geometric construction chain generation module, a mask matching and index calculation module, a sub-item evaluation module, and a total score correction and visualization feedback module, wherein: The data input module acquires the radar manual plotting electronic drawing, performs optical character recognition on the question table fields and answer fields in the electronic drawing, obtains the pre-filled results of the question data and answer data, and establishes a data input chain for automatic evaluation of radar manual plotting. The mapping area positioning and coordinate conversion module is used to determine the radar mapping area range based on the electronic mapping diagram in the data input chain, combined with the radar image green grid positioning result and the auxiliary detection result of the target detection model, and to establish the conversion relationship between radar plane coordinates and pixel plane coordinates based on the mapping area boundary, the ship's origin position and radar range. The ink extraction module is used to extract the student's black ink marks within the drawing area and generate an answer ink mark mask. And preserve the lines, points, and compositional traces; The standard answer calculation module is used to calculate the target ship's motion parameters, nearest encounter distance, arrival time at the nearest encounter point, and avoidance standard answer based on the three observation data in the question data, the ship's course, the ship's speed, and the rules of nautical manual plotting; The standard geometric construction chain generation module is used to generate a standard geometric construction chain based on the avoidance standard answer, which includes three observation points, the original relative motion line, the nearest encounter distance perpendicular line, the relative motion triangle construction line, the action point, the relative motion line after avoidance, and the construction line in the direction change or speed change scenario, and to generate a corresponding standard element mask for each standard geometric element. The union of all standard element masks is denoted as the standard geometric region mask. ; The mask matching and index calculation module is used to mask the answer ink marks. Masks with standard elements Perform matching and calculate element coverage. and geometric precision And calculate the coverage of key plotting steps; The sub-evaluation module is used to calculate the numerical error based on the answer data and the standard answer, and to integrate the numerical error and element coverage. Geometric accuracy And the coverage of key mapping steps, generate sub-scores according to the ten evaluation items of radar manual plotting; The total score correction and visualization feedback module corrects the total score according to the geometric consistency upper limit rule, and outputs the total score, the scores of each item, the reasons for deductions for each item, and a visualization feedback chart that overlays the student's ink marks with the standard geometric construction chain.

[0064] like Figure 11 As shown, the radar manual plotting automatic evaluation system in this embodiment includes functions such as answer data input, image and field preprocessing, standard answer and geometric drawing chain generation, ink matching and index calculation, evaluation decision and feedback output, and result storage and verification. Figure 11 Used to describe the system's execution Figure 1 The data flow and functional hierarchy in the evaluation method shown.

[0065] In one implementation, the system is implemented as a standalone desktop evaluation program, without embedding the calculation controls of the radar manual plotting training software or relying on the internal data interface of the training software. The desktop evaluation program includes a single evaluation page, a batch evaluation page, a question bank management page, a simulation sample page, and an experimental statistics page. The single evaluation page is used to import electronic plotting diagrams, read or load question data, manually verify fields, and perform evaluations. The batch evaluation page is used to perform batch evaluations on a directory of screenshots containing data files of questions with the same name and output evaluation result files. The question bank management page is used to browse the standard question bank, preview simulation diagrams, and load single evaluations. The simulation sample page is used to generate realistic template simulation screenshots and controllable error samples. The experimental statistics page is used to read batch evaluation result files and statistically analyze the number of samples, average score, score range, location overlap, and number of candidate line segments.

[0066] In one embodiment, the system further includes a batch evaluation module, a question bank management module, a simulation sample module, and an experimental statistics module, wherein: The batch evaluation module is used to read the question data file corresponding to the electronic drawing diagram, call the data input module, drawing area positioning and coordinate conversion module, ink extraction module and sub-item evaluation module to complete the batch evaluation, and output the batch evaluation details file and summary statistics file.

[0067] The question bank management module is used to read the preset standard question bank, display the question number, question data, standard answer and simulation plot, and load the selected questions into a single evaluation page.

[0068] The simulation sample module is used to generate standard simulation screenshots and controllable error samples with point offset, angle deviation, distance ratio deviation, omission and extra drawing based on blank templates of real radar manual plotting training software.

[0069] The experimental statistics module is used to read batch evaluation result files and count the number of evaluation samples, the number of successful evaluations, the average score, the highest score, the lowest score, the degree of location overlap, and the number of candidate line segments, providing data support for the verification of evaluation rules.

[0070] In summary, this invention can output the total score, sub-item scores, reasons for deductions, and overlay a visual feedback chart, enabling objective and interpretable automatic evaluation of radar manual plotting responses. While ensuring the reliability of the plotting area positioning, it achieves the objectivity and interpretability of the evaluation.

[0071] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A radar manual plotting automatic evaluation method based on structured geometric chain constraint drawing, characterized in that, include: S1. Obtain the radar manual plotting electronic drawing diagram, perform optical character recognition on the question table fields and answer fields in the electronic drawing diagram, obtain the pre-filled results of question data and answer data, and establish a data input chain for automatic evaluation of radar manual plotting. S2. Based on the electronic plotting diagram in the data input chain, combined with the radar image green grid positioning results and the auxiliary detection results of the target detection model, determine the radar plotting area range, and establish the conversion relationship between radar plane coordinates and pixel plane coordinates based on the plotting area boundary, the ship's origin position and radar range. S3, extracting the black plotting ink of the student in the plotting area to generate the answer ink mask and retaining the line, point and composition traces; S4. Based on the three observation data in the question, the ship's course, the ship's speed, and the rules of manual navigation plotting, calculate the target ship's motion parameters, the nearest encounter distance, the time to reach the nearest encounter point, and the standard answer for avoidance. S5. Generate a standard geometric construction chain based on the standard answer for obstacle avoidance, including three observation points, the original relative motion line, the nearest encounter distance perpendicular line, the relative motion triangle construction line, the action point, the relative motion line after avoidance, and the construction line in the direction change or speed change scenario, and generate corresponding standard element masks for each standard geometric element. The union of all standard element masks is denoted as the standard geometric region mask. ; S6, apply answer ink mask with each standard element mask perform matching, calculate element coverage and geometric accuracy and calculate key composition step coverage; S7. Calculate the numerical error based on the answer data and the standard answer, and integrate the numerical error and element coverage. Geometric accuracy And the coverage of key mapping steps, generate sub-scores according to the ten evaluation items of radar manual plotting; S8. Correct the total score according to the geometric consistency upper limit rule, and output the total score, the scores of each item, the reasons for the deduction of each item, and a visual feedback chart that overlays the student's ink marks with the standard geometric construction chain.

2. The automatic evaluation method for radar manual plotting based on structured geometric drawing chain constraints according to claim 1, characterized in that, Step S1 includes: S11. Receive electronic plotting diagrams from radar manual plotting training software; S12. In the independent desktop evaluation interface, perform optical character recognition on the question table fields of the electronic drawing diagram to obtain the ship's heading, ship's speed and the target ship's three observed bearings and distances, as the question data pre-filling results. S13. In the independent desktop evaluation interface, perform optical character recognition on the answer fields of the electronic drawing diagram to obtain the target ship's true course, true speed, nearest encounter distance, arrival time at the nearest encounter point, and avoidance maneuver parameters, as the answer data pre-fill result; S14. Display the pre-filled results of the question data and the pre-filled results of the answer data in the independent desktop assessment interface, and provide a manual verification and correction interface for assessors to verify and correct the fields. S15. Use the corrected question data and the corrected answer data as the data input chain.

3. The automatic evaluation method for radar manual plotting based on structured geometric drawing chain constraints according to claim 1, characterized in that, Step S2 includes: S21. Based on the concentric circles, azimuth scale, grid line color and geometric distribution characteristics of the radar map in the electronic drawing, detect green grids and determine candidate areas of the drawing area; S22. Use an object detection model to detect the electronic drawing and output the candidate bounding box of the drawing area and the candidate bounding box of the preset geometric elements. S23. The green grid positioning results are fused with the candidate bounding boxes output by the target detection model to determine the boundary of the radar plotting area, the origin position of the ship, and the effective plotting radius. S24. Based on the boundary of the drawing area, the location of the ship's origin, and the radar range read from the data in the question, establish the conversion relationship between the radar plane coordinates and the pixel plane coordinates with the ship's origin as the coordinate origin.

4. The automatic evaluation method for radar manual plotting based on structured geometric drawing chain constraints according to claim 1, characterized in that, Step S3 includes: S31. Extract the binary foreground of the black ink marks in the drawing area based on the color threshold, and distinguish the dark handwriting drawn by the students from the light background, green grid and question table. S32. Dilate the binary foreground to connect the tiny gaps caused by screenshot compression, thin handwriting, or broken lines. S33. Within the candidate bounding box of geometric elements output by the target detection model, perform a closing operation on the dilated binary foreground to enhance the continuity of points, line segments, and local composition traces. S34. Use Hough transform to detect line segments from the binary foreground, and merge the line segment detection results with the binary processing results to obtain the answer ink mask. .

5. The automatic evaluation method for radar manual plotting based on structured geometric drawing chain constraints according to claim 1, characterized in that, Step S4 includes: S41. Based on the three observation azimuths and distances in the question data, and combined with the established conversion relationship between radar plane coordinates and pixel plane coordinates, determine the positions of the target ship's first observation point A1, second observation point A2, and last observation point A3 in the pixel plane. S42. Determine the original relative motion line by connecting the first observation point A1 and the last observation point A3, and use the intermediate observation point A2 for the evaluation of the integrity and coverage of the observation points. S43. Use the second observation point A2 as an intermediate observation point for evaluation of the completeness and coverage of subsequent observation points; S44. Draw a perpendicular line from the origin of this ship to the original relative motion line, and obtain the nearest meeting distance DCPA and the time to reach the nearest meeting point TCPA based on the position of the foot of the perpendicular. S45. Construct a relative motion triangle based on the ship's motion vector and the original relative motion line, and calculate the target ship's true heading and true speed. S46. Determine the action point A4 based on the avoidance maneuver parameters, construct the relative motion line after avoidance, and construct a new motion triangle based on the relative motion line after avoidance and the motion vector of the ship itself. S47. Based on the new motion triangle, calculate the new course or speed of the ship after the avoidance maneuver.

6. The automatic evaluation method for radar manual plotting based on structured geometric drawing chain constraints according to claim 1, characterized in that, Step S5 includes: S51. Based on the calculated nearest encounter distance, time to reach the nearest encounter point, target ship's true course, target ship's true speed, action point, relative motion line after avoidance, and the ship's new course or speed after avoidance, generate a standard geometric drawing chain composed of several standard geometric elements in the order of drawing. Among them, the standard geometric elements include the first observation point A1, the second observation point A2, the last observation point A3, the original relative motion line determined by the first observation point A1 and the last observation point A3, the DCPA perpendicular line drawn from the ship's origin to the original relative motion line, the A1-M construction line and the M-A3 construction line in the relative motion triangle, action point A4, the relative motion line after avoidance starting from action point A4, and the Q construction line in the course change scenario or the R construction line in the speed change scenario. S52. Generate a corresponding standard element mask for each standard geometric element. The observation point generates a circular mask area with a preset radius, the line segment generates a strip mask area with a preset line width, the perpendicular line generates a strip mask area with a preset line width, and the construction line generates a strip mask area with a preset line width. S53, Mask all standard elements Perform a union operation to obtain a standard geometric region mask. .

7. The automatic evaluation method for radar manual plotting based on structured geometric drawing chain constraints according to claim 1, characterized in that, Step S6 includes: S61. Mask the ink marks of student answers. Masks with standard elements Perform pixel-level matching element by element, and calculate the element coverage for each standard geometric element. : In the above formula, This indicates the number of foreground pixels in the mask. Indicates the student's ink mask and the first The intersection of standard element masks; element coverage This is used to measure whether the standard geometric element has been completely drawn by the student. A low coverage rate indicates that the corresponding point, line segment or construction step has been omitted, missing or incomplete. S62. Calculate the ink mask used for answering questions. Relative to standard geometric region mask geometric precision : Geometric accuracy Used to measure the proportion of a student's ink marks that fall within the standard geometric area; when a student draws too many, draws randomly, or the line segments deviate significantly from the standard geometric area, the area falling within the standard geometric area is masked. In addition to increased ink marks, geometric precision It then declined.

8. The automatic evaluation method for radar manual plotting based on structured geometric drawing chain constraints according to claim 1, characterized in that, Step S7 includes: S71. Extract the answer values ​​for each numerical item from the answer data, extract the standard values ​​for the corresponding numerical items from the standard answers, and calculate the error for each numerical item. ; S72. Set full tolerance for each numerical item. and zero threshold ,when When the error score for this numerical item is... ;when When the error score for that numerical item is 0, the score is 0. When this value is calculated linearly, the scoring function is: ; S73. Integrate the error scores of each numerical item with the element coverage, geometric accuracy and key mapping step coverage. According to the ten evaluation items of radar manual plotting, match each evaluation item with the corresponding geometric elements and numerical quantities to generate sub-item scores.

9. The automatic evaluation method for radar manual plotting based on structured geometric drawing chain constraints according to claim 1, characterized in that, Step S8 includes: S81. Calculate the geometric consistency score based on the obtained element coverage, geometric accuracy, and key composition step coverage. S82. Set the total score correction upper limit according to the different levels of the geometric consistency score, where each level corresponds to a different total score upper limit value; S83. Determine the current level of geometric consistency score. If key geometric elements are missing, geometric accuracy is too low, or key composition steps are not covered enough, then the sum of the sub-scores generated in S7 is corrected according to the upper limit of the total score corresponding to this level to obtain the corrected total score. S84. Overlay the generated answer ink mask with the generated standard geometric drawing chain to generate a visual feedback diagram, in which the student ink and standard geometric elements are displayed in different visual styles. S85. Output the corrected total score, the score for each sub-item, the reason for deduction for each sub-item, and a visual feedback chart.

10. An automatic evaluation system for radar manual plotting based on structured geometric construction chain constraints, implemented based on the automatic evaluation method for radar manual plotting based on structured geometric construction chain constraints as described in any one of claims 1-9, characterized in that, include: The system includes a data input module, a drawing area positioning and coordinate conversion module, an ink extraction module, a standard answer calculation module, a standard geometric construction chain generation module, a mask matching and index calculation module, a sub-item evaluation module, and a total score correction and visualization feedback module. Among these modules: The data input module acquires the radar manual plotting electronic drawing, performs optical character recognition on the question table fields and answer fields in the electronic drawing, obtains the pre-filled results of the question data and answer data, and establishes a data input chain for automatic evaluation of radar manual plotting. The mapping area positioning and coordinate conversion module is used to determine the radar mapping area range based on the electronic mapping diagram in the data input chain, combined with the radar image green grid positioning result and the auxiliary detection result of the target detection model, and to establish the conversion relationship between radar plane coordinates and pixel plane coordinates based on the mapping area boundary, the ship's origin position and radar range. The ink extraction module is used to extract the student's black ink marks within the drawing area and generate an answer ink mark mask. And preserve the lines, points, and compositional traces; The standard answer calculation module is used to calculate the target ship's motion parameters, nearest encounter distance, arrival time at the nearest encounter point, and avoidance standard answer based on the three observation data in the question data, the ship's course, the ship's speed, and the rules of nautical manual plotting; The standard geometric construction chain generation module is used to generate a standard geometric construction chain based on the avoidance standard answer, which includes three observation points, the original relative motion line, the nearest encounter distance perpendicular line, the relative motion triangle construction line, the action point, the relative motion line after avoidance, and the construction line in the direction change or speed change scenario, and to generate a corresponding standard element mask for each standard geometric element. The union of all standard element masks is denoted as the standard geometric region mask. ; The mask matching and index calculation module is used to mask the answer ink marks. Masks with standard elements Perform matching and calculate element coverage. and geometric precision And calculate the coverage of key plotting steps; The sub-evaluation module is used to calculate the numerical error based on the answer data and the standard answer, and to integrate the numerical error and element coverage. Geometric accuracy And the coverage of key mapping steps, generate sub-scores according to the ten evaluation items of radar manual plotting; The total score correction and visualization feedback module corrects the total score according to the geometric consistency upper limit rule, and outputs the total score, the scores of each item, the reasons for deductions for each item, and a visualization feedback chart that overlays the student's ink marks with the standard geometric construction chain.

Images