Unmanned aerial vehicle aerial survey data error intelligent detection system

Abstract

The invention discloses an intelligent detection system for aerial survey data errors of an unmanned aerial vehicle, and relates to the technical field of picture error detection. A temporary storage unit is arranged for temporarily storing all pictures which are collected by a trusted data source and are not subjected to error detection; a pre-processing unit is set to periodically pre-process undetected pictures to obtain corresponding intra-group verification sequences, mapping sequences and inter-group verification sequences, and a security detection platform is set to perform integrity verification on all transmitted undetected pictures based on the received intra-group verification sequences, mapping sequences and component verification sequences. The waste of the detection efficiency and the resource utilization rate of a public platform caused by error detection of transmitted missing pictures is avoided, binary conversion and preprocessing rules are performed on the pictures, picture data can be verified from multiple dimensions, the integrity and the accuracy of the data in the detection process are ensured, and the detection efficiency and the resource utilization rate of the public platform are improved. And the accuracy of picture error detection is greatly improved.

Description

Technical Field

[0001] The present invention relates to the technical field of picture error detection, and specifically provides an intelligent detection system for errors in UAV aerial survey data. Background Art

[0002] With the rapid development of UAV technology, UAV aerial survey has been widely used in many fields. UAV aerial survey can quickly and efficiently obtain large-area geospatial information, and the data obtained contains a large amount of valuable information. Among them, picture data is an important component, and these picture data can be used for topographic mapping, urban planning, agricultural monitoring, disaster assessment and other aspects.

[0003] However, in the actual process of UAV aerial survey, due to various factors, aerial survey data, especially picture data, is prone to errors. For example, vibrations during UAV flight, weather changes (such as strong winds, rainfall, fog, etc.), sensor failures, and interference during data transmission may all cause problems such as blurring, distortion, missing parts of information, and splicing errors in the obtained pictures. If these incorrect data are used without being detected and corrected, it will seriously mislead subsequent analysis and decision-making based on these data, resulting in waste of resources, decision-making mistakes and other adverse consequences.

[0004] Currently, when detecting errors in picture data included in UAV aerial survey data, a public platform is often used. The public platform generally has powerful computing capabilities and rich detection algorithm resources, and can quickly and effectively analyze the uploaded pictures and quickly detect and identify errors such as blurring and distortion.

[0005] However, this method of relying on a public platform for picture error detection also has some problems. First, if the content of the uploaded picture is missing, it may cause deviations in the detection results. In addition, the missing content of the picture will also waste the verification resources of the public platform. When the public platform detects each uploaded picture, it needs to consume a certain amount of computing resources and time. If the picture content is missing, the investment of these resources may not get effective returns, reducing the detection efficiency and resource utilization rate of the public platform.

[0006] To solve the above problems, the present invention proposes a solution. Summary of the Invention

[0007] The purpose of the present invention is to provide an intelligent detection system for errors in UAV aerial survey data to solve the problems raised in the above background art.

[0008] The present invention provides an intelligent detection system for errors in UAV aerial survey data, including a trusted source end and a security detection platform:

[0009] A trusted source end for a trusted data source to provide pictures for error detection to a security detection platform. The trusted source end includes a temporary storage unit and a preprocessing unit;

[0010] The preprocessing unit periodically obtains a number of pictures to be detected stored in the temporary storage unit. The preprocessing unit obtains all the pictures to be detected stored in the temporary storage unit in the current detection period, and then performs binary conversion on each obtained picture to be detected, and calibrates the converted binary data as the binary data of the corresponding picture to be detected;

[0011] For the binary data of each picture to be detected, the preprocessing unit generates an in-group check sequence, a mapping sequence, and an inter-group check sequence of the picture to be detected according to a preset preprocessing rule, summarizes the in-group check sequences, mapping sequences, and inter-group check sequences of all pictures to be detected to generate a security picture set for the current detection period, and transmits it to the security detection platform;

[0012] A security detection platform for providing picture error detection services to a trusted information source.

[0013] Further, the preprocessing rules for generating an in-group check sequence, a mapping sequence, and an inter-group check sequence of a picture to be detected are as follows:

[0014] S11: Take every four characters in the binary data as a comparison array in order from left to right, and several comparison arrays can be obtained;

[0015] Then, according to the positions of the characters in each comparison array in the binary data, all the comparison arrays obtained from the binary data are sequentially marked as A1, A2... Aa in order from left to right, where a≥1;

[0016] S12: Remove duplicates from the comparison arrays A1, A2... Aa, re-calibrate all the obtained comparison arrays after duplicate removal as feature arrays, and sequentially mark all the group feature arrays as B1, B2... Bb in the order of the comparison arrays A1, A2... Aa, where 1≤b<a. During the marking process, each feature array is marked according to its first occurrence position in the comparison arrays A1, A2... Aa;

[0017] S13: Calculate and obtain the mapping comparison amounts of all the feature arrays B1, B2... Bb according to a preset calculation rule, sort all the group feature arrays according to the numerical magnitudes of the mapping comparison amounts and re-mark them to obtain feature arrays D1, D2... Db;

[0018] S14: Map the numbers 0, 1, 00, 01, 10, 11, 000, 001, 010, 011, 100, 101, 110, 111 to the mapping queues of the feature arrays D1, D2... Db-2 in the order of 0, 1, 00, 01, 10, 11, 000, 001, 010, 011, 100, 101, 110, 111;

[0019] For the feature arrays Db-1 and Db: The mapping queue of the feature array Db-1 is itself, which is Db-1; the mapping queue of the feature array Db is itself, which is Db;

[0020] S15: Determine the total number of characters in the mapping queues that make up the feature arrays D1, D2... Db in sequence according to the preset determination acquisition rule, and obtain the corresponding in-group verification queue based on the determination result;

[0021] S16: Obtain the classification lists F1, F2... Ff according to the classification rule based on the comparison arrays A1, A2... Aa, where 1 ≤ f ≤ a;

[0022] S17: Calculate and obtain the verification feature value K1 of the classification list F1 according to the preset calculation rule;

[0023] S18: Obtain the verification feature values K2, K3... Kf of the classification lists F2, F3... Ff in sequence according to S17;

[0024] The preprocessing unit splices the corresponding in-group verification queues according to the order of the comparison arrays A1, A2... Aa to obtain the in-group verification sequence of the to-be-detected image; splices the corresponding mapping queues according to the order of the comparison arrays A1, A2... Aa to obtain the mapping sequence of the to-be-detected image; splices the corresponding verification feature values according to the order of the classification lists F1, F2,..., Ff to obtain the between-group verification sequence of the to-be-detected image.

[0025] Furthermore, for S13, the calculation rule for calculating and obtaining the mapping comparison amounts of all feature arrays B1, B2... Bb is as follows:

[0026] S131: Traverse the comparison arrays A1, A2... Aa, obtain the total number of comparison arrays that are consistent with the feature array B1, and label the total number as the mapping comparison amount of the feature array B1;

[0027] S132: Calculate and obtain the mapping comparison amounts of the feature arrays B2, B3... Bb in sequence according to S131, and relabel all the feature arrays as D1, D2... Db in the order from largest to smallest mapping comparison amount.

[0028] Further, S17, the calculation rule for calculating the verification characteristic value K1 classified into the list F1 is as follows:

[0029] S172: Generate a check matrix H1 assigned to the list F1 according to the comparison arrays G1, G2...Gg, wherein the check matrix H1 is a matrix with g rows and 4 columns, and the 1st, 2nd...g rows in the check matrix H1 correspond to the comparison arrays G1, G2...Gg from top to bottom, respectively;

[0030] From left to right, the 4 characters in the first row of the check matrix H1 are the 4 characters that constitute the comparison array G1 from left to right, and the 4 characters in the second, third, ...g rows of the check matrix H1 are analogous to this;

[0031] S173: Calculate and obtain the eigenvalue J1 of the check matrix H1 using the equation |H1-I 1×J1|=0, where I1 is the identity matrix of the check matrix H1, the number of rows and columns in the identity matrix I1 is the same as the number of rows and columns in the check matrix H1, and the main diagonal elements in the identity matrix I1 are all 1, and the other elements are all 0;

[0032] The absolute value of the characteristic value J1 is obtained, and the absolute value is rounded up to obtain the verification characteristic value K1 assigned to the list F1.

[0033] Compared with the prior art, it has the following beneficial effects:

[0034] (1) The present invention sets a temporary storage unit to temporarily store all pictures collected by the trusted data source that have not been error-checked, sets a pre-processing unit to periodically pre-process the undetected pictures to obtain corresponding intra-group verification sequences, mapping sequences and inter-group verification sequences, sets a security detection platform to perform integrity verification on all transmitted undetected pictures based on the received intra-group verification sequence, mapping sequence and component verification sequence, and performs error detection on pictures that pass the verification, thereby avoiding the waste of detection efficiency and resource utilization of the public platform caused by error detection of transmitted missing pictures;

[0035] (2) The present invention generates intra-group verification sequences, mapping sequences and inter-group verification sequences by performing binary conversion on the image and applying complex pre-processing rules. It can verify the image data from multiple dimensions, ensure the integrity and accuracy of the data during the detection process, and greatly improve the accuracy of image error detection. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] Figure 1 It is a system block diagram of the present invention. DETAILED DESCRIPTION

[0037] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0038] Please refer to Figure 1 , this application provides an intelligent detection system for UAV aerial survey data errors, including a trusted source end and a security detection platform;

[0039] The trusted source end is used to provide a set of pictures for error detection to the security detection platform from a trusted data source, where "trusted" refers to enterprise users, individual users, websites, etc. that have passed the identity and credit authentication of the security detection platform and are authorized to apply for picture error detection services;

[0040] The trusted source end includes a temporary storage unit, a preprocessing unit, and an atlas storage unit. The temporary storage unit temporarily stores a number of pictures to be detected that have been collected by a trusted data source and have not been subjected to error detection. In this application, the pictures to be detected refer to pictures obtained by UAV aerial survey;

[0041] The preprocessing unit periodically obtains a number of pictures to be detected stored in the temporary storage unit. The preprocessing unit obtains all the pictures to be detected stored in the temporary storage unit in the current detection period, and then performs binary conversion on each obtained picture to be detected, and calibrates the converted binary data as the binary data of the corresponding picture to be detected;

[0042] For the binary data of each picture to be detected, the preprocessing unit generates an in-group check sequence, a mapping sequence, and an inter-group check sequence for the picture to be detected according to a preset preprocessing rule. The preprocessing rule is as follows:

[0043] S11: Take every four characters in the binary data as a comparison array in the order from left to right, and several comparison arrays can be obtained;

[0044] Then, according to the positions of the characters in each comparison array in the binary data, all the comparison arrays obtained from the binary data are sequentially marked as A1, A2...Aa in the order from left to right, where a≥1;

[0045] S12: Remove duplicates from the comparison arrays A1, A2... Aa. Re-label all the comparison arrays obtained after duplicate removal as feature arrays, and sequentially label all the groups of feature arrays as B1, B2... Bb in the order of the comparison arrays A1, A2... Aa, where 1 ≤ b < a. During the labeling process, each feature array is labeled according to its position of first appearance in the comparison arrays A1, A2... Aa.

[0046] S13: Calculate and obtain the mapping comparison amounts of all the feature arrays B1, B2... Bb according to a preset calculation rule. Sort and re-label all the groups of feature arrays according to the numerical magnitudes of the mapping comparison amounts to obtain the feature arrays D1, D2... Db. The calculation rule is as follows:

[0047] S131: Traverse the comparison arrays A1, A2... Aa to obtain the total number of comparison arrays that are consistent with the feature array B1, and label the total number as the mapping comparison amount of the feature array B1.

[0048] S132: Sequentially calculate and obtain the mapping comparison amounts of the feature arrays B2, B3... Bb according to S131, and re-label all the feature arrays as D1, D2... Db in the order from largest to smallest of the mapping comparison amounts.

[0049] S14: Sequentially use 0, 1, 00, 01, 10, 11, 000, 001, 010, 011, 100, 101, 110, 111 as the mapping queues of the feature arrays D1, D2... Db-2 in the order of the string 0, 1, 00, 01, 10, 11, 000, 001, 010, 011, 100, 101, 110, 111.

[0050] For the feature arrays Db-1 and Db: The mapping queue of the feature array Db-1 is itself, which is Db-1; the mapping queue of the feature array Db is itself, which is Db.

[0051] S15: Sequentially determine the total number of characters in the mapping queues that make up the feature arrays D1, D2... Db according to a preset determination rule, and obtain the corresponding in-group verification queues based on the determination results. The determination rule is as follows:

[0052] S151: Determine the total number of characters E1 in the mapping queue that makes up the feature array D1. If E1 = 1, then use the string 0 as the in-group verification queue of the feature array D1. If E1 = 2, then use the string 1 as the in-group verification queue of the feature array D1. If E1 = 3, then use the string 10 as the in-group verification queue of the feature array D1. If E1 = 4, then use the string 11 as the in-group verification queue of the feature array D1.

[0053] S152: Obtain the in-group verification queues of the feature arrays D2, D3... Db in sequence according to S151;

[0054] S16: According to the preset classification rules, classify and obtain f classification lists F1, F2... Ff from the comparison arrays A1, A2... Aa, where 1 ≤ f ≤ a, and the grouping rules are as follows:

[0055] S161: Take the comparison array A1 as the classification identifier. According to the order of the comparison arrays A1, A2... Aa, extract all the comparison arrays with subscripts less than or equal to d, and count the number of comparison arrays that are consistent with the feature arrays D1, D2... Db in sequence among all the extracted comparison arrays, and mark them as E1, E2... Eb respectively, where d is the preset standard extraction quantity threshold;

[0056] S162: Use the formula F1 = E1 + E2 +... + Eb to calculate and obtain the redundancy evaluation quantity F1 of the classification identifier. The redundancy evaluation quantity is artificially defined and used to evaluate the complexity quantity of the types of the extracted comparison arrays;

[0057] S163: If F1 ≤ F, add all the extracted comparison arrays to an empty list to obtain a classification list. In the classification list, all the comparison arrays are arranged from left to right in the order of A1, A2... Aa;

[0058] S164: If F1 > F, according to the order of the comparison arrays A1, A2... Aa, extract all the comparison arrays with subscripts less than or equal to d - 1, d - 2... 3 in sequence. Each time an extraction is made, calculate the redundancy evaluation quantity F1 of the classification identifier according to the extracted comparison arrays in step S162. Each time the redundancy evaluation quantity F1 is calculated, compare the magnitudes of F1 and F until F1 ≤ F, and obtain the classification list at this time;

[0059] S165: Obtain the maximum subscript e of the marks of all the comparison arrays in the classification list in S164, and take the comparison array Ae as the classification identifier, and obtain the second classification list according to the steps from S161 to S164;

[0060] S166: According to the steps from S161 to S165, several classification lists can be obtained from the comparison arrays A1, A2... Aa, and all the classification lists are marked as F1, F2... Ff in sequence according to the order in which the classification lists are obtained, where 1 ≤ f ≤ a;

[0061] S17: Calculate and obtain the verification feature value K1 of the classification list F1 according to the preset calculation rules. The calculation rules are as follows:

[0062] S171: Extract all comparison arrays included in the classification list F1 in the order from left to right, and re-label all the extracted comparison arrays as G1, G2... Gg in the order of extraction, where g ≥ 1;

[0063] S172: Generate a parity-check matrix H1 of the classification list F1 according to the comparison arrays G1, G2... Gg. The parity-check matrix H1 is a matrix with g rows and 4 columns. From top to bottom, the 1st, 2nd... gth rows in the parity-check matrix H1 correspond to the comparison arrays G1, G2... Gg respectively;

[0064] From left to right, the 4 characters in the 1st row of the parity-check matrix H1 are respectively the 4 characters that form the comparison array G1 from left to right, and the 4 characters in the 2nd, 3rd... gth rows of the parity-check matrix H1 are the same by analogy;

[0065] S173: Use the equation |H1 - I1×J1| = 0 to calculate and obtain the eigenvalue J1 of the parity-check matrix H1. In the equation, I1 is the identity matrix of the parity-check matrix H1. The number of rows and columns in the identity matrix I1 is the same as the number of rows and columns in the parity-check matrix H1, and the main diagonal elements in the identity matrix I1 are all 1, and other elements are all 0;

[0066] Obtain the absolute value of the eigenvalue J1, and round up the absolute value to obtain the parity-check eigenvalue K1 of the classification list F1;

[0067] S18: Obtain the parity-check eigenvalues K2, K3... Kf of the classification lists F2, F3... Ff in sequence according to S17;

[0068] The preprocessing unit splices the corresponding in-group parity-check queues in the order of the comparison arrays A1, A2... Aa to obtain the in-group parity-check sequence of the to-be-detected picture;

[0069] Splice the corresponding mapping queues in the order of the comparison arrays A1, A2... Aa to obtain the mapping sequence of the to-be-detected picture;

[0070] Splice the corresponding parity-check eigenvalues in the order of the classification lists F1, F2,... Ff to obtain the inter-group parity-check sequence of the to-be-detected picture;

[0071] After the preprocessing unit generates the in-group parity-check sequence, mapping sequence and inter-group parity-check sequence of all the to-be-detected pictures stored in the temporary storage unit in the current detection period, it generates the security picture set of the current detection period according to them and the feature arrays D1, D2... Db, and transmits the security picture set to the security detection platform;

[0072] The described security detection platform is used to provide picture error detection services to trusted information sources. After receiving the security picture set of the current detection cycle transmitted, the security detection platform obtains the in-group verification sequence, mapping sequence, component verification sequence, and feature arrays D1, D2... Db of all the pictures to be detected included therein;

[0073] For a picture to be detected obtained, in combination with the feature arrays D1, D2... Db, first restore the mapping sequence of the picture to be detected to obtain corresponding comparison arrays, and then calculate and obtain the in-group verification sequence and component verification sequence of the picture to be detected based on the obtained comparison data;

[0074] Perform a consistency comparison between the calculated in-group verification sequence and inter-group verification sequence of the picture to be detected and the received in-group verification sequence and inter-group verification sequence of the picture to be detected. If the comparisons are all consistent, the verification passes;

[0075] After the verification passes, use a preset error detection algorithm to perform error detection on it. If an error is detected, generate error detection information of the picture to be detected according to the error type and the picture to be detected;

[0076] In this application, the error detection algorithm can be any one of the Laplace operator, Sobel operator, Brenner gradient function, and Tenengrad gradient function. These algorithms have the advantages of fast detection speed and high accuracy, and can effectively identify errors such as blurring and distortion of pictures;

[0077] After the detection of the picture to be detected is completed, transmit the generated error detection information of the pictures to be detected to the trusted source end. The trusted source end feeds back these pictures to be detected to the management personnel, and deletes these pictures from the temporary storage unit, and transmits all the pictures to be detected that have not received error detection information to the picture set storage unit for storage;

[0078] Some of the data in the above formulas are all numerically calculated by removing their dimensions, and the content not described in detail in this specification belongs to the prior art well-known to those skilled in the art.

[0079] The above embodiments are only used to illustrate the technical method of the present invention and not to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those of ordinary skill in the art should understand that the technical method of the present invention can be modified or equivalently replaced without departing from the spirit and scope of the technical method of the present invention.

Claims

1. An intelligent detection system for UAV aerial survey data errors, characterized in that, It includes a trusted source end and a security detection platform: The trusted source end is used for the trusted data source to provide pictures for error detection to the security detection platform. The trusted source end includes a temporary storage unit and a preprocessing unit; The preprocessing unit periodically obtains a number of pictures to be detected stored in the temporary storage unit. The preprocessing unit obtains all the pictures to be detected stored in the temporary storage unit in the current detection cycle, and then performs binary conversion on each obtained picture to be detected, and calibrates the converted binary data as the binary data of the corresponding picture to be detected; For the binary data of each picture to be detected, the preprocessing unit generates an in-group check sequence, a mapping sequence, and an inter-group check sequence of the picture to be detected according to a preset preprocessing rule, summarizes the in-group check sequences, mapping sequences, and inter-group check sequences of all pictures to be detected to generate a security picture set in the current detection cycle, and transmits it to the security detection platform; The security detection platform is used to provide picture error detection services to the trusted information source.

2. An intelligent detection system for UAV aerial survey data errors according to claim 1, characterized in that, The preprocessing rules for generating an in-group check sequence, a mapping sequence, and an inter-group check sequence of a picture to be detected are as follows: S11: Take every four characters in the binary data as a comparison array in order from left to right, and several comparison arrays can be obtained; Then, according to the positions of the characters in each comparison array in the binary data, all the comparison arrays obtained from the binary data are sequentially marked as A1, A2... Aa in order from left to right, where a≥1; S12: Remove duplicates from the comparison arrays A1, A2... Aa, re-calibrate all the comparison arrays obtained after deduplication as feature arrays, and sequentially mark all the group feature arrays as B1, B2... Bb in the order of the comparison arrays A1, A2... Aa, where 1≤b<a. During the marking process, each feature array is marked according to its first appearance position in the comparison arrays A1, A2... Aa; S13: Calculate and obtain the mapping comparison amounts of all the feature arrays B1, B2... Bb according to a preset calculation rule, sort all the group feature arrays according to the numerical sizes of the mapping comparison amounts and re-mark them to obtain the feature arrays D1, D2... Db; S14: Take 0, 1, 00, 01, 10, 11, 000, 001, 010, 011, 100, 101, 110, 111 as the mapping queues of the feature arrays D1, D2... Db-2 in the order of the strings 0, 1, 00, 01, 10, 11, 000, 001, 010, 011, 100, 101, 110, 111; For the feature arrays Db-1 and Db: The mapping queue of the feature array Db-1 is itself, which is Db-1; the mapping queue of the feature array Db is itself, which is Db; S15: Sequentially determine the total number of characters forming the mapping queues of the feature arrays D1, D2... Db according to a preset determination rule, and obtain the corresponding in-group check queue based on the determination result; S16: According to the preset classification rules, classify and obtain classification lists F1, F2... Ff based on comparison arrays A1, A2... Aa, where 1 ≤ f ≤ a; S17: Calculate and obtain the verification characteristic value K1 of classification list F1 according to the preset calculation rules; S18: Obtain the verification characteristic values K2, K3... Kf of classification lists F2, F3... Ff in sequence according to S17; The preprocessing unit splices the corresponding in-group verification queues in the order of comparison arrays A1, A2... Aa to obtain the in-group verification sequence of the to-be-detected image; splices the corresponding mapping queues in the order of comparison arrays A1, A2... Aa to obtain the mapping sequence of the to-be-detected image; splices the corresponding verification characteristic values in the order of classification lists F1, F2,... Ff to obtain the between-group verification sequence of the to-be-detected image.

3. An intelligent detection system for UAV aerial survey data errors according to claim 2, characterized in that, S13. The calculation rules for calculating and obtaining the mapping comparison amounts of all feature arrays B1, B2... Bb are as follows: S131: Traverse comparison arrays A1, A2... Aa, obtain the total number of comparison arrays that are consistent with feature array B1 among them, and label the total number as the mapping comparison amount of feature array B1; S132: Calculate and obtain the mapping comparison amounts of feature arrays B2, B3... Bb in sequence according to S131, and relabel all feature arrays as D1, D2... Db in the order from largest to smallest mapping comparison amount.

4. An intelligent detection system for UAV aerial survey data errors according to claim 2, characterized in that S15. Determine the total number of characters in the mapping queues that make up feature arrays D1, D2... Db, and obtain the corresponding determination acquisition rules for in-group verification queues based on the determination results as follows: S151: Determine the total number of characters E1 in the mapping queue that makes up feature array D1. If E1 = 1, use the string 0 as the in-group verification queue of feature array D1. If E1 = 2, use the string 1 as the in-group verification queue of feature array D1. If E1 = 3, use the string 10 as the in-group verification queue of feature array D1. If E1 = 4, use the string 11 as the in-group verification queue of feature array D1; S152: Obtain the in-group verification queues of feature arrays D2, D3... Db in sequence according to S151.

5. An intelligent detection system for UAV aerial survey data errors according to claim 2, characterized in that, S16. The classification rules for obtaining f classification lists F1, F2... Ff are as follows: S161: Use comparison array A1 as the classification identifier. According to the order of comparison arrays A1, A2... Aa, extract all comparison arrays with subscripts less than or equal to d, and count the number of comparison arrays that are consistent with feature arrays D1, D2... Db in the extracted comparison arrays in sequence and label them as E1, E2... Eb, where d is a preset standard extraction quantity threshold; S162: Calculate and obtain the redundancy evaluation quantity F1 of the classification identifier using the formula F1 = E1 + E2 +... + Eb. The redundancy evaluation quantity is defined manually and is used to evaluate the complexity of the types of the extracted comparison arrays; S163: if F1≤F, then add all the extracted comparison arrays into an empty list to obtain a classification list, in which all the comparison arrays are arranged from left to right in the order of A1, A2...Aa; S164: If F1>F, then all comparison arrays with subscripts less than or equal to d-1, d-2...3 are extracted in sequence according to the comparison arrays A1, A2...Aa. Each time the comparison arrays are extracted, the redundant evaluation amount F1 of the classification identifier is calculated according to all the extracted comparison arrays according to step S162. Each time the redundant evaluation amount F1 is calculated, F1 is compared with F until F1≤F, and the classification list at this time is obtained. S165: Obtain the maximum subscript e of the marks of all comparison arrays in the classification list in S164, and use the comparison array Ae as the classification identifier, and obtain the second classification list according to the steps from S161 to S164; S166: According to steps S161 to S165, a plurality of classification lists can be obtained by comparing arrays A1, A2...Aa, and all classification lists are marked as F1, F2...Ff in sequence according to the order in which the classification lists are obtained.

6. An intelligent detection system for UAV aerial survey data errors according to claim 2, characterized in that, S17, the calculation rule for obtaining the verification characteristic value K1 classified into the list F1 is as follows: S172: Generate a check matrix H1 assigned to the list F1 according to the comparison arrays G1, G2...Gg, wherein the check matrix H1 is a matrix with g rows and 4 columns, and the 1st, 2nd...g rows in the check matrix H1 correspond to the comparison arrays G1, G2...Gg from top to bottom, respectively; From left to right, the 4 characters in the first row of the check matrix H1 are the 4 characters that constitute the comparison array G1 from left to right, and the 4 characters in the second, third, ...g rows of the check matrix H1 are analogous to this; S173: Calculate and obtain the eigenvalue J1 of the check matrix H1 using the equation |H1-I 1×J1|=0, where I1 is the identity matrix of the check matrix H1, the number of rows and columns in the identity matrix I1 is the same as the number of rows and columns in the check matrix H1, and the main diagonal elements in the identity matrix I1 are all 1, and the other elements are all 0; The absolute value of the characteristic value J1 is obtained, and the absolute value is rounded up to obtain the verification characteristic value K1 assigned to the list F1.

7. An intelligent detection system for UAV aerial survey data errors according to claim 2, characterized in that, The generated safety atlas of the current detection cycle also includes feature arrays D1, D2, ..., Db.

8. An intelligent detection system for UAV aerial survey data errors according to claim 1, characterized in that, After receiving the transmitted security atlas of the current detection cycle, the security detection platform obtains the intra-group verification sequence, mapping sequence, component verification sequence and feature arrays D1, D2...Db of all the images to be detected contained therein; For an acquired image to be detected, the mapping sequence of the image to be detected is first restored in combination with the feature arrays D1, D2...Db to obtain a number of corresponding comparison arrays, and then based on the restored comparison data, the intra-group check sequence and component check sequence of the image to be detected are calculated and obtained; Perform a consistency comparison between the calculated intra-group check sequence and inter-group check sequence of the image to be detected and the received intra-group check sequence and inter-group check sequence of the image to be detected. If the comparisons are all consistent, the verification passes; otherwise, the verification fails.

Images