UAV mapping system based on deep learning

Abstract

The present invention relates to the field of surveying and mapping technology, and specifically to a deep learning-based unmanned aerial vehicle (UAV) surveying and mapping system; the system comprises: an UAV surveying and mapping terminal, wherein the UAV surveying and mapping terminal has a built-in demand analysis module and a surveying and mapping optimization module; by real-time monitoring and comprehensive analysis of the data volume and transmission performance in the UAV surveying and mapping process, optimization instructions can be automatically generated, thereby improving the stability of data acquisition and the performance of network transmission, and providing effective support for subsequent surveying and mapping optimization; by reasonable priority sorting and transmission strategy, the efficiency and reliability of data transmission are improved, and the utilization of system resources is optimized; by quantitatively analyzing the degree of repeated surveying and mapping of each section in the flight path to determine the optimal degree of repeated surveying and mapping of each section, and optimizing each section accordingly; the flight path of the UAV is dynamically adjusted to avoid repeated surveying and mapping to the greatest extent, reduce repeated data transmission, and improve the effective collection and effective transmission of surveying and mapping data.

Description

Technical Field

[0001] The present invention relates to the field of surveying and mapping technology, and in particular to an unmanned aerial vehicle surveying and mapping system based on deep learning. Background Art

[0002] Drone mapping is the use of unmanned aircraft (drones) to measure and record the surface or objects. It combines advanced technologies such as aerial photography and lidar to efficiently and accurately obtain large-scale geographic data and image information. Map making, disaster assessment and rescue, infrastructure construction and environmental monitoring all rely on drone mapping technology, so drone mapping is particularly important.

[0003] The amount of data collected during current drone mapping is huge, especially high-resolution images and lidar point clouds, which will occupy a large amount of bandwidth during the transmission process, causing data delays or packet loss, or even failure to transmit all data in a timely manner. Summary of the invention

[0004] The purpose of the present invention is to provide a UAV surveying and mapping system based on deep learning to solve the problems mentioned in the above background technology.

[0005] The purpose of the present invention can be achieved through the following technical solutions: A UAV surveying and mapping system based on deep learning, comprising: a UAV surveying and mapping terminal; the UAV surveying and mapping terminal comprises: a demand analysis module and a surveying and mapping optimization module;

[0006] The demand analysis module monitors and analyzes the amount of data collected during the UAV mapping process and the transmission performance during the data transmission process to obtain the data evaluation index and performance evaluation index, which are recorded as SG and ZX respectively; SG and ZX are substituted into the set formula Calculate and obtain the optimization requirement value Yd, where d1 and d2 are respectively the set proportional coefficients; generate optimization instructions according to the optimization requirement value and send them to the surveying and mapping optimization module;

[0007] The surveying and mapping optimization module optimizes the data transmission and the UAV flight path during the surveying and mapping process based on the received optimization instructions. The specific steps of flight path optimization are:

[0008] Step 1: Obtain the survey area map, flight path, and current drone position of this surveying mission; divide the survey area into several blocks according to the coverage boundary of the drone surveying, and mark the blocks that have been surveyed and the blocks that have not been surveyed on the survey area map;

[0009] Step 2: Import the flight path, current drone position, and survey area map into the GIS system, divide the flight path into several sections, and perform spatial overlay analysis on each section with the blocks in the survey area map to obtain the overlapping blocks of each section and the overlapping area of ​​each overlapping block;

[0010] Step 3: Count the number of overlapping blocks in each road section, and sum up the overlapping areas of each overlapping block to get the total overlapping area; calculate the area covered when mapping each road section; substitute the number of overlapping blocks M1, the total overlapping area M2 and the area covered when mapping the road section M3 into the set formula Calculate and obtain the overlap value Mg of each road section, where g1 and g2 are respectively the set proportional coefficients; when the overlap value is greater than or equal to the set overlap threshold, the road section is recorded as a repeated road section;

[0011] Step 4: Select the flight start point and flight end point of the repeated section, and use the genetic algorithm to generate several alternative sections, thereby obtaining the alternative sections of each repeated section; select any one of the repeated sections, repeat steps 2 to 3 for its corresponding alternative section to obtain the overlap value of each alternative section, when the overlap value of each alternative section is less than the set overlap threshold, select the alternative section with the smallest overlap value as the target section, and replace the corresponding repeated section with the target section; when the overlap value of each alternative section is greater than or equal to the set overlap threshold, obtain the length of the repeated section and each alternative section, and perform weighted calculation with the corresponding overlap value to obtain the optimized value, select the section with the smallest optimized value as the target section, and replace the corresponding repeated section with the target section;

[0012] Step 5: The optimized path after replacement can be obtained, and the UAV is controlled to perform flight mapping according to its optimized path.

[0013] Preferably, the specific process of data transmission optimization is as follows:

[0014] Preferably, the specific process of data transmission optimization is as follows:

[0015] Retrieve the data evaluation index of each type of data, set each type of data to correspond to a type value, and compare the data category with all the set data categories to match the corresponding type value;

[0016] The data evaluation index and type value corresponding to each type of data are taken as the numerical value, and the numerical value is multiplied to obtain the priority index of each type of data; each type of data is arranged in descending order according to its corresponding priority index to obtain a data transmission list, and data transmission is performed according to the data transmission list.

[0017] Preferably, the specific process of monitoring and analyzing the amount of data collected during the drone mapping process is as follows:

[0018] 301: Classify the data collected by the UAV surveying and mapping, where any category is denoted as k, k=1,2,3...K, K is a positive integer, and K represents the total category; obtain the data volume of each type of data collected by each UAV at each collection time and denoted as Lkj, where j=1,2,3...J, J is a positive integer, J represents the total number of collection times, and j represents the sequence number of any collection time;

[0019] 302: Take any one type of data, take time as the horizontal axis and data volume as the vertical axis to obtain a curve graph of data volume changing with time; draw a tangent line at the data point, obtain a tangent line expression by data fitting, and calculate the derivative of the tangent line expression to obtain the data derivative, which is recorded as Dkj; sum the data derivatives greater than zero to obtain the quantity increase value, which is recorded as S1k, and sum the data derivatives less than zero and take the absolute value to obtain the data volume decrease value, which is recorded as S2k;

[0020] 303: Record the data derivative equal to zero as a stable derivative, extract the data amount of the data point corresponding to the stable derivative, and record it as the stable value of the stable derivative, thereby obtaining the stable value of each stable derivative, and performing quantitative analysis on it to obtain a stable level value recorded as Gak;

[0021] 304: Substitute the data derivative Dkj, the quantity increase value S1k, the data quantity decrease value S2k, the stable level value Gak and the data quantity Lkj into the set formula The data evaluation index SG corresponding to each type of data is calculated, where b1, b2, b3, b4, and b5 are the set proportional coefficients respectively. is the mean value of the data derivative of each data point; the data evaluation index corresponding to each type of data is summed up to obtain the data evaluation index of UAV mapping.

[0022] Preferably, the specific process of quantitatively analyzing the stable value of each stable derivative is as follows:

[0023] Compare and analyze the stable value with the set stable interval to divide the data points corresponding to the stable value into high level points, middle level points and low level points, and count the cumulative number of high level points, middle level points and low level points respectively, and record them as G1k, G2k and G3k respectively; sum the data corresponding to the high level points, middle level points and low level points to obtain the high level stable value, middle level stable value and low level stable value, and record them as G4k, G5k and G6k respectively;

[0024] Substitute the cumulative number of high-level points G1k, the cumulative number of mid-level points G2k, the cumulative number of low-level points G3k, the high-level stability value G4k, the mid-level stability value G5k, and the low-level stability value G6k into the setting formula The stable level value Gak is calculated, wherein a1, a2, and a3 are proportional coefficients set respectively, and a1>a2>a3>0.

[0025] Preferably, the specific process of monitoring and analyzing the transmission performance during data transmission is as follows:

[0026] 501: Obtain data transmission information at each acquisition time, where the data transmission information includes network bandwidth (network bandwidth here refers to the maximum bandwidth of the network connection, indicating the actual bandwidth capacity of the network connection) and delay (delay here is a value indicating the length of time, which indicates the round-trip time of the Ping command, i.e., RRT), and record them as Wj and Pj respectively; substitute Wj and Pj into the set formula The transmission performance value WPj is calculated, where c1 and c2 are the set proportional coefficients respectively;

[0027] 502: construct a two-dimensional rectangular coordinate system with time as the horizontal coordinate and transmission performance value as the vertical coordinate, plot the transmission performance value on the coordinate axis according to its corresponding acquisition time to obtain the performance point; use a smooth curve to connect the performance points in sequence to obtain a curve graph of the transmission performance value changing with time; draw a tangent line at the performance point, use data fitting to obtain a tangent expression, and perform a derivative calculation on the tangent expression to obtain a performance derivative, recorded as Xj; sum the performance derivatives greater than zero to obtain a performance value-added, recorded as Z1, and sum the performance derivatives less than zero and take the absolute value to obtain a performance value-added, recorded as Z2;

[0028] 503: Substitute the performance derivative Xj, the performance increase value Z1, the performance decrease value Z2 and the transmission performance value WPj into the set formula The performance evaluation index ZX is calculated, where c3, c4, and c5 are the set proportional coefficients. is the mean of the performance derivatives of each performance point.

[0029] Beneficial effects of the present invention:

[0030] 1. Through real-time monitoring and comprehensive analysis of the data volume and transmission performance during the UAV mapping process, optimization instructions can be automatically generated to improve the stability of data collection and the performance of network transmission; the automated processing method reduces the need for manual intervention, improves the intelligence level of the system, and provides effective support for subsequent mapping optimization;

[0031] 2. Through reasonable priority sorting and transmission strategies, not only the efficiency and reliability of data transmission are improved, but also the timely processing of key data is guaranteed, the utilization of system resources is optimized, and the overall task execution effect is improved;

[0032] 3. Quantitatively analyze the degree of repeated mapping of each section in the flight path to determine the degree of repeated mapping of each section, and optimize each section accordingly; dynamically adjust the flight path of the drone to avoid repeated mapping to the greatest extent, reduce repeated data transmission, and improve the effective collection and transmission of mapping data. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] The present invention will be further described below in conjunction with the accompanying drawings.

[0034] Figure 1 It is a schematic diagram of system module connection of the present invention. DETAILED DESCRIPTION

[0035] It should be noted that, in the absence of conflict, the embodiments and features in the embodiments of the present application can be combined with each other. The present application will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.

[0036] In order to enable those skilled in the art to better understand the solution of the present application, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present application.

[0037] It should be noted that the terms "first", "second", etc. in the specification and claims of the present application and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequential order. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments of the present application described here. In addition, the terms "including" and "having" and any of their variations are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device comprising a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0038] See also Figure 1 As shown, the present invention is a UAV surveying and mapping system based on deep learning, comprising: a UAV surveying and mapping terminal, wherein the UAV surveying and mapping terminal has a built-in demand analysis module and a surveying and mapping optimization module;

[0039] The demand analysis module conducts a comprehensive analysis based on the amount of data collected by the UAV surveying and the data transmission information to determine whether surveying and mapping optimization is needed, and generates optimization instructions accordingly and sends them to the surveying and mapping optimization module; specifically:

[0040] Data volume analysis:

[0041] The data collected by the drone mapping is classified, and any category is denoted as k, k=1,2,3...K, K is a positive integer, and K represents the total category; for example, the data categories include image data (including resolution, color depth and other information), sensor data (GPS data, temperature and humidity sensor data, spectral data, etc.), video data and lidar data, etc.;

[0042] The amount of data collected by each drone at each collection time is recorded as Lkj, where j=1,2,3...J, J is a positive integer, J represents the total number of collection times, and j represents the sequence number of any collection time; take any one type of data, use time as the horizontal coordinate and data volume as the vertical coordinate to construct a two-dimensional rectangular coordinate system, plot the data volume on the coordinate axis according to its corresponding collection time to obtain data points, use smooth curves to connect the data points in sequence to obtain a curve graph of data volume changing with time; draw a tangent line at the data point, and use data The tangent expression is obtained by fitting, and the derivative of the tangent expression is calculated and recorded as Dkj; it should be noted that when the data derivative is greater than zero, it means that the amount of data at the data point shows an increasing trend, that is, the amount of data collected at this time increases greatly; when the data derivative is less than zero, it means that the amount of data at the data point shows a decreasing trend, that is, the amount of data collected at this time is decreasing; the data derivatives greater than zero are summed up to obtain the value of increase in quantity, recorded as S1k, and the data derivatives less than zero are summed up and the absolute value is taken to calculate the value of decrease in data quantity, recorded as S2k;

[0043] The data derivative equal to zero is recorded as the stable derivative, the data volume of the data point corresponding to the stable derivative is extracted, and it is recorded as the stable value of the stable derivative, thereby obtaining the stable value of each stable derivative; the stable value is compared and analyzed with the set stable interval. When the stable value is greater than the set upper limit of the stable interval, it means that the data volume is maintained at a relatively large level at this time, and a high level point is accumulated; when the stable value is in the set stable interval, a medium level point is accumulated; when the stable value is less than the set lower limit of the stable interval, a low level point is accumulated; the number of high level points, medium level points and low level points is counted respectively, and they are recorded as G1k, G2k and G3k respectively; the data volume corresponding to the high level point, medium level point and low level point is summed up to obtain the high level stable value, medium level stable value and low level stable value, and they are recorded as G4k, G5k, G6k respectively; using the set formula The stable level value Gak is calculated, where a1, a2, and a3 are proportional coefficients set respectively, and a1>a2>a3>0;

[0044] The data derivative Dkj, the quantity increase value S1k, the data quantity decrease value S2k, the stable level value Gak and the data quantity Lkj are calculated by the formula The data evaluation index SG corresponding to each type of data is calculated, where b1, b2, b3, b4, and b5 are the set proportional coefficients respectively. is the mean of the data derivatives of each data point; from the formula, we can see that when the amount of data is larger and the increasing trend is greater, the data evaluation index is larger; the correspondingly smaller the amount of data is, the greater the decreasing trend is, the smaller the data evaluation index is; when the amount of data is more volatile (i.e. The larger the value, the more unstable the amount of collected data is. When the amount of data suddenly becomes large, it is easy to cause data packet loss or delay, etc., and the larger the data evaluation index is. The data evaluation index corresponding to each type of data is summed up and calculated to obtain the data evaluation index of UAV mapping, which is recorded as SG.

[0045] Data transmission information analysis:

[0046] Obtain the data transmission information at each acquisition time, where the data transmission information includes network bandwidth (network bandwidth here refers to the maximum bandwidth of the network connection, indicating the actual bandwidth capacity of the network connection) and delay (delay here is a value indicating the duration, which indicates the round-trip time of the Ping command, i.e. RRT), and record them as Wj and Pj respectively; use the set formula The transmission performance value WPj is calculated, where c1 and c2 are the set proportional coefficients respectively;

[0047] A two-dimensional rectangular coordinate system is constructed with time as the horizontal coordinate and the transmission performance value as the vertical coordinate. The transmission performance value is plotted on the coordinate axis according to its corresponding collection time to obtain the performance point; a smooth curve is used to connect the performance points in sequence to obtain a curve graph of the transmission performance value changing with time; a tangent line is drawn at the performance point, and a tangent expression is obtained by data fitting. The tangent expression is derived to obtain the performance derivative, recorded as Xj; the performance derivatives greater than zero are summed to obtain the performance value-added, and the performance derivatives less than zero are summed and the absolute value is taken to obtain the performance increase value, recorded as Z1, and the performance derivatives less than zero are summed to obtain the performance reduction value, recorded as Z2; using the set formula The performance evaluation index ZX is calculated, where c3, c4, and c5 are the set proportional coefficients. is the mean of the performance derivatives of each performance point; from the formula, it can be seen that when the transmission performance value increases, it means that its transmission capacity is greater, the possibility of data transmission congestion or packet loss is relatively low, and the performance evaluation index is greater; when the transmission performance value fluctuates more (i.e. The larger the value, the more unstable the transmission performance is. When the transmission performance suddenly decreases, it is easy to cause data transmission packet loss or delay, and the smaller the performance evaluation value is.

[0048] The data evaluation index SG and performance evaluation index ZX of drone mapping are calculated by the set formula Calculate and obtain the optimized demand value Yd, where d1 and d2 are respectively the set proportional coefficients; compare and analyze the optimized demand value with the set demand threshold. When the optimized demand value is greater than the set demand threshold, it indicates that there is a relatively large risk of data transmission congestion and packet loss, and then generate an optimization instruction and send it to the surveying and mapping optimization module;

[0049] By real-time monitoring and comprehensive analysis of the data volume and transmission performance during the UAV mapping process, optimization instructions can be automatically generated to improve the stability of data collection and the performance of network transmission; the automated processing method reduces the need for manual intervention, improves the intelligence level of the system, and provides effective support for subsequent mapping optimization.

[0050] The surveying and mapping optimization module optimizes the surveying and mapping data transmission and surveying and mapping path of the UAV based on the received optimization instructions, so as to avoid repeated collection, reduce repeated data transmission, improve effective data transmission and reduce repeated surveying and mapping while ensuring comprehensive surveying and mapping data collection; specifically:

[0051] Optimization of surveying and mapping data transmission:

[0052] Retrieving the data evaluation index of each type of data. It should be noted that the larger the data evaluation index of a certain type of data is, the greater the probability of information congestion or packet loss will be;

[0053] Each type of data is set to correspond to a type value. It should be noted that the type value is set by a person skilled in the art according to actual needs. The specific setting method is: the more important the data of this type is for this surveying and mapping task, the larger the corresponding type value is; the data category is compared with all the set data categories to match the corresponding type value;

[0054] The data evaluation index and type value corresponding to each type of data are taken as the numerical value, and the numerical value is multiplied to obtain the priority index of each type of data; each type of data is arranged in descending order according to its corresponding priority index to obtain a data transmission list, and data transmission is performed according to the data transmission list;

[0055] Through reasonable priority sorting and transmission strategies, not only the efficiency and reliability of data transmission are improved, but also the timely processing of key data is guaranteed, the utilization of system resources is optimized, and the overall task execution effect is improved.

[0056] Mapping path optimization:

[0057] Step 1: Get the mapping area map, flight path (the flight path here is the pre-set path for the drone to fly next, excluding the path the drone has already flown) and the current drone position of this mapping mission; divide the mapping area into several blocks according to the coverage boundary of the drone mapping, and mark the blocks that have been mapped and the blocks that have not been mapped with different colors on the mapping area map;

[0058] Step 2: Import the flight path, current drone position, and survey area map into the GIS system, divide the flight path into several sections, and each section needs to be spatially overlaid with the blocks in the survey area map separately; when the completed survey block overlaps with the area covered by the section survey, the completed survey block is recorded as the overlapping block, and the overlapping area is calculated; thus, the overlapping blocks of each section and the overlapping area of ​​each overlapping block can be obtained;

[0059] Step 3: Count the number of overlapping blocks of each road section and record it as M1, and sum up the overlapping areas of each overlapping block to get the total overlapping area and record it as M2; calculate the area covered by the surveying of each road section and record it as M3; use the set formula Calculate and obtain the overlap value Mg of each road section, where g1 and g2 are respectively the set proportional coefficients; it can be seen from the formula that when the number of overlapping blocks is greater, the overlap degree between the road section and the blocks that have been mapped is higher, the degree of repeated mapping is greater, and the overlap value is greater; when the area covered during the mapping of the road section is closer to the total overlapping area, the overlap degree between the road section and the blocks that have been mapped is higher, the degree of repeated mapping is greater, and the overlap value is greater; compare and analyze the overlap value of each road section with the set overlap threshold value, when the overlap value is greater than or equal to the set overlap threshold value, it means that there is a large degree of repeated mapping when flying from the road section, and the road section needs to be optimized; then the road section is recorded as a repeated road section;

[0060] Step 4: Select the flight start and end points of the repeated sections, and use the genetic algorithm to generate several alternative sections, thereby obtaining the alternative sections of each repeated section; select any of the repeated sections, and repeat steps 2 to 3 for its corresponding alternative section to obtain the overlap value of each alternative section. When the overlap value of each alternative section is less than the set overlap threshold, select the alternative section with the smallest overlap value as the target section, and replace the corresponding repeated section with the target section; when the overlap values ​​of each alternative section are greater than or equal to When the overlap threshold is equal to the set overlap threshold, the length of the repeated section and each alternative section is obtained, and the weighted calculation is performed with the corresponding overlap value to obtain the optimization value, and the section with the smallest optimization value (the section here refers to the repeated section and its corresponding alternative section) is selected as the target section, and the target section replaces the corresponding repeated section; it should be noted that if the optimization value of the repeated section is the smallest, the repeated section is selected as the target section, which is equivalent to still flying according to the original section; in the actual experimental process, this situation exists, but the probability of this situation is relatively small;

[0061] Step 5: The optimized path after replacement can be obtained, and the UAV is controlled to perform flight mapping according to the optimized path;

[0062] The degree of repeated mapping of each section in the flight path is quantitatively analyzed to determine the degree of repeated mapping of each section, and each section is optimized accordingly; the flight path of the drone is dynamically adjusted to avoid repeated mapping to the greatest extent, reduce repeated data transmission, and improve the effective collection and transmission of mapping data.

[0063] It should also be noted that the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, commodity or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, commodity or device. In the absence of more restrictions, the elements defined by the sentence "comprises a ..." do not exclude the existence of other identical elements in the process, method, commodity or device including the elements.

[0064] The above are only embodiments of the present application and are not intended to limit the present application. For those skilled in the art, the present application may have various changes and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application should be included within the scope of the claims of the present application.

Claims

1. UAV mapping system based on deep learning, including: The UAV surveying and mapping terminal is characterized in that the UAV surveying and mapping terminal includes: a demand analysis module and a surveying and mapping optimization module; The demand analysis module monitors and analyzes the amount of data collected during the UAV surveying and the transmission performance during the data transmission process to obtain the data evaluation index and the performance evaluation index, normalizes them and takes their values, and analyzes the values ​​to obtain the optimization demand value; generates optimization instructions based on the optimization demand value and sends them to the surveying and mapping optimization module; The surveying and mapping optimization module optimizes the data transmission and the UAV flight path during the surveying and mapping process based on the received optimization instructions. The specific steps of flight path optimization are: Step 1: Obtain the survey area map, flight path, and current drone position of this surveying mission; divide the survey area into several blocks according to the coverage boundary of the drone surveying, and mark the blocks that have been surveyed and the blocks that have not been surveyed on the survey area map; Step 2: Import the flight path, current drone position, and survey area map into the GIS system, divide the flight path into several sections, and perform spatial overlay analysis on each section with the blocks in the survey area map to obtain the overlapping blocks of each section and the overlapping area of ​​each overlapping block; Step 3: Count the number of overlapping blocks of each road section, and sum up the overlapping areas of each overlapping block to obtain the total overlapping area; calculate the area covered when each road section is surveyed; normalize the number of overlapping blocks, the total overlapping area and the area covered when the road section is surveyed and taken as values, and analyze the values ​​to obtain the overlapping value of each road section; when the overlapping value is greater than or equal to the set overlapping threshold, the road section is recorded as a repeated road section; Step 4: Select the flight start point and flight end point of the repeated section, and use the genetic algorithm to generate several alternative sections, thereby obtaining the alternative sections of each repeated section; select any one of the repeated sections, repeat steps 2 to 3 for its corresponding alternative section to obtain the overlap value of each alternative section, when the overlap value of each alternative section is less than the set overlap threshold, select the alternative section with the smallest overlap value as the target section, and replace the corresponding repeated section with the target section; when the overlap value of each alternative section is greater than or equal to the set overlap threshold, obtain the length of the repeated section and each alternative section, and perform weighted calculation with the corresponding overlap value to obtain the optimized value, select the section with the smallest optimized value as the target section, and replace the corresponding repeated section with the target section; Step 5: The optimized path after replacement can be obtained, and the UAV is controlled to perform flight mapping according to the optimized path; The specific process of monitoring and analyzing the amount of data collected during drone mapping is as follows: 301: classify the data collected by the UAV surveying and mapping, wherein any category is denoted as k, k=1, 2, 3...K, K is a positive integer, and K represents the total category; obtain the data volume of each category of data collected by each UAV at each collection time; 302: Take any one type of data, use time as the horizontal axis and data volume as the vertical axis to obtain a curve graph of data volume changing with time; draw a tangent line at the data point, obtain a tangent line expression by data fitting, and calculate the derivative of the tangent line expression to obtain the data derivative; sum the data derivatives greater than zero to obtain the quantity increase value, and sum the data derivatives less than zero and take the absolute value to calculate the data volume decrease value; 303: Recording the data derivative equal to zero as a stable derivative, extracting the data amount of the data point corresponding to the stable derivative, and recording it as the stable value of the stable derivative, thereby obtaining the stable value of each stable derivative, and performing quantitative analysis on it to obtain a stable level value; 304: normalize the data derivative, the quantity increase value, the data quantity decrease value, the stable level value and the data quantity and obtain their values, analyze the values ​​to obtain the data evaluation index corresponding to each type of data, and sum the data evaluation index corresponding to each type of data to obtain the data evaluation index of the UAV mapping; The specific process of data transmission optimization is as follows: Retrieve the data evaluation index of each type of data, set each type of data to correspond to a type value, and compare the data category with all the set data categories to match the corresponding type value; The data evaluation index and type value corresponding to each type of data are taken as the numerical value, and the numerical value is multiplied to obtain the priority index of each type of data; each type of data is arranged in descending order according to its corresponding priority index to obtain a data transmission list, and data transmission is performed according to the data transmission list.

2. The deep learning-based UAV mapping system according to claim 1, characterized in that: The specific process of quantitative analysis of the stable values ​​of each stable derivative is as follows: Compare and analyze the stable value with the set stable interval to divide the data points corresponding to the stable value into high level points, middle level points and low level points, count the cumulative number of high level points, middle level points and low level points respectively, and sum the data corresponding to the high level points, middle level points and low level points to obtain the high level stable value, middle level stable value and low level stable value respectively; The cumulative number of high-level points, the cumulative number of medium-level points, the cumulative number of low-level points, the high-level stable value, the medium-level stable value and the low-level stable value are calculated and analyzed by formulating.

3. The deep learning-based UAV mapping system according to claim 1, characterized in that: The specific process of monitoring and analyzing the transmission performance during data transmission is as follows: 501: obtaining data transmission information at each collection moment, wherein the data transmission information includes network bandwidth and delay, normalizing the information and obtaining a value thereof, and analyzing the value to obtain a transmission performance value at each collection moment; 502: construct a two-dimensional rectangular coordinate system with time as the horizontal coordinate and the transmission performance value as the vertical coordinate, plot the transmission performance value on the coordinate axis according to its corresponding collection time to obtain performance points; use a smooth curve to connect the performance points in sequence to obtain a curve graph of the transmission performance value changing with time; Draw a tangent line of the curve at the performance point, use data fitting to obtain a tangent expression, and perform a derivative calculation on the tangent expression to obtain a performance derivative; sum the performance derivatives greater than zero to obtain a performance increase value, and sum the performance derivatives less than zero and take the absolute value to obtain a performance decrease value; 503: The performance derivative, the performance increase value, the performance decrease value and the transmission performance value are calculated and analyzed in a formulaic manner to obtain a performance evaluation index.

Images