An ecological destruction monitoring system based on satellite remote sensing and unmanned aerial vehicle cooperation

Abstract

The application provides an ecological damage monitoring system based on satellite remote sensing and cooperation of unmanned aerial vehicles, and belongs to the technical field of ecological environment monitoring. The system comprises a remote sensing data analysis unit, an unmanned aerial vehicle scheduling unit and an ecological monitoring unit. The remote sensing data analysis unit determines at least one target monitoring area based on regional traversal remote sensing difference data; the unmanned aerial vehicle scheduling unit determines an unmanned aerial vehicle scheduling scheme based on the position of the at least one target monitoring area, and schedules at least one unmanned aerial vehicle to the at least one target monitoring area based on the unmanned aerial vehicle scheduling scheme; and the ecological monitoring unit judges whether the target monitoring area appears ecological damage phenomenon through real-time image data of the target monitoring area returned by the at least one unmanned aerial vehicle. The technical scheme of the application fully gives play to the cooperative effect of satellite remote sensing data and real-time image data of unmanned aerial vehicles, and proposes multiple effective technical means for identifying ecological damage, so that the ecological damage area can be quickly and effectively identified and continuously and effectively tracked.

Description

TECHNICAL FIELD

[0001] The present application belongs to the technical field of ecological environment monitoring, and particularly relates to an ecological destruction monitoring system based on satellite remote sensing and unmanned aerial vehicle cooperation. BACKGROUND

[0002] Ecological destruction refers to the damage caused by human activities or natural factors to the ecosystem, which usually manifests as the degradation of the structure and function of the ecosystem, including large-scale deforestation, leading to decreased vegetation coverage, intensified soil erosion, and reduced biodiversity; land desertification, salinization, and water and soil loss, resulting in decreased land productivity and damaged ecosystem function; pollution of rivers, lakes, oceans, and other water bodies by industrial wastewater, agricultural non-point source pollution, and domestic sewage, leading to decreased water quality and the death of aquatic organisms; reclamation, drainage, and pollution of wetlands, resulting in reduced wetland area and lost ecological function; and mining leading to destruction of surface vegetation, soil erosion, and heavy metal pollution.

[0003] Image analysis is an important means of ecological destruction monitoring, and through satellite remote sensing images and unmanned aerial vehicle images, the types and degrees of ecological destruction can be quickly and accurately identified. In related technologies, patent publication CN109684929A proposes a terrestrial plant ecological environment monitoring method based on multi-source remote sensing data fusion, which is suitable for terrestrial plant ecological environment monitoring; and patent publication CN118608972A proposes a grassland ecological hazard intelligent monitoring and early warning method based on big data technology, which uses a generative idea to effectively solve the problem of identifying grassland ecological hazard types from satellite remote sensing images.

[0004] However, in related technologies, unmanned aerial vehicle images and satellite remote sensing data do not truly cooperate, and there is no problem of unmanned aerial vehicle scheduling scheme; at the same time, the different roles of satellite remote sensing data and unmanned aerial vehicle returned data are not reflected, and there is room for improvement in regional identification pertinence, image acquisition effectiveness, and destruction scene recognition accuracy. SUMMARY

[0005] To solve the above technical problems, the present application proposes an ecological destruction monitoring system based on satellite remote sensing and unmanned aerial vehicle cooperation.

[0006] The technical solution of the present application is as follows:

[0007] An ecological destruction monitoring system based on satellite remote sensing and unmanned aerial vehicle cooperation, the system comprising a remote sensing data analysis unit, an unmanned aerial vehicle scheduling unit, and an ecological monitoring unit;

[0008] The remote sensing data analysis unit determines at least one target monitoring region based on regionally traversed remote sensing difference data;

[0009] The unmanned aerial vehicle scheduling unit determines an unmanned aerial vehicle scheduling scheme based on the location of the at least one target monitoring area, and schedules at least one unmanned aerial vehicle to the at least one target monitoring area based on the unmanned aerial vehicle scheduling scheme;

[0010] The ecological monitoring unit determines whether the target monitoring area has an ecological damage phenomenon based on real-time image data of the target monitoring area returned by the at least one unmanned aerial vehicle.

[0011] The remote sensing data analysis unit comprises a remote sensing data receiving unit, a remote sensing difference calculation unit, and a target area identification unit;

[0012] The remote sensing data receiving unit receives remote sensing picture data corresponding to each area according to a preset period;

[0013] The remote sensing difference calculation unit calculates difference data of remote sensing picture data corresponding to the nearest adjacent period for each area, and when the difference data meets a preset condition, sends a region ID to the target area identification unit;

[0014] The target area identification unit determines whether to take the region ID as a target monitoring area based on a historical identification result.

[0015] The remote sensing data receiving unit receives remote sensing picture data corresponding to N areas according to N preset periods; N>2;

[0016] Among them, the preset period corresponding to the i-th area Areai is Ti, i=1, 2, …, N;

[0017] If the difference degree of remote sensing picture data corresponding to the nearest adjacent period of the i-th area Areai is less than a preset difference degree threshold, the preset period Ti is increased.

[0018] The historical identification result includes whether the target monitoring area corresponding to the region ID has ever had an ecological damage phenomenon;

[0019] The target area identification unit determines whether to take the region ID as a target monitoring area based on a historical identification result, and specifically comprises:

[0020] If the target monitoring area corresponding to the region ID has never had an ecological damage phenomenon, the region ID is taken as a target monitoring area.

[0021] The ecological monitoring unit determines whether the target monitoring area has an ecological damage phenomenon, and specifically comprises:

[0022] The ecological monitoring unit receives real-time image data of the target monitoring area returned by the unmanned aerial vehicle, and obtains a plurality of sub-image areas after edge segmentation on the real-time image data.

[0023] After performing the wavelet transform on each sub-image region, high-frequency information of each sub-image region is obtained;

[0024] Based on the high-frequency information, it is determined whether the target monitoring region has an ecological damage phenomenon.

[0025] When the ecological monitoring unit determines that the target monitoring region has an ecological damage phenomenon, the region ID of the target monitoring region is recorded, and the region ID is sent to the historical identification result database of the target region identification unit.

[0026] If the difference degree of the remote sensing image data corresponding to the nearest adjacent period of the i-th region Areai is greater than the preset difference degree threshold, it is further determined whether the region ID of the i-th region Areai is saved in the historical identification result database of the target region identification unit.

[0027] If so, the preset period Ti is reduced.

[0028] The ecological monitoring unit determines whether the target monitoring region has an ecological damage phenomenon, specifically including:

[0029] The ecological monitoring unit receives real-time image data of the target monitoring region returned by the unmanned aerial vehicle, performs edge recognition on the real-time image data, and obtains a plurality of edge trajectory line segments;

[0030] For each edge trajectory line segment, a turning point thereof is calculated;

[0031] Each edge trajectory line segment is cut according to the turning point to obtain a plurality of edge trajectory sub-line segments;

[0032] According to the relative position relationship of the plurality of edge trajectory sub-line segments corresponding to different edge line segments, it is determined whether the target monitoring region has an ecological damage phenomenon.

[0033] The ecological monitoring unit determines whether the target monitoring region has an ecological damage phenomenon, specifically including:

[0034] An image semantic recognition model for recognizing ecological damage objects is pre-trained;

[0035] The real-time image data of the target monitoring region returned by the unmanned aerial vehicle is input into the image semantic recognition model;

[0036] Based on the output result of the image semantic recognition model, it is determined whether the target monitoring region has an ecological damage phenomenon.

[0037] The at least one unmanned aerial vehicle is configured with image acquisition devices of multiple resolution modes;

[0038] The at least one unmanned aerial vehicle receives the judgment result of the ecological monitoring unit and determines whether to enable the second resolution to collect the second real-time image data of the target monitoring area based on the judgment result.

[0039] The second resolution is greater than the first resolution.

[0040] The technical scheme of the present application fully utilizes the synergistic effect of satellite remote sensing data and real-time image data of unmanned aerial vehicles, and proposes a variety of effective identification of ecological destruction technical means, which can quickly and effectively identify the ecological destruction area and continuously track it. The specific advantages and implementation principles will be further described in detail in the specific embodiment part combined with the drawings. BRIEF DESCRIPTION OF DRAWINGS

[0041] In order to more clearly illustrate the technical scheme in the embodiments of the present application or the prior art, the drawings needed in the embodiments will be briefly introduced below. Obviously, the drawings in the following description are only some embodiments of the present application, and those skilled in the art can obtain other drawings according to these drawings without creative labor.

[0042] Figure 1 is a main function module composition schematic diagram of an ecological destruction monitoring system based on satellite remote sensing and unmanned aerial vehicle cooperation according to an embodiment of the present application;

[0043] Figure 2 is Figure 1 is an internal composition schematic diagram of a remote sensing data analysis unit in the system;

[0044] Figure 3 is Figure 1 is a preferred embodiment schematic diagram of the remote sensing data analysis unit in the system based on region traversal remote sensing difference data to determine at least one target monitoring area;

[0045] Figures 4-5 is Figure 1 is a two different technical principle diagram of the state monitoring unit judging whether the target monitoring area appears ecological destruction phenomenon in the system; DETAILED DESCRIPTION

[0046] In the specific embodiments of the present application, if the embodiments of the related technical scheme involve user-related data, when the embodiments of the present application are applied to specific products or technologies, the user's permission or consent is required, and the collection, use and processing of related data need to comply with relevant laws, regulations and standards of relevant countries and regions.

[0047] Referring to Figure 1 , Figure 1is a main function module composition schematic diagram of an ecological destruction monitoring system based on satellite remote sensing and unmanned aerial vehicle cooperation according to an embodiment of the present application.

[0048] Figure 1 The system comprises a remote sensing data analysis unit, an unmanned aerial vehicle scheduling unit and an ecological monitoring unit.

[0049] The remote sensing data analysis unit determines at least one target monitoring area based on regionally traversed remote sensing difference data;

[0050] The unmanned aerial vehicle scheduling unit determines an unmanned aerial vehicle scheduling scheme based on the location of the at least one target monitoring area, and schedules at least one unmanned aerial vehicle to the at least one target monitoring area based on the unmanned aerial vehicle scheduling scheme;

[0051] Through real-time image data returned by the at least one unmanned aerial vehicle, the ecological monitoring unit determines whether the target monitoring area has ecological destruction phenomenon.

[0052] Next, referring to Figure 2 , Figure 2 is Figure 1 An internal composition schematic diagram of the remote sensing data analysis unit in the system.

[0053] In Figure 2 , it is shown that the remote sensing data analysis unit comprises a remote sensing data receiving unit, a remote sensing difference calculation unit and a target area identification unit;

[0054] The remote sensing data receiving unit receives remote sensing picture data corresponding to each region according to a preset period;

[0055] The remote sensing difference calculation unit calculates difference data of remote sensing picture data corresponding to the nearest adjacent period for each region, and when the difference data meets a preset condition, sends a region ID to the target area identification unit;

[0056] The target area identification unit determines whether to take the region ID as a target monitoring area based on historical identification results.

[0057] Preferably, referring to Figure 3 , Figure 3 is Figure 1 A preferred embodiment schematic diagram in which the remote sensing data analysis unit in the system determines at least one target monitoring area based on regionally traversed remote sensing difference data.

[0058] Specifically, the scheme can be automatically realized in the form of computer program code, and the relevant computer pseudo code flow language is as follows:

[0059] Input:

[0060] Area IDs corresponding to N regions: {Area1, Area2, …, AreaN}

[0061] Remote sensing data receiving periods T corresponding to N regions: {T1, T2, T3, …, TN};

[0062] Wherein, under the initial condition, T1=T2=T3=…=TN=T0, that is, when initialized, the N regions receive remote sensing data according to the same period T0;

[0063] Output:

[0064] Target monitoring region AreaK.

[0065] Specifically, the computer program code flow language is described as follows:

[0066] Step 1: Let i=1;

[0067] Step 2: Obtain the latest remote sensing image ImageAreaT of Areai according to the preset period Ti Cur ;

[0068] Step 3: Obtain the historical stored previous remote sensing image ImageAreaT of Areai pre ;

[0069] Step 4: Calculate the difference degree of the latest remote sensing image ImageAreaT cur and the previous remote sensing image ImageAreaT pre ;

[0070] Step 5: When the difference degree meets the preset condition, the target monitoring region Areai is taken as the target monitoring region; Let i=i+1, return to step 2.

[0071] The above gives the process of whether a certain monitoring region Areai can be taken as a target monitoring region output, when i=1, 2, …, N, that is, the process of determining at least one target monitoring region based on region traversal of the embodiment described in the remote sensing difference data. Figure 1

[0072] For the convenience of understanding, see Figure 3 , further describe the identification steps for a certain specific monitoring region Areai as follows( Figure 3 Step numbers are omitted in the following):

[0073] S310: Obtain the remote sensing picture data ImageTi of the Ti period for the ith region Areai;

[0074] ​S320: calculate the difference degree D between the remote sensing picture data ImageTi of the i-th region Areai in the T i-1 i-th period and the remote sensing picture data ImageT i-1 i-1 of the i-th region Areai in the T i i-1 period;

[0075] Since the remote sensing data receiving unit receives the remote sensing picture data corresponding to the N regions in N preset periods, in step S310, the remote sensing picture data of the latest period of the i-th region Areai is obtained, which is described as "remote sensing picture data ImageTi of the i-th region Areai in the T

[0076] For example, assuming that for the third region Area3, the preset period is 3 days (T3=3), that is, the remote sensing data of the region Area3 is obtained once every three days, and assuming that the acquisition starts from January 1.

[0077] Then, the remote sensing data of the region Area3 is obtained once on January 4, and so on.

[0078] Assuming that the current time is January 10.

[0079] Therefore, the remote sensing picture data ImageT3 of the third region Area3 in the T3 period obtained by step S310 on January 10 is the latest remote sensing data image of the region Area3 obtained on that day.

[0080] Therefore, step S310 can also be described as: obtaining the latest remote sensing picture data ImageTi of the i-th region Areai in the T

[0081] The remote sensing picture data of the i-th region Areai in the T i-1 i-1 period is called ImageT i-1 i-1.

[0082] In the foregoing example, that is, the remote sensing picture data of the third region Area3 obtained on January 7.

[0083] Therefore, step S320 can also be described as: calculating the difference degree D between the latest remote sensing data image ImageTi and the remote sensing picture data ImageT i-1 i-1. i

[0084] Next, step S330 is entered: judging whether the difference degree Di is greater than a preset threshold PresetD.

[0085] If the difference Di is not greater than the preset threshold PresetD, it means that the i-th region Areai has hardly changed in this period, that is, the ecological status remains good, and it is highly possible that no ecological destruction (mutation) phenomenon occurs, and the monitoring of the region can be regarded as normal, and the region also has a trend of keeping stable. Therefore, at this time, the preset period Ti can be increased.

[0086] Taking the third region Area3 with an initial preset period of 3 days (T3=3) as an example, at this time, T3 can be increased, for example, T3=4, that is, remote sensing data of the region Area3 is acquired every 4 days for analysis next time;

[0087] On the contrary, if the difference Di is greater than the preset threshold PresetD, it means that the i-th region Areai has changed greatly in this period, that is, the ecological status is destroyed, and it is highly possible that an ecological destruction (mutation) phenomenon occurs, and at this time, the region needs further observation and test acceptance.

[0088] However, if it is directly and simply taken as a target monitoring region, there may be a problem of repeated processing, because the region may have been listed as a key observation region in the previous processing period.

[0089] Therefore, the embodiment of the present application further improves this as follows: when the difference Di is greater than the preset threshold PresetD, it is further judged whether the region ID of Areai is saved in the historical identification result database;

[0090] If yes, it means that the region has been listed as a key observation region in the previous processing period, and at this time, the preset period Ti needs to be reduced;

[0091] Taking the third region Area3 with an initial preset period of 3 days (T3=3) as an example, at this time, T3 can be reduced, for example, T3=2, that is, remote sensing data of the region Area3 is acquired every 2 days for analysis next time.

[0092] In this step, whether the period is reduced or increased, it means that the judgment of the region Areai has ended, and it is necessary to return to S310;

[0093] If the difference Di is greater than the preset threshold PresetD, and the region ID of Areai is not saved in the historical identification result database, it means that Areai is indeed a newly discovered region that needs to be paid attention to, and therefore, it is taken as a target monitoring region.

[0094] When the ecological monitoring unit determines that the target monitoring area has an ecological destruction phenomenon, the area ID of the target monitoring area is recorded, and the area ID is sent to the historical identification result database of the target area identification unit.

[0095] In Figure 3 In the embodiment, the focus is on the setting of the preset threshold PresetD.

[0096] According to long-term observation practice, the stability and variability of different target areas change greatly over time, especially over seasons.

[0097] For example, for forest vegetation, if it is in a normal and healthy ecological state, the remote sensing image change value in different periods in the preset period of midsummer and early autumn is relatively slow (there is a change, but the change is not large), at this time, a lower preset threshold PresetD can be set; in other seasons, for example, at the junction of late spring and early summer, the change of forest vegetation can be very large within one day (including coverage area, color profile, etc.), at this time, although the change is large, it is still in a normal state, at this time, a larger preset threshold PresetD can be set.

[0098] That is, Figure 3 In the embodiment, the preset threshold dynamically changes over time, especially over seasons, and can be pre-set by professionals according to the ecological species and type of the target area.

[0099] It should be noted that if the ecology of an area is initially in a destruction state, the state of the area entering the destruction state is mostly unstable, and the difference between the remote sensing images obtained in the adjacent two periods is necessarily large, much larger than the normal vegetation change, at this time, the above-mentioned embodiment of the present application can still accurately list it as a target monitoring area.

[0100] After one or more target monitoring areas are determined, the unmanned aerial vehicle scheduling unit determines an unmanned aerial vehicle scheduling scheme based on the location of the at least one target monitoring area, and schedules at least one unmanned aerial vehicle to the at least one target monitoring area based on the unmanned aerial vehicle scheduling scheme.

[0101] When there are multiple target monitoring areas, the multiple target monitoring areas can be position clustered, and target monitoring areas in adjacent positions can be classified into one class; the same unmanned aerial vehicle is scheduled for target monitoring areas in the same class, so as to avoid waste of unmanned aerial vehicle resources.

[0102] The ecological monitoring unit determines whether the target monitoring area has an ecological destruction phenomenon through the real-time image data of the target monitoring area returned by the at least one unmanned aerial vehicle.

[0103] To further improve the recognition accuracy, as a more preferred example, the at least one unmanned aerial vehicle is configured with multiple resolution mode image acquisition devices;

[0104] The at least one unmanned aerial vehicle collects first real-time image data of the target monitoring area at a first resolution and sends it to the ecological monitoring unit. After receiving the judgment result of the ecological monitoring unit, it is determined whether to enable the collection of second real-time image data of the target monitoring area at a second resolution based on the judgment result; the second resolution is greater than the first resolution.

[0105] Specifically, after the at least one unmanned aerial vehicle collects first real-time image data of the target monitoring area at a first resolution and sends it to the ecological monitoring unit, if the ecological monitoring unit identifies that the target monitoring area does not have ecological damage phenomenon, the unmanned aerial vehicle continues to enable the second resolution to continue collecting second real-time image data of the target monitoring area and sending it to the ecological monitoring unit for further judgment.

[0106] The above preferred embodiment further ensures that the target monitoring area with large differences in remote sensing data in adjacent periods will not be misjudged due to the shooting mode of the unmanned aerial vehicle through secondary confirmation.

[0107] Preferably, the unmanned aerial vehicle in the second resolution mode shoots the target monitoring area at a second position and a second angle different from the first resolution mode to obtain second real-time image data of the target monitoring area.

[0108] Next, the technical principle of how the ecological monitoring unit judges whether the target monitoring area has an ecological damage phenomenon is introduced.

[0109] The ecological monitoring unit judges whether the target monitoring area has an ecological damage phenomenon based on the real-time image data of the target monitoring area returned by the unmanned aerial vehicle.

[0110] For different target monitoring areas, the forms of ecological damage are different. After preliminary induction, the following situations are mainly included:

[0111] 1. Forest vegetation: mainly manifested as large-scale deforestation, leading to decreased vegetation coverage, intensified soil erosion, and reduced biodiversity.

[0112] At this time, in the real-time image returned by the unmanned aerial vehicle, the forest area will appear obvious patchy or strip-shaped bare land, and the color will change from green to brown or gray. The unmanned aerial vehicle image can more clearly see the stumps and residual wood after deforestation.

[0113] 2. Basic land farmland: mainly manifested as land desertification, salinization, water and soil loss, etc., leading to decreased land productivity and damaged ecosystem function.

[0114] Land degradation areas usually appear light yellow or white, in sharp contrast to the surrounding green vegetation. UAV images can observe soil erosion gullies and saline-alkaline surface features in more detail.

[0115] 3. Water bodies (rivers, lakes, oceans): mainly manifested as: rivers, lakes, oceans and other water bodies are polluted by industrial wastewater, agricultural non-point source pollution, domestic sewage and other pollutants, water quality declines, and aquatic organisms die.

[0116] Contaminated water bodies usually present abnormal colors such as black, red or green (eutrophication). UAV images can more clearly see the oil film, garbage and other pollutants on the surface of the water body.

[0117] 4. Wetland destruction: mainly manifested as: wetlands are reclamation, drainage, pollution, leading to a decrease in wetland area and loss of ecological function.

[0118] The color and texture of the wetland area will change from green or blue to brown or gray. UAV images can more clearly see the boundaries and internal structure of the wetland.

[0119] 5. Mining: mainly manifested as destruction of surface vegetation, soil erosion, and heavy metal pollution.

[0120] Mining areas usually present obvious patchy bare land with brown or gray color. UAV images can more clearly observe the mining face, waste rock pile and tailings pond of the mine.

[0121] Based on the above principles, an image semantic recognition model for recognizing ecological damage objects can be pre-trained;

[0122] The real-time image data of the target monitoring area returned by the UAV is input into the image semantic recognition model;

[0123] Based on the output result of the image semantic recognition model, it is determined whether the target monitoring area has ecological damage phenomenon.

[0124] The advantage of the image semantic recognition model is that the target monitoring area can be pre-classified, and then the training model is adjusted to make it more suitable for semantic recognition and judgment of UAV returned images of a specific type of target monitoring area.

[0125] However, when the type of the target monitoring area is unknown, the effect of the image semantic recognition model may need to be further confirmed by artificial, resulting in low accuracy and efficiency.

[0126] Therefore, the present application proposes further improvement schemes as follows, see Figures 4-5 , Figures 4-5 is Figure 1Two different technical principle diagrams for judging whether the target monitoring area appears ecological damage phenomenon by the system state monitoring unit.

[0127] Firstly referring to Figure 4 , the ecological monitoring unit judges whether the target monitoring area appears ecological damage phenomenon, specifically including: Figure 4

[0128] The ecological monitoring unit receives the real-time image data of the target monitoring area returned by the unmanned aerial vehicle, and obtains a plurality of sub-image areas after performing edge segmentation on the real-time image data;

[0129] After performing wavelet transform on each sub-image area, the high-frequency information of each sub-image area is obtained;

[0130] Based on the high-frequency information, it is judged whether the target monitoring area appears ecological damage phenomenon.

[0131] Research has proved that when the region appears ecological damage, the information carried by the region image after wavelet transform is distributed in a higher proportion in the high-frequency band, because ecological damage will lead to complex and variable characteristics of the region image, thus showing a larger proportion of high-frequency information.

[0132] Therefore, in Figure 4 , after performing edge segmentation on the real-time image data, a plurality of sub-image areas are obtained; after performing wavelet transform on each sub-image area, the high-frequency information of each sub-image area is obtained;

[0133] Based on the distribution proportion of the high-frequency information, it is judged whether the target monitoring area appears ecological damage phenomenon.

[0134] If the distribution proportion is higher than a preset proportion value, it is confirmed that the target monitoring area appears ecological damage phenomenon.

[0135] Preferably, the high-frequency information and the low-frequency information of each sub-image area are obtained, and when the proportion of the high-frequency information exceeds three times the proportion of the low-frequency information, it is confirmed that the target monitoring area appears ecological damage phenomenon.

[0136] It can be understood that the high-frequency and low-frequency in the image information after wavelet transform have a well-known definition in the art.

[0137] The improved method avoids the problem that the effect of using an image semantic recognition model may need to be further confirmed manually, resulting in low accuracy and efficiency.

[0138] In another aspect, referring to Figure 5 , the ecological monitoring unit judges whether the target monitoring area appears ecological damage phenomenon, specifically including:

[0139] ​The ecological monitoring unit receives real-time image data of the target monitoring area returned by the unmanned aerial vehicle, performs edge recognition on the real-time image data, and obtains a plurality of edge trajectory line segments;

[0140] For each edge trajectory line segment, a turning point thereof is calculated;

[0141] Each edge trajectory line segment is split according to the turning point, and a plurality of edge trajectory sub-line segments are obtained;

[0142] According to the relative position relationship of the plurality of edge trajectory sub-line segments corresponding to different edge line segments, it is determined whether the target monitoring area has an ecological destruction phenomenon.

[0143] Figure 5 A schematic diagram of three edge trajectory line segments (red, blue and yellow) is shown. Among them, the red edge trajectory line segment has one turning point, and the red and blue edge trajectory line segments each have two turning points.

[0144] The turning point is defined in the art as a point where data "mutates". It can usually be determined by finding local or global extreme points, and the specific turning point determination method belongs to the common knowledge in the art, which will not be expanded in this embodiment.

[0145] As mentioned above, ecological destruction will lead to complex and variable regional image features. According to statistics, ecological destruction usually means more complex contours, and random crossing without rules, resulting in more contour mutation crossing situations.

[0146] Therefore, based on the above features, according to the relative position relationship of the plurality of edge trajectory sub-line segments corresponding to different edge line segments, it is determined whether the target monitoring area has an ecological destruction phenomenon.

[0147] In the above process, first, the turning point corresponding to each edge line segment is determined, and then each edge trajectory line segment is split according to the turning point to obtain a plurality of edge trajectory sub-line segments.

[0148] If two edge trajectory sub-line segments belonging to different edge line segments intersect, and the intersection point is not any turning point, it is determined that the target monitoring area has an ecological destruction phenomenon.

[0149] This is because, if the intersection point of two edge trajectory sub-line segments belonging to different edge line segments is exactly a turning point, it is very likely that the two edge trajectory sub-line segments of different edge line segments form a connected contour line, which is relatively common in diversified vegetation of different levels (color levels), and should not be considered as ecological destruction.

[0150] If two edge track sub-segments belonging to different edge line segments intersect and the intersection point is not any inflection point, it is likely to be an abnormal ecological mutation, and thus it is determined that the ecological damage phenomenon occurs in the target monitoring area.

[0151] Of course, in practical applications, the above methods can be combined, for example, the image semantic recognition model, Figure 4 , Figure 5 The judgment process of whether the ecological damage phenomenon occurs in the target monitoring area is determined by performing the methods of the image semantic recognition model,

[0152] respectively in parallel, and when the judgment results are completely consistent, it is determined that the ecological damage phenomenon occurs in the target monitoring area; when the judgment results are not completely consistent, a majority over minority mode is used for determination.

[0153] It should be noted that, Figure 4 , Figure 5 The process method of determining whether the ecological damage phenomenon occurs in the target monitoring area according to the UAV back image is an improved technical means first summarized by the inventor after long-term observation, and there is no related record in the prior art, which is one of the important improvements of the present application; and long-term application proves that the recognition and judgment effect is in most cases due to the general artificial intelligence model, such as the aforementioned image semantic recognition model.

[0154] Although not shown in the drawings, a preferred and more product embodiment can also be an electronic device, which includes a memory and one or more processors. The memory has one or more application programs stored therein, and the one or more application programs are adapted to be executed by the one or more processors Figures 3-5 The method steps of the embodiments.

[0155] Although not shown in the drawings, more embodiments also include a computer-readable storage medium, which stores a computer program that, when executed, Figures 3-5 The method steps of the embodiments are implemented.

[0156] It can be understood that the system, product, device, medium embodiments and method embodiments correspond to each other and can be mutually referred to, and the principles are similar or the same, so they will not be repeated.

[0157] The technical solution of the present application fully utilizes the synergistic effect of satellite remote sensing data and real-time image data of unmanned aerial vehicles, and proposes multiple effective identification of ecological damage technology, which can quickly and effectively identify the ecological damage area and realize continuous tracking.

[0158] Other techniques, principles, algorithms or models not detailed herein can be found in the prior art.

[0159] The foregoing has outlined rather broadly the method embodiments and system of the present application so that those skilled in the art can appreciate the principles of the present application. It is contemplated that those skilled in the art will be able to devise various changes, modifications, substitutions and equivalents of the embodiments of the present application without departing from the principles and spirit of the present application. It is intended that the present application embrace all such changes, modifications, substitutions and equivalents as fall within the scope of the appended claims and their equivalents.

Claims

1. A satellite remote sensing and unmanned aerial vehicle (UAV) cooperative ecological damage monitoring system, the system comprising a remote sensing data analysis unit, a UAV scheduling unit, and an ecological monitoring unit, characterized in that: the remote sensing data analysis unit determines at least one target monitoring area based on regionally traversed remote sensing difference data; the UAV scheduling unit determines a UAV scheduling scheme based on the location of the at least one target monitoring area, and schedules at least one UAV to the at least one target monitoring area based on the UAV scheduling scheme; the ecological monitoring unit determines whether the target monitoring area has ecological damage phenomenon based on real-time image data returned by the at least one UAV, comprising: the ecological monitoring unit receives real-time image data returned by the UAV for the target monitoring area, performs edge recognition on the real-time image data, and obtains a plurality of edge trajectory line segments; for each edge trajectory line segment, a turning point thereof is calculated; each edge trajectory line segment is divided into a plurality of edge trajectory sub-line segments according to the turning points; if two edge trajectory sub-line segments belonging to different edge line segments intersect, and the intersection point is not any turning point, it is determined that the target monitoring area has ecological damage phenomenon. 2.The satellite remote sensing and UAV cooperative ecological damage monitoring system of claim 1, characterized in that: the remote sensing data analysis unit comprises a remote sensing data receiving unit, a remote sensing difference calculation unit, and a target area identification unit; the remote sensing data receiving unit receives remote sensing picture data corresponding to each region according to a preset period; the remote sensing difference calculation unit calculates difference data of remote sensing picture data corresponding to the nearest adjacent period for each region, and sends a region ID to the target area identification unit when the difference data meets a preset condition; the target area identification unit determines whether to take the region ID as a target monitoring area based on historical identification results. 3.The satellite remote sensing and UAV cooperative ecological damage monitoring system of claim 2, characterized in that: the remote sensing data receiving unit receives remote sensing picture data corresponding to N regions according to N preset periods; N>2; wherein the preset period corresponding to the i th region Areai is Ti, i=1, 2, …, N; if the difference degree of remote sensing picture data corresponding to the nearest adjacent period of the i th region Areai is less than a preset difference degree threshold, the preset period Ti is increased. 4.The satellite remote sensing and UAV cooperative ecological damage monitoring system of claim 2, characterized in that: the historical identification results include whether the target monitoring area corresponding to the region ID has ever had ecological damage phenomenon; the target area identification unit determines whether to take the region ID as a target monitoring area based on historical identification results, specifically comprising: if the target monitoring area corresponding to the region ID has never had ecological damage phenomenon, the region ID is taken as a target monitoring area. 5.The satellite remote sensing and UAV cooperative ecological damage monitoring system of claim 1, characterized in that: The ecological monitoring unit determines whether the target monitoring area has an ecological destruction phenomenon, specifically comprising: The ecological monitoring unit receives real-time image data of the target monitoring area returned by the unmanned aerial vehicle, and obtains a plurality of sub-image regions after edge segmentation of the real-time image data; After performing wavelet transform on each sub-image region, high-frequency information of each sub-image region is obtained; Based on the high-frequency information, it is determined whether the target monitoring area has an ecological destruction phenomenon.

6. The ecological destruction monitoring system based on satellite remote sensing and unmanned aerial vehicle cooperation according to claim 1, characterized in that: When the ecological monitoring unit determines that the target monitoring area has an ecological destruction phenomenon, the region ID of the target monitoring area is recorded and sent to the historical identification result database of the target region identification unit.

7. The ecological destruction monitoring system based on satellite remote sensing and unmanned aerial vehicle cooperation according to claim 3, characterized in that: If the difference degree of the remote sensing image data corresponding to the nearest adjacent period of the i-th region Areai is greater than the preset difference degree threshold, it is further determined whether the region ID of the i-th region Areai is saved in the historical identification result database of the target region identification unit; If yes, the preset period Ti is reduced.

8. The ecological destruction monitoring system based on satellite remote sensing and unmanned aerial vehicle cooperation according to claim 1, characterized in that: The ecological monitoring unit determines whether the target monitoring area has an ecological destruction phenomenon, further comprising: pre-training an image semantic recognition model for identifying ecological destruction objects; inputting the real-time image data of the target monitoring area returned by the unmanned aerial vehicle into the image semantic recognition model; determining whether the target monitoring area has an ecological destruction phenomenon based on the output result of the image semantic recognition model.

9. The ecological destruction monitoring system based on satellite remote sensing and unmanned aerial vehicle cooperation according to claim 1, characterized in that: The at least one unmanned aerial vehicle is equipped with image acquisition devices of multiple resolution modes; After the first real-time image data of the target monitoring area collected by the at least one unmanned aerial vehicle at a first resolution is sent to the ecological monitoring unit, the judgment result of the ecological monitoring unit is received, and based on the judgment result, it is determined whether to enable the second real-time image data of the target monitoring area collected at a second resolution; The second resolution is greater than the first resolution.

Images