A method for monitoring and early warning of grassland caterpillar larvae based on UAV remote sensing technology

Abstract

This invention discloses a method for monitoring and early warning of grassland caterpillar larvae based on UAV remote sensing technology. The method involves: utilizing the flexibility and ease of operation of UAVs, along with corresponding aerial photography, recording, verification, and statistical methods, to acquire large-scale visible light image data with millimeter-level measurement accuracy based on UAV remote sensing; conducting model construction, verification, and structural parameter comparison analysis in different grassland types, typical climate zones, and areas with high levels of human interference; and using the YOLOv5 algorithm to calculate the number and density of grassland caterpillar larvae in the images. This method uses small, inexpensive UAVs to acquire large-scale spatiotemporal distribution characteristics of grassland caterpillar larvae, requiring minimal manpower, and is highly efficient and low-cost; it is non-destructive; and it enables long-term, fixed-point, and real-time monitoring under specific environments.

Description

Technical Field

[0001] This invention belongs to the field of grassland caterpillar larvae monitoring technology, and in particular relates to a method for monitoring and early warning of grassland caterpillar larvae based on UAV remote sensing technology. Background Technology

[0002] Grassland ecosystems are among the most widespread and important terrestrial ecosystem types globally. With increasing global climate change and human disturbance, grasslands are experiencing widespread degradation. The feeding habits of grassland caterpillars highly overlap with those of livestock. Large numbers of grassland caterpillars directly consume these larvae, exacerbating grassland degradation and intensifying the conflict between livestock and grassland ecosystems. Therefore, conducting large-scale monitoring of grassland caterpillar larvae and improving early warning mechanisms are of irreplaceable importance and practical significance for guiding livestock production and environmental protection.

[0003] Research on grassland caterpillar larvae has primarily focused on the species, damage levels, and food choices of larvae in specific study areas. Current methods for estimating grassland caterpillar larvae populations mainly rely on traditional ground surveys to investigate and statistically analyze larval numbers and distributions in individual areas. However, such methods are inefficient, require significant human and material resources, and are unsuitable for studying, forecasting, and controlling the spatiotemporal distribution patterns of grassland caterpillar larvae over large areas.

[0004] The existing technology for monitoring grassland caterpillar larvae has the following drawbacks:

[0005] 1. High cost and low efficiency: The traditional sampling method requires a lot of labor for sample collection and the work progress is slow, resulting in high cost for this method;

[0006] 2. Difficulty in long-term, fixed-point repeated monitoring: Due to the large workload and the damage to grassland caused by human trampling during the data survey process, it is difficult to monitor at a high frequency; because the survey personnel are different each time, it is difficult to determine the exact location of the previous survey, thus making it impossible to achieve long-term fixed-point monitoring of grassland caterpillar larvae.

[0007] 3. Poor timeliness and limited representativeness: Due to the low survey efficiency, the long time required for field operations and the short time available for surveys, it is impossible to conduct large-scale surveys of grassland caterpillar larvae in a short period of time.

[0008] 4. Difficulty in conducting surveys in specific environments: Because traditional sampling methods require on-site investigation and recording, it is difficult to conduct measurements in specific environments such as swamps, mountains, and areas restricted by human factors (e.g., privately owned ranches require permission, and there are dog attacks, etc.), which limits the sample size and representativeness. Summary of the Invention

[0009] The purpose of this invention is to provide a method for monitoring and early warning of grassland caterpillar larvae based on UAV remote sensing technology, so as to solve the problems of low efficiency, limited applicability, destructive sampling and high operating costs in traditional grassland caterpillar larvae monitoring.

[0010] To achieve the above objectives, this invention provides a method for monitoring and early warning of grassland caterpillar larvae based on unmanned aerial vehicle (UAV) remote sensing technology, comprising:

[0011] Acquire several aerial photographs of the target area sample plots, and use the aerial photographs acquired along each flight route as a monitoring unit;

[0012] An automatic identification model for grassland caterpillar larvae was constructed based on YOLOv5. The automatic identification model for grassland caterpillar larvae was trained based on training samples of grassland caterpillar larvae circled in several aerial photos. The samples were manually corrected and used for model training until the accuracy of the automatic identification result was higher than the preset value. The trained and corrected model was then used to identify and mark all aerial photos of the target area sample plot.

[0013] The markers in all aerial photographs of the target area were statistically analyzed to obtain the frequency of grassland caterpillar larvae in each monitoring unit. Based on the frequency of occurrence in each monitoring unit, the average density and uniformity of grassland caterpillar larvae were obtained.

[0014] Based on the average density, all monitoring units are divided into several hazard levels;

[0015] Using a niche model, key influencing factors are obtained and their current spatial distribution characteristics are characterized based on the distribution areas of each level of harm and the corresponding natural and anthropogenic impact datasets. Combined with future climate predictions, the occurrence areas of each level of disaster under future climate scenarios are realized to achieve early warning.

[0016] Optionally, the process of acquiring several aerial photos includes: selecting a target plot, identifying several aerial photography points within the target plot, setting a drone flight path based on the aerial photography points, and taking photos vertically downwards when the drone flies along the flight path to the aerial photography point.

[0017] Optionally, the aerial photography points are evenly distributed in a matrix within the target sample plot.

[0018] Optionally, the drone is equipped with a terrain following system and a high-resolution camera, and its flight altitude is fixed at 2m.

[0019] Optionally, the training standard for the grassland caterpillar larvae automatic identification model is an automatic identification accuracy rate of over 90%.

[0020] Optionally, the uniformity of grassland caterpillar larvae can be obtained by taking the maximum and minimum frequency values. The process includes: taking the difference between the maximum and minimum frequency values, taking the sum of the maximum and minimum frequency values, and taking the ratio of the difference to the sum as the uniformity.

[0021] The technical effects of this invention are as follows:

[0022] 1. Large-scale monitoring: By utilizing the flexibility and ease of operation of drones, along with corresponding aerial photography, recording, and statistical methods, large-scale spatial distribution monitoring of grassland caterpillar larvae has been achieved; effectively solving the problems of low efficiency, destructiveness, high operating costs, and difficulty in large-scale long-term monitoring in traditional grassland caterpillar larvae spatiotemporal distribution measurement.

[0023] 2. High efficiency and low cost: It eliminates the sample collection and walking sampling process in traditional methods, saving working time and manpower; the purchase of equipment and subsequent costs are low, and it can be reused for a long time, which greatly saves costs;

[0024] 3. Non-destructive: Non-destructive sampling is used to ensure the accuracy and representativeness of long-term fixed-point monitoring;

[0025] 4. Enables long-term, fixed-point, and environmentally specific monitoring: With the flexibility and mobility of drones, measurement work can be carried out in environments such as high altitude, swamps, and mountains, especially in areas with complex terrain, high altitude, and difficult working conditions;

[0026] 5. It is simple to operate, breaks through the limitations of traditional measurement methods, and is suitable for widespread application. Attached Figure Description

[0027] The accompanying drawings, which form part of this application, are used to provide a further understanding of this application. The illustrative embodiments and descriptions of this application are used to explain this application and do not constitute an undue limitation of this application. In the drawings:

[0028] Figure 1 This is a schematic diagram of the method flow in an embodiment of the present invention;

[0029] Figure 2 This is a schematic diagram of a drone acquiring low-altitude aerial photos over a grassland according to a set route in an embodiment of the present invention. Detailed Implementation

[0030] It should be noted that, unless otherwise specified, the embodiments and features described in this application can be combined with each other. This application will now be described in detail with reference to the accompanying drawings and embodiments.

[0031] It should be noted that the steps shown in the flowchart in the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions, and although a logical order is shown in the flowchart, in some cases the steps shown or described may be executed in a different order than that shown here.

[0032] Example 1

[0033] like Figure 1-2 As shown, this embodiment provides a method for monitoring and early warning of grassland caterpillar larvae based on UAV remote sensing technology, including:

[0034] S1: Acquire monitoring unit.

[0035] Specifically, step S1 involves taking several photos perpendicular to the ground within the target area, with each photo serving as a basic monitoring unit.

[0036] The process of taking several photographs perpendicular to the ground within the target sample plot includes:

[0037] First, select target monitoring area 01 and evenly distribute 16 waypoints. The drone can automatically take aerial photos one by one. Set the drone's flight path 02 so that the drone passes through the 16 aerial photo points along the prescribed route. When the drone reaches the aerial photo point, it should not be blown to the ground by the wind generated by the drone, affecting the observation accuracy, without affecting the caterpillar's state (e.g., lying on vegetation to feed). It should also be able to be clearly identified (the camera resolution is sufficient). In this embodiment, 2m is suitable according to the standard of DJI consumer-grade models. However, if the drone has a high resolution or zoom function, aerial photos can also be taken at different altitudes. The main considerations are terrain following function and camera resolution. Consumer-grade models are less expensive and facilitate large-scale collaborative monitoring. Lightweight drones or small drones equipped with RGB cameras can also be used.

[0038] In this embodiment, a large number of standardized samples are obtained by uniformly using aerial photography to extract samples with uniform standards.

[0039] Photos are taken vertically downwards at a height of 2 meters above the ground. The flight path covers a 40m x 40m target plot 03, with aerial photography points evenly distributed in a matrix pattern within the target plot. The set flight path can be repeated, and the flight altitude remains at 2 meters above the ground. Fixed-point aerial photographs taken by the UAV can be obtained through repeated shooting at fixed locations determined by the control system. The target monitoring area should be a larger area requiring monitoring and early warning, while the target plot refers to the area covered by a single flight path.

[0040] Because the drone takes photos vertically downwards from a constant altitude, the area captured at each aerial photography point is essentially the same. The drone is equipped with a terrain-following system, ensuring a consistent vertical distance from the ground and that the lens remains perpendicular to the shooting surface to guarantee consistent area in each photo. This ensures data comparability and consistency between the area captured in each photo and the total monitored area. It also prevents damage to the drone due to terrain or fences. The drone is equipped with a camera with at least 20 megapixels for taking photos.

[0041] S2: Automatic identification and statistics.

[0042] Specifically, based on a large number of standardized aerial photographs obtained in the early stage, training samples of grassland caterpillar larvae were acquired and manual correction methods were used to obtain an automatic identification model of grassland caterpillar larvae based on YOLOv5, ensuring that the accuracy of automatic identification is greater than 90%, which is used to automatically identify and count the frequency of grassland caterpillar larvae in each monitoring unit.

[0043] S3: Based on the frequency of grassland caterpillar larvae in each monitoring unit, the average density and uniformity of grassland caterpillar larvae are calculated.

[0044] Specifically, based on the occurrence of grassland caterpillar larvae recorded in step S2, the frequency of grassland caterpillar larvae is statistically obtained, and the average density and uniformity of grassland caterpillar larvae are obtained based on the frequency of occurrence in each monitoring unit.

[0045] The process of statistically determining the frequency of occurrence of each species and then obtaining the average density and evenness of grassland caterpillar larvae includes:

[0046] S51: Get the number of times the grassland caterpillar larvae are marked in each photo, i.e., the frequency of a single photo.

[0047] S52: Obtain the total number of times grassland caterpillar larvae were marked, i.e., the total frequency of the monitoring unit's appearance;

[0048] S53: Using the data obtained in step S51 and step S52, obtain the average density and uniformity of grassland caterpillar larvae.

[0049] In step S5, the average density and uniformity of grassland caterpillar larvae are calculated using the following formula:

[0050]

[0051] E = (x max -x min ) / (x max +x min )

[0052] In the above formula, x iM represents the frequency of grassland caterpillars appearing in each aerial photograph obtained in step S33. M represents the average density of grassland caterpillar larvae; E represents the uniformity of grassland caterpillar larvae.

[0053] The basic monitoring units are divided into n hazard levels based on the average density, where n≤5.

[0054] By utilizing the niche model, key influencing factors are extracted and their current spatial distribution characteristics are characterized based on the distribution areas of each level of harm and the corresponding natural and anthropogenic impact datasets. Combined with future climate and anthropogenic impacts, the occurrence areas of each level of disaster under future climate scenarios are predicted, thereby achieving high-precision and operable early warning.

[0055] Using the BIOMOD model, spatial distribution data of grassland caterpillars in the study area (which can be classified into levels or simply data on their presence or absence) is input along with biometeorological data, topographic data, and human disturbance data from different time periods (such as the present or a future time). The model can then depict the spatiotemporal distribution pattern of grassland caterpillars.

[0056] This invention replaces traditional ground survey methods by extracting the density and evenness of caterpillar larvae from 16 photographs. Utilizing a drone's GPS positioning device, it enables long-term, fixed-point monitoring of observation points. This method requires low fieldwork intensity, has high sampling efficiency, and can acquire a large number of samples in a short time, making it suitable for large-scale field species surveys and monitoring. Furthermore, the data can be stored long-term and reviewed regularly. This non-destructive method allows for long-term, high-frequency monitoring and effectively overcomes the limitations of traditional measurement methods, making it suitable for widespread application. Based on the obtained average density, the current caterpillar larvae damage level is classified, key data for each level's corresponding area are obtained, and a niche model is used to obtain the spatiotemporal distribution pattern. Combined with climate and other information, this enables pest early warning, reducing pest control costs. It plays an irreplaceable and important role in guiding livestock production and environmental protection, maintaining the stability of grassland ecosystems.

[0057] Those skilled in the art will also understand that the various illustrative logic blocks, modules, circuits, and algorithm steps described in conjunction with the embodiments herein can be implemented as electronic hardware, computer software, or a combination thereof. To clearly illustrate the interchangeability between hardware and software, the various illustrative components, blocks, modules, circuits, and steps described above are generally described in terms of their functionality. Whether such functionality is implemented as hardware or software depends on the specific application and the design constraints imposed on the overall system. Those skilled in the art can implement the described functionality in alternative ways for each specific application; however, such implementation decisions should not be construed as departing from the scope of this disclosure.

[0058] The above description is merely a preferred embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A method for monitoring and early warning of grassland caterpillar larvae based on unmanned aerial vehicle (UAV) remote sensing technology, characterized in that, Includes the following steps: Acquire several aerial photographs of the target area sample plots, and use the aerial photographs acquired along each flight route as a monitoring unit; An automatic identification model for grassland caterpillar larvae was constructed based on YOLOv5. The automatic identification model for grassland caterpillar larvae was trained on training samples of grassland caterpillar larvae circled in several aerial photos until the accuracy of the automatic identification result was higher than the preset value. The trained and corrected model was then used to identify and mark all aerial photos of the target area sample plot. The markers in all aerial photographs of the target area were statistically analyzed to obtain the frequency of grassland caterpillar larvae in each monitoring unit. Based on the frequency of occurrence in each monitoring unit, the average density and uniformity of grassland caterpillar larvae were obtained. Based on the average density, all monitoring units are divided into several hazard levels; Using a niche model, key influencing factors are obtained and their current spatial distribution characteristics are characterized based on the distribution areas of each level of harm and the corresponding natural and anthropogenic impact datasets of the distribution areas. Combined with future climate predictions, the occurrence areas of each level of disaster under future climate scenarios are realized to achieve early warning. The uniformity of grassland caterpillar larvae is obtained by taking the maximum and minimum frequency values. The process includes: taking the difference between the maximum and minimum frequency values, taking the sum of the maximum and minimum frequency values, and taking the ratio of the difference to the sum as the uniformity.

2. The method for monitoring and early warning of grassland caterpillar larvae based on UAV remote sensing technology according to claim 1, characterized in that, The process of acquiring several aerial photos includes: selecting a target plot, identifying several aerial photography points within the target plot, setting a drone flight path based on the aerial photography points, and taking photos vertically downwards when the drone flies along the flight path to the aerial photography point.

3. The grassland caterpillar larvae monitoring and early warning method based on UAV remote sensing technology according to claim 2, characterized in that, The aerial photography points are evenly distributed in a matrix within the target sample plot.

4. The grassland caterpillar larvae monitoring and early warning method based on UAV remote sensing technology according to claim 1, characterized in that, The drone is equipped with a terrain-following system and a high-resolution camera, and its flight altitude is fixed at 2m.

5. The method for monitoring and early warning of grassland caterpillar larvae based on UAV remote sensing technology according to claim 1, characterized in that, The training standard for the automatic identification model of grassland caterpillar larvae is that the automatic identification accuracy exceeds 90%.

Images