A method for detecting the number of aphids in a green manure orchard ecosystem by multi-dimensional sampling

Abstract

The application discloses a kind of multi-dimension sampled green manure orchard ecosystem aphid quantity detection methods, it is related to green manure orchard aphid quantity detection technical field, including: by green manure orchard ecosystem is divided into fruit tree crown wood layer, surface layer and ground root layer according to vertical structure, and corresponding detection method is used to different levels respectively;While starting hierarchical collection sample and aphid identification and quantity detection in daily preset time period;In fruit tree crown wood layer, orchard is divided into several units and generates flight path collection leaf image according to rules, in surface layer, data is collected by gridding, in ground root layer, focus root dense area is distributed sampling point according to preset interval.It realizes the multi-dimension accurate detection of green manure orchard aphid, effectively improves the comprehensiveness and accuracy of aphid quantity statistics, and provides reliable data support for orchard aphid scientific control.

Description

TECHNICAL FIELD

[0001] The present application belongs to the technical field of green manure orchard aphid quantity detection, and relates to a multi-dimensional sampling green manure orchard ecosystem aphid quantity detection method. BACKGROUND

[0002] The green manure orchard ecosystem refers to a sustainable orchard management mode in which specific green manure crops are planted between rows or under trees in an orchard to improve soil fertility and enhance ecological regulation function. The microenvironment formed by green manure plants and fruit trees in this system provides suitable habitat and breeding conditions for aphids. Aphids directly affect the growth and development of fruit trees by sucking their sap, so accurate monitoring of their population size is an important foundation for early warning and green prevention and control of pests.

[0003] At present, although certain progress has been made in the existing aphid quantity detection technology applied to the green manure orchard ecosystem, there are still the following shortcomings: first, aphids at different growth stages are distributed in different areas of the green manure orchard. The existing technology for monitoring the green manure orchard ecosystem mainly focuses on single-dimensional detection, usually limited to observation of nymphs or adults above the ground surface, and cannot completely cover the distribution range of aphids, resulting in obvious deviation of the detection results from the true number of aphids in the orchard.

[0004] Secondly, the behavior and body temperature of aphids are deeply affected by the daily variation cycle of environmental factors, showing obvious circadian rhythm. However, the existing technology is mainly for stage or single detection, and cannot realize continuous and multi-time point dynamic monitoring, so it cannot capture the activity rules and number fluctuations of aphids at different times. The data obtained can only reflect the pest population at a certain moment and in a certain place, and it is difficult to construct the time series trend of aphid population size, which affects the timeliness and accuracy of early warning of pests.

[0005] In addition, the existing method mainly relies on random sampling of a small number of leaves or setting limited monitoring points on the ground, and the sampling strategy lacks systematicness and representativeness. Due to the limited coverage of samples, it is difficult to fully reflect the real spatial distribution pattern of aphids in the orchard, and local high-risk outbreak areas may be missed. As a result, the overall occurrence pattern of aphids cannot be accurately reflected, leading to lack of pertinence of prevention and control measures, which may delay the management opportunity of key areas. SUMMARY

[0006] In view of the above problems in the background art, the present application provides a multi-dimensional sampling green manure orchard ecosystem aphid quantity detection method.

[0007] The object of the present application can be achieved by the following technical solution: a multi-dimensional sampling green manure orchard ecosystem aphid quantity detection method, comprising: S1. Dividing the green manure orchard ecosystem according to the vertical structure into a fruit tree canopy layer, a ground surface layer and a ground bottom root layer, and starting hierarchical sampling in a daily preset time period.

[0008] S2. Collecting leaf images for the fruit tree canopy layer, identifying aphid secretion spectrum characteristics in the images, identifying insect spot distribution areas, counting insect spot distribution density in a unit canopy volume, and calculating the total amount of aphids in the layer.

[0009] S3. Collecting temperature distribution data of the ground surface layer by gridding, identifying the ground surface area with a temperature value higher than a preset temperature difference threshold as an aphid activity hot spot area, calculating the coverage rate of the hot spot area to a unit ground surface area to obtain the total amount of aphids in the layer.

[0010] S4. Drilling soil samples in the root dense area of the ground bottom root layer, collecting soil particle surface images after hierarchical dissociation of the samples, identifying the number of aphid nymphs based on their morphological characteristics, and calculating the total amount of aphids in the layer by combining the sampling volume and the total volume of soil.

[0011] S5. Superimposing the aphid quantities calculated by the fruit tree canopy layer, the ground surface layer and the ground bottom root layer to generate a green manure orchard aphid quantity representing the vertical structure of the orchard.

[0012] Compared with the prior art, the present application has the following advantages: (1) The present application divides the green manure orchard ecosystem into a fruit tree canopy layer, a ground surface layer and a ground bottom root layer according to the vertical structure, and uses corresponding detection methods for different layers, covering the complete distribution range of aphids in different growth stages in the orchard, fundamentally overcoming the limitation of single dimension of existing monitoring means, and can truly and completely reflect the actual population base of aphids in the orchard, significantly eliminating the estimation bias caused by incomplete monitoring range.

[0013] (2) The present application starts hierarchical sampling and aphid identification and quantity detection in a daily preset time period, effectively eliminating the interference of environmental factors such as day and night temperature difference and light change on the monitoring results, and significantly improving the timeliness and accuracy of early warning.

[0014] (3) The present application divides the orchard into several units in the fruit tree canopy layer and generates flight paths according to rules to collect leaf images, collects data by gridding in the ground surface layer, and focuses on the root dense area in the ground bottom root layer to arrange sampling points at a preset interval, forming a systematic and full-coverage sampling strategy, avoiding the defects of random sampling and greatly reducing the probability of missing high-risk hot spot areas of local outbreak, avoiding delay of prevention and control opportunity and waste of resources. BRIEF DESCRIPTION OF DRAWINGS

[0015] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings needed for the description of the embodiments will be briefly introduced as follows. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative effort on the basis of these drawings.

[0016] Figure 1 The method embodiment of the present application is implemented by the step diagram.

[0017] Figure 2 The leaf image acquisition flowchart of the crown layer of the present application is shown in the figure.

[0018] Figure 3 The leaf image insect spot identification flowchart of the present application is shown in the figure. DETAILED DESCRIPTION

[0019] The technical solutions in the embodiments of the present application will be described clearly and completely in the following with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only some embodiments of the present application, but not all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative effort belong to the protection scope of the present application.

[0020] Please refer to Figure 1 As shown in the figure, the present application provides a multi-dimensional sampling green manure orchard ecosystem aphid quantity detection method, which comprises: S1. Dividing the green manure orchard ecosystem into a fruit tree crown layer, a ground surface layer and a ground bottom root layer according to the vertical structure, and starting hierarchical sample collection in a daily preset time period.

[0021] The green manure orchard ecosystem is divided into a fruit tree crown layer, a ground surface layer and a ground bottom root layer according to the vertical structure, and the specific implementation is as follows: taking the trunk of the fruit tree as the reference to measure the vertical distance from the trunk branch point to the ground, and defining the area from the trunk branch point to the top of the tree crown as the fruit tree crown layer.

[0022] Specifically, the trunk branch point formed by the first bifurcation of the trunk of the fruit tree is taken as the vertical reference point P1, the height interval [P1, P2] is formed by extending from the P1 point to the vertex P2 of the newly grown branches and leaves at the top of the tree crown, and the maximum envelope body of all the woody branches growing outward is covered around the trunk in three-dimensional space.

[0023] It should be noted that this layer contains the main feeding sites of aphids, such as young leaves, new shoots and honeydew secretion attachment areas, and adult aphids mainly move in the crown layer above the branch point of the trunk in the area with dense leaves and sufficient light. It is the core area of the concentrated growth of fruit tree leaves and the vigorous photosynthesis, and is also the main place for aphid adults to feed and reproduce. According to the branch point of the trunk to the top of the crown, the main habitat and activity range of aphid adults can be accurately covered, laying a foundation for subsequent targeted detection of the number of aphids in this layer.

[0024] With the ground as the horizontal reference surface, the space region extending upward from the ground to below the branch point of the trunk is defined as the ground surface layer.

[0025] Specifically, the root point P3 at the junction of the main trunk base and the ground surface is taken as the plane coordinate origin, and P3 is taken as the center to extend outward to the vertical projection outer edge boundary B1 of the crown, and extend upward from the ground interface S0 to the top height H1 of the green manure crop canopy.

[0026] It should be noted that the green manure vegetation community, the bare ground surface and the soil and atmosphere interface are the transition zone for aphid migration and natural enemy activity, and the ground surface layer is the core area for short-distance migration, temporary habitat and feeding of green manure crops. The space extending upward from the ground to below the branch point of the trunk covers not only the green manure growth area of the orchard ground surface, but also the low vegetation area near the base of the trunk. This division method can completely enclose the main activity space of aphids in the ground surface and near ground surface area.

[0027] With the ground as the horizontal reference surface, the vertical downward extension to the lower limit of the main root distribution depth of the fruit tree forms a three-dimensional soil space below the reference point, which is defined as the ground bottom root layer.

[0028] Specifically, the root point P3 of the ground surface layer is taken as the starting point, and P3 is vertically extended downward to the lower limit depth D1 of the dense root zone of the fruit tree, and the main trunk is taken as the center to cover the maximum horizontal distribution radius R1 of the root system.

[0029] It should be noted that this layer contains the overwintering place of aphid nymphs and the root exudate enrichment area. Aphid nymphs have the habit of underground habitat and feeding, and mainly gather in the soil around the roots of fruit trees to suck root juice for nutrition. The division method of vertically extending downward to the lower limit of the main root distribution depth of the fruit tree can completely enclose the soil space where the roots grow densely. This space is the key area for aphid nymphs to obtain food and avoid external interference.

[0030] The specific implementation process of starting hierarchical sampling in the preset time period of each day is as follows: in the early stage, the active peak period of each level of aphids is recorded through 1-2 weeks of orchard observation, combined with the best working period of the detection equipment, to determine and fix the active peak period of each level of aphids as the preset sampling time period each week. According to seasonal changes, such as advancing in summer and delaying in winter, the time period is fine-tuned by one to two hours to ensure adaptability. All data in the sampling process are stored according to level, unit, and time stamp, and the original data of each level are automatically summarized after the preset time period ends, to prepare for subsequent aphid identification and quantity calculation.

[0031] It should be noted that the preset time period of each day is combined with the activity rule of aphids and the technical requirement of detection, and is preferably 9:00-11:00 in the morning or 15:00-17:00 in the afternoon. Taking a temperate orchard as an example, the time period can be fine-tuned according to latitude and season, such as advancing to 8:30-10:30 in summer in the north and delaying to 10:00-12:00 in winter. The temperature in this time period is moderate, and the activity of aphids is vigorous. The adult aphids in the canopy layer frequently feed on leaves, and the characteristics of aphid spots are clear. The ground layer is in the transition stage, and the aphids have a wide range of activities, and are easy to form a temperature hotspot area. The nymphs in the ground root layer are also in the active feeding period, and the probability of occurrence in the soil sample is high. The data collected at this time can truly reflect the actual quantity and distribution of aphid population, and avoid the low quantity caused by aphid dormancy.

[0032] In the life cycle of aphids, adult aphids have phototaxis and mainly inhabit the fruit tree canopy layer with dense leaves and sufficient light to feed on leaf juice. The nymphs in the transition stage to adult aphids are active on the ground and in the near-ground space, relying on the ground green manure crops for feeding and avoiding natural enemies. The low-age nymphs in the ground layer prefer to suck the juice of fruit tree roots and are concentrated in the ground root layer. According to the vertical structure, the core distribution area of each growth stage of aphids can be accurately covered, and detection omission can be avoided.

[0033] S2. Collecting leaf images of the fruit tree canopy layer, identifying the spectral characteristics of aphid secretions in the images, identifying the distribution area of aphid spots, counting the distribution density of aphid spots in the unit canopy volume, and calculating the total quantity of aphids in the layer.

[0034] Reference Figure 2 As shown, the specific process of collecting leaf images of the fruit tree canopy layer is as follows: dividing the green manure orchard into a plurality of units, and taking a reference sample fruit tree as the center of each unit to define the area range including a certain number of surrounding fruit trees.

[0035] Specifically, taking the base of the trunk of each reference sample fruit tree as the geometric center point, extending a preset distance in the horizontal direction, usually 1.5-2 times the distance between adjacent fruit trees, to construct a square sampling unit, to ensure that the unit covers the complete canopy layer of the center fruit tree and its surrounding adjacent fruit trees.

[0036] It should be noted that this division method can make the fruit trees in each unit consistent in age, growth state, and green manure crop accompanying conditions, avoiding sampling bias caused by the special growth conditions of single fruit trees. At the same time, the unit division can cover different areas of the orchard, ensuring that the collected leaf images and aphid quantity data can represent the overall situation of the orchard, rather than the accidental results of a special local area.

[0037] With the orchard operation entrance as the starting point of the path, the planar rectangular position coordinates of all unit center points are obtained, and the straight line distances of each center point from the starting point are calculated.

[0038] A unique digital identification code is assigned to each unit, and the identification code sequence is rearranged from near to far according to the straight line distance. The flight path is generated by connecting the center points in the new order.

[0039] Specifically, a planar rectangular coordinate system with the orchard operation entrance as the origin (0, 0) is established, the coordinates of the center points of each collection unit are imported into the coordinate system, the straight line distances of each point from the origin are calculated, the identification code sequence of the collection unit is rearranged in ascending order of distance, and the initial flight polyline is generated by connecting the center points in sequence. The cross product method is used to detect the intersection of adjacent path segments. If there is an intersection, the identification code sequence of the end point is exchanged until the intersection is eliminated.

[0040] Arranging the identification codes from near to far according to the straight line distances of the unit center points from the starting point and generating the path can make the UAV cover all units in order from near to far, avoiding path intersection, U-turn or long-distance jumping, such as flying to a distant unit first and then flying back to a near unit, and minimizing the total flight distance.

[0041] The UAV carrying a multispectral imager adjusts the hovering height when it reaches the target unit center point along the flight path, and collects leaf images of the fruit tree canopy layer at multiple angles.

[0042] Specifically, when flying to the target collection unit center point, the hovering height is automatically adjusted according to the reference sample fruit tree canopy layer marker points bound to this unit, satisfying the minimum canopy lower boundary height to the maximum canopy upper boundary height. The gimbal rotation mechanism is started for multi-angle scanning, covering the azimuth and pitch angle range, and the multispectral sensor in the spectral imager is used to simultaneously collect leaf images in the visible light and near-infrared bands.

[0043] It should be noted that multispectral imaging can capture the reflection characteristics of leaves in non-visible light bands such as red edge and near-infrared, enhancing the recognition of aphid damage spots, honeydew and other traces. Multi-angle rotation scanning avoids missing detection caused by leaf obstruction, especially for dense canopies, and close-range collection ensures that the image resolution meets the detection needs of small aphids.

[0044] Reference Figure 3As shown, the specific steps for identifying the aphid secretion spectrum feature in the image are as follows: screening the clear leaf image without obstruction and reflection from the collected leaf image, and classifying and arranging according to the unit and fruit tree number.

[0045] It should be noted that the interference factors such as obstruction and reflection are excluded to ensure that the subsequent extracted spectrum information is a true reflection of the leaf and its attached objects, and to avoid noise introduction.

[0046] The full-surface spectrum information of each leaf image is compared with the preset healthy leaf spectrum reference and aphid secretion spectrum reference respectively.

[0047] Specifically, in the green manure orchard to be detected, healthy fruit trees with the same tree age, growth state and reference sample fruit trees in the orchard are selected, and healthy leaves without damage, spots and insect bite marks are randomly picked up from different directions and different heights of the crown layer of each healthy sample tree, such as the upper layer, the middle layer and the lower layer, to ensure that the samples cover each area of the crown layer and avoid accidental deviation of single position samples. The spectral reflectance curve data of all healthy leaves are preprocessed, and abnormal data caused by operation errors are removed, such as obvious jump and peak abnormality of spectral curve. The average value of spectral reflectance corresponding to each group of wavelengths in the remaining effective data is calculated to generate the average spectral reflectance curve of healthy leaves, and the standard deviation of each wavelength reflectance is calculated to determine the allowable range of reflectance fluctuation. The average spectral reflectance curve and the allowable range of fluctuation are used as the preset healthy leaf spectrum reference.

[0048] In the green manure orchard to be detected, the fruit trees with mild aphid infestation are selected, and the aphid secretion on the leaf back and the top of the tender branch is wiped, the secretion on the cotton swab is transferred to a sterile centrifuge tube, and the aphid population consistent with the aphid species in the orchard is cultured in the laboratory. The aphids are placed on healthy leaves in sterile culture, and the secretion is collected in the same way after the aphids secrete honeydew.

[0049] The collected aphid secretion samples are filtered through a microporous filter to remove impurities such as aphid carcasses and leaf debris that may be mixed in the samples, and the filtered pure secretion forms a stable secretion film. The pixel area where the spectral reflectance curve deviates from the healthy leaf spectrum reference and the spectral reflectance curve falls into the preset matching interval of the aphid secretion spectrum reference is marked.

[0050] The spectral reflectance curve data of the aphid secretions is subtracted from the background spectral data of the blank quartz slide to obtain pure aphid secretion spectral data. After removing abnormal data, the average spectral reflectance curve of the remaining data is calculated, and the characteristic peak value in the curve is determined, such as the characteristic reflection peak and valley of the aphid secretions near 550 nm. Based on the fluctuation range of the average curve, a preset matching interval of the spectral reflectance curve is set. The average spectral reflectance curve, the characteristic peak, and the preset matching interval are used as the preset aphid secretion spectral reference.

[0051] The marked pixel region is subjected to boundary continuity analysis, and adjacent pixels are merged to form an independent contour region, and the contour broken or fragmented region is screened out.

[0052] Specifically, the marked candidate pixel points are subjected to boundary continuity analysis, and a region growing algorithm is used to take a single candidate pixel as a starting point, and adjacent candidate pixels with consistent spectral characteristics are merged to form an independent contour region. A contour area threshold is set to automatically screen out contour broken or fragmented regions with an area less than the threshold, and the complete contour region is retained.

[0053] If the regional spectral reflection peak position coincides with the typical spectral reflection peak of the leaf disease, it is determined to be a disease interference region.

[0054] It should be noted that different types of leaf diseases, such as anthracnose and leaf spot disease, can cause specific changes in the internal tissue structure of the leaf, such as chlorophyll content and cell arrangement. Such changes will form a typical spectral reflection peak in the spectral reflection characteristics. For example, anthracnose can cause local necrosis of the leaf, and the necrotic area will lose chlorophyll, resulting in a significant reflection peak in the wavelength range of 650-680 nm. Leaf spot disease will cause damage to the epidermal cells of the leaf, forming a characteristic reflection peak in the wavelength range of 520-550 nm.

[0055] The spectral reflection peak of the aphid secretion covered insect spot area is derived from the chemical composition of the secretion itself, such as sugar and amino acids, and the characteristic wavelength is essentially different from the disease reflection peak. Therefore, when the spectral reflection peak position of the region to be identified coincides with the typical reflection peak position of a certain type of leaf disease, it indicates that the spectral characteristics of the region are caused by the disease rather than the aphid secretions.

[0056] If the regional spectral absorption valley position coincides with the spectral absorption valley of the stain, it is determined to be an environmental interference region.

[0057] It should be noted that the environmental stains on the surface of the leaves of the orchard, such as soil, pesticide residues, and dust accumulation, have specific chemical compositions, and these compositions can strongly absorb light at specific wavelengths to form typical spectral absorption valleys. For example, the iron oxide contained in the soil can form a significant absorption valley in the wavelength range of 800-850 nm, and the organic phosphorus component in the pesticide residues can form a characteristic absorption valley in the wavelength range of 900-920 nm.

[0058] The spectral absorption valley of aphid secretions is determined by its own chemical structure, and the wavelength position of the stain absorption valley is significantly different. Therefore, when the spectral absorption valley position of the region to be identified coincides with the typical absorption valley position of a certain type of stain, it indicates that the spectral characteristics of the region are caused by environmental stains rather than aphid secretions.

[0059] The remaining independent contour region that is not excluded is marked as the final insect spot distribution region.

[0060] The insect spot distribution density in the statistical unit canopy volume is calculated as follows: the total number of pixels of all leaf images in the unit that are marked as insect spot distribution regions is accumulated.

[0061] The total number of pixels is converted to the actual total insect spot coverage area based on the spatial resolution of the multispectral imager.

[0062] The multispectral imager device parameters currently used for fruit tree canopy layer leaf image acquisition are retrieved, and the spatial resolution is extracted and recorded. The actual area corresponding to a single pixel is calculated based on the spatial resolution. The total number of pixels in all insect spot contour regions is accumulated through the image pixel counting function of the detection system and recorded as the total number of insect spot pixels. The actual total insect spot coverage area is obtained by multiplying the total number of insect spot pixels by the actual area of a single pixel.

[0063] The canopy layer vertical height data of the reference sample fruit tree is obtained, and the canopy layer vertical height of the corresponding reference sample is matched based on the current unit center point position.

[0064] It should be noted that the selection of the reference sample fruit tree follows the principle of average growth level. The matching of the corresponding reference sample height through the center point position of each unit can make the canopy volume calculation of all units based on a unified height reference standard, avoiding the calculation caliber differences caused by different canopy height data acquisition methods among units.

[0065] The canopy volume is calculated based on the matched canopy layer vertical height and the projection range of the unit geographical boundary as the bottom surface.

[0066] The total area of the unit insect spots is compared with the volume of the unit canopy to output the insect spot distribution density value per unit canopy volume.

[0067] The mapping relationship parameter between the unit area of the insect spot and the number of aphids is called, and the generated insect spot distribution density value is converted into the number of aphids in the unit crown volume.

[0068] It should be noted that the mapping relationship parameter is obtained by collecting long-term aphid monitoring historical data of green manure orchard, which contains information of insect spot area and aphid number in different periods and different environmental conditions, and using data analysis methods such as regression analysis to explore the correlation between the unit area of the insect spot and the number of aphids.

[0069] The number of aphids in the corresponding unit is obtained by multiplying the number of aphids in the corresponding unit by the total volume of the crown layer.

[0070] After traversing all the units, the estimated number of aphids is accumulated to generate the total amount of aphids in the fruit tree crown layer.

[0071] S3. Collecting ground layer temperature distribution data by gridding, identifying the ground area with temperature value higher than the preset temperature difference threshold as the aphid activity hotspot area, and calculating the coverage rate of the hotspot area per unit ground area to obtain the total amount of aphids in this layer.

[0072] The specific steps for collecting ground layer temperature distribution data by gridding are as follows: dividing the ground layer plane into a plurality of independent grid units according to a preset area threshold, and setting a fixed infrared thermal imager in each grid unit.

[0073] It should be noted that the preset area threshold is combined with the green manure vegetation distribution density of the green manure orchard, the distance between fruit trees and the monitoring range of the infrared thermal imager.

[0074] Specifically, a plane rectangular coordinate system is established with the geometric center of the orchard boundary contour as the coordinate origin (0, 0), the x-axis is parallel to the boundary line where the orchard operation entrance is located, and the y-axis is perpendicular to the x-axis.

[0075] According to the preset area threshold and the coordinate system parameters, the side length of the grid unit is calculated, and the grid is divided along the x and y axis directions according to the side length size in turn: starting from the origin (0, 0), a vertical line segment is marked every preset area side length along the positive direction of the x-axis, and a horizontal line segment is marked every preset area side length along the positive direction of the y-axis, the horizontal and vertical line segments intersect to form a plurality of square independent grid units with a preset area, and each grid unit is assigned an initial number. If 2-3 adjacent grid units are in the area with sparse green manure and no fruit tree shade, they can be combined into a larger grid unit, and if a single grid unit contains multiple fruit trees or has a large area of fruit tree shade, the grid unit can be split into 2 smaller grid units to ensure that the temperature monitoring in each unit is not disturbed by the complex environment.

[0076] A plane rectangular coordinate system covering all grid units is established with the geometric center of the orchard boundary contour as the reference point, and the coordinate data of the center point of each grid unit is recorded.

[0077] According to the difference in the distribution density of green manure vegetation in the orchard and the projection shading range of the fruit trees, the grid units are combined and adjusted to form combined collection areas with different areas.

[0078] Each combined collection area is assigned a device identification code, and the infrared thermal imagers in each area are started in sequence according to the identification code.

[0079] All infrared thermal imagers are triggered synchronously within a preset time period each day, and the infrared temperature sensors of the infrared thermal imagers are used to collect ground temperature data to generate temperature distribution maps of each combined collection area.

[0080] The temperature distribution maps are spliced and integrated according to coordinate data to generate a temperature distribution data set covering the entire ground surface layer of the orchard.

[0081] The coverage rate of the hot spot area per unit ground surface area is calculated to obtain the total amount of aphids in the layer, and the specific content is as follows: the actual ground surface area of each hot spot area is accumulated to obtain the total area of the hot spot area, and the actual monitoring total area of the corresponding ground surface layer is obtained.

[0082] The ratio of the total area of the hot spot area to the actual monitoring total area is calculated to obtain the hot spot area coverage rate.

[0083] The pre-established corresponding relationship parameters of the number of aphids per unit area of the hot spot area are retrieved.

[0084] It should be noted that different environmental conditions are simulated in the laboratory or experimental field, and different densities of aphid populations are set to observe the reproduction, survival and distribution of aphids per unit area. By controlling the variable experiment, the influence of different factors on the number of aphids is analyzed, and the relationship between the number of aphids per unit area and environmental factors is established.

[0085] The hot spot area coverage rate value is multiplied by the number of aphids per unit area of the hot spot area parameter.

[0086] The operation result is multiplied by the actual monitoring total area to obtain the total amount of aphids in the ground surface layer.

[0087] S4. Drill soil samples in the root dense area of the subterranean root layer, collect soil particle surface images after layering and dissociation of the samples, identify the number of aphid nymphs based on their morphological characteristics, and calculate the total amount of aphids in the layer by combining the sampling volume and the total volume of soil.

[0088] The specific steps for drilling soil samples in the root dense area of the subterranean root layer are as follows: a root core distribution area is delineated around the trunk of the fruit tree, and a series of sampling points are arranged along the outer contour line at a preset interval distance based on the boundary of the root core distribution area.

[0089] It should be noted that the aphid nymphs rely on the root system of fruit trees for survival, and mainly gather in the root system intensive area around the main stem of fruit trees, i.e. the root system core distribution area. The sampling points are arranged based on the boundary of the area to ensure that the sampling range accurately covers the main habitat and feeding area of the nymphs, avoids missing detection due to deviation of the sampling range from the root system intensive area, and ensures that the number of nymphs in the subsequent soil samples truly reflects the distribution of aphids in the root system layer of the ground.

[0090] The preset interval distance is determined by referring to the verified effective preset interval distance of the historical data of sampling and detection of the root system layer of the orchard of the same fruit tree variety and similar green manure planting mode.

[0091] The soil drilling equipment is used to drill soil column samples vertically downward at each sampling point in turn, and the drilling depth and geographic coordinates of each sampling point are recorded.

[0092] The obtained continuous soil column samples are transversely cut at the preset depth interval to separate them into a plurality of independent layered soil samples.

[0093] It should be noted that the preset depth interval is determined by referring to the depth interval range recommended in the relevant technical standards for sampling of the root system layer of the green manure orchard, and combining the maximum cutting accuracy of the soil drilling equipment, under the premise of the standard range and the accuracy allowed by the equipment.

[0094] The dispersed soil particle surface is imaged using a microscopic imaging device.

[0095] It should be noted that the aphid nymphs are extremely small in size and inhabit the interstitial space between soil particles, and their morphological details such as antennae structure and abdominal stripes cannot be clearly captured by the naked eye or ordinary imaging equipment. The microscopic imaging device has high magnification and high resolution, and can clearly present the morphological characteristics of the nymphs on the surface of the soil particles.

[0096] The specific content of identifying the number of aphid nymphs based on their morphological characteristics is as follows: the particle surface images of each layered soil sample are sequentially retrieved, and the contour objects identified in the images are judged according to a preset biological morphological primary screening threshold, and the objects meeting the threshold are determined as suspected targets.

[0097] Specifically, core morphological parameters for distinguishing between the nymphs and the interference objects are extracted from the sample images, including: contour area, aspect ratio, i.e. the ratio of body length to body width, and contour circularity.

[0098] The extracted parameters are statistically analyzed to calculate the mean value ± 2 times the standard deviation of each parameter of the aphid nymphs, determine the reasonable distribution interval of the nymph morphological parameters, and mark the overlap range of the interference object parameters and the nymph interval to clearly define the parameter critical value that can distinguish between the two.

[0099] Based on the distribution interval of the nymph shape parameter, combined with the interference material exclusion requirement, the preliminary screening threshold is set: the area threshold value excludes too small or too large particle fragments, the length-width ratio threshold value excludes irregular interference materials, and the roundness threshold value excludes too sharp or round interference materials.

[0100] A large number of group number mixed soil sample images containing nymphs and interference materials are selected, the set threshold is applied for screening, the correct retention nymph rate and the error retention interference material rate are counted, if the threshold is not adjusted, the verification is repeated until the detection accuracy requirement is met, and finally the preset biological shape preliminary screening threshold is determined.

[0101] The suspected target morphological characteristics are compared with the pre-stored typical morphological parameter set of aphid nymphs.

[0102] If the suspected target characteristics meet the parameter set defined combination, the nymph individual is determined and marked for counting.

[0103] After traversing all the images, the marked counting results are accumulated to obtain the total number of nymphs in the sample.

[0104] According to the drilling equipment parameters, the standard cross-sectional area of the soil column sample is obtained, combined with the layered cutting depth, the volume fraction of each independent layered sample is calculated, and the total sampling volume is output by accumulating and summing each volume fraction.

[0105] The obtained total number of nymphs in the sample and the total sampling volume are subjected to division operation to obtain the nymph density basis of unit volume of soil.

[0106] The specific content of calculating the total amount of aphids in the layer combined with the sampling volume and the total volume of soil is as follows: the three-dimensional space boundary data of the root layer of the ground is obtained, the root layer depth value of each sample is extracted, the arithmetic mean value is calculated as the depth, and the horizontal projection area of the ground surface layer is obtained. The depth and the horizontal projection area are multiplied to obtain the total volume of soil.

[0107] The nymph density basis and the total volume of soil are subjected to multiplication operation, and the total amount of aphids in the root layer of the ground is output.

[0108] S5. Superimpose the calculated number of aphids of the fruit tree canopy layer, the ground surface layer and the root layer of the ground to generate the green manure orchard aphid number representing the vertical structure of the orchard.

[0109] The above embodiments can be realized by software, hardware, firmware or any combination thereof, in whole or in part. When realized by software, the above embodiments can be realized in the form of a computer program product in whole or in part.

[0110] Those skilled in the art can understand that the modules and algorithm steps of each example described in combination with the embodiments disclosed herein can be realized in electronic hardware or a combination of computer software and electronic hardware. Whether the functions are realized in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to realize the described functions for each specific application, but such implementation should not be considered beyond the scope of the present application.

[0111] In addition, each functional module in each embodiment of the present application can be integrated in one processing module, or each module can exist physically alone, or two or more modules can be integrated in one module.

[0112] The above is merely specific embodiments of the present application, but the protection scope of the present application is not limited thereto, any skilled in the art within the technical scope disclosed in the present application can easily think of changes or replacements, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be subject to the protection scope of the claims.

[0113] Finally, the above is only the preferred embodiments of the present application, and is not used to limit the present application, any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application should be included in the protection scope of the present application.

Claims

1. A method for detecting aphid populations in a green manure orchard ecosystem using multi-dimensional sampling, characterized in that: include: S1. The green manure orchard ecosystem is divided into the fruit tree canopy layer, the surface layer and the underground root system layer according to the vertical structure, and samples are collected at each level during the preset time period each day. S2. Collect leaf images of the fruit tree canopy layer, identify the spectral characteristics of aphid secretions in the images to identify the distribution area of ​​aphid spots, count the distribution density of aphid spots within a unit volume of the canopy layer, and calculate the total number of aphids in that layer. S3. Collect surface temperature distribution data in a grid, identify surface areas with temperature values ​​higher than the preset temperature difference threshold as aphid activity hotspots, and calculate the coverage rate of the hotspots per unit surface area to obtain the total number of aphids in that layer. S4. Drill soil samples in the dense root zone of the root system, and collect soil particle surface images after separating the samples into layers. Identify the number of aphid nymphs based on their morphological characteristics, and calculate the total number of aphids in the layer by combining the sampling volume with the total soil volume. S5. The number of aphids calculated by superimposing the canopy layer, surface layer and root system layer of fruit trees is used to generate the number of aphids in green manure orchards that characterize the vertical structure of the orchard.

2. The method for detecting aphid numbers in a green manure orchard ecosystem using multi-dimensional sampling as described in claim 1, characterized in that: The specific implementation of dividing the green manure orchard ecosystem according to its vertical structure is as follows: The vertical distance from the branching point of the trunk to the ground is measured with the main trunk of the fruit tree as the reference. The area from the branching point of the trunk to the top of the canopy is defined as the canopy layer of the fruit tree. Using the ground as a horizontal reference plane, the space extending upwards from the ground to below the branching point of the tree trunk is defined as the surface layer. Using the ground as a horizontal reference plane, extending vertically downwards to the lower limit of the main root distribution depth of fruit trees, the three-dimensional soil space formed below this reference plane is defined as the underground root layer.

3. The method for detecting aphid numbers in a green manure orchard ecosystem using multi-dimensional sampling as described in claim 2, characterized in that: The specific process for acquiring leaf images from the canopy layer of fruit trees is as follows: The green manure orchard is divided into several units. Each unit is centered on a reference sample fruit tree and includes a certain number of surrounding fruit trees. Taking the orchard operation entrance as the starting point of the path, obtain the plane rectangular position coordinates of the center points of all units, and calculate the straight-line distance between each center point and the starting point; Each unit is assigned a unique digital identifier, and the identifiers are rearranged in order of straight-line distance from nearest to farthest. The flight path is generated by connecting the center points in the new order. When the drone carrying the multispectral imager arrives at the center point of the target unit along the flight path, it adjusts the hovering height and performs multi-angle rotational scanning to collect images of the fruit tree canopy layers within the unit.

4. The method for detecting aphid numbers in a green manure orchard ecosystem using multi-dimensional sampling as described in claim 1, characterized in that: The specific steps for identifying the distribution area of ​​aphid spots in the image by recognizing the spectral features of aphid secretions are as follows: Clear leaf images without obstruction or reflection were selected from the collected leaf images and then classified and organized according to unit and fruit tree number. The full surface spectral information of each leaf image was extracted and compared with the preset spectral benchmarks of healthy leaves and aphid secretions. Mark the pixel regions whose spectral reflectance curves deviate from the spectral reference of healthy leaves and whose spectral reflectance curves fall within the preset matching range of the spectral reference of aphid secretions. Perform boundary continuity analysis on the marked pixel region, merge adjacent pixels to form independent contour regions, and filter out contour breakage or fragmented regions. If the location of the regional spectral reflectance peak coincides with the typical spectral reflectance peak of leaf diseases, it is determined to be a disease interference area. If the location of the regional spectral absorption valley coincides with the spectral absorption valley of the stain, it is determined to be an environmental interference zone. Independent outline regions that were not excluded were retained and marked as the final insect spot distribution areas.

5. The method for detecting aphid numbers in a green manure orchard ecosystem using multi-dimensional sampling as described in claim 1, characterized in that: The distribution density of insect spots within the canopy layer of the statistical unit, and the calculation of the total number of aphids in that layer, are detailed below: The total number of pixels in all leaf images within the summation unit is marked as the total number of pixels in the insect spot distribution area; Based on the spatial resolution of the multispectral imager, the total number of pixels is converted into the total actual insect spot coverage area; Obtain the vertical height data of the canopy layer of the reference sample fruit tree marker, and match the vertical height of the canopy layer of the corresponding reference sample according to the current unit center point position; The three-dimensional volume of the canopy layer is calculated by taking the geographical boundary of the unit as the bottom projection range and combining it with the vertical height of the matching canopy layer. The ratio of the total area of ​​insect spots in a unit to the volume of the canopy layer in a unit is calculated, and the distribution density value of insect spots per unit volume of canopy layer is output. The mapping relationship parameter between the unit area of ​​insect spots and the number of aphids is retrieved, and the generated insect spot distribution density value is converted into the base number of aphids per unit volume of canopy layer. Multiply this base number by the total volume of the canopy layer of the corresponding unit to obtain the estimated number of aphids in the corresponding unit; The total estimated number of aphids after traversing all units yields the total number of aphids in the canopy layer of the fruit tree.

6. The method for detecting aphid numbers in a green manure orchard ecosystem using multi-dimensional sampling as described in claim 1, characterized in that: The specific details of the gridded surface temperature distribution data are as follows: The surface plane is divided into several independent grid units according to a preset area threshold, and a fixed infrared thermal imager is set in each grid unit. Using the geometric center of the orchard boundary outline as the reference point, a Cartesian coordinate system covering all grid cells is established, and the coordinate data of the center point of each grid cell is recorded. Based on the differences in the density of green manure vegetation distribution in the orchard and the shading range of fruit tree projections, the grid units are merged and adjusted to form merged collection areas of varying sizes. Assign a device identification code to each merged acquisition area, and start the infrared thermal imagers in each area in sequence according to the identification code order; All infrared thermal imagers are simultaneously triggered to collect surface temperature data within a preset time period each day, generating temperature distribution maps for each merged collection area. The temperature distribution maps are stitched together according to coordinate data to generate a temperature distribution dataset covering the entire surface layer of the orchard.

7. The method for detecting aphid numbers in a green manure orchard ecosystem using multi-dimensional sampling as described in claim 1, characterized in that: The calculation of the coverage rate of the hotspot area per unit surface area to obtain the total number of aphids in that layer is as follows: The total area of ​​the hotspot area is obtained by summing up the actual surface areas of each hotspot area, and the total actual monitored area of ​​the corresponding surface layer is also obtained. The coverage rate of hotspot areas is obtained by calculating the ratio of the total area of ​​hotspot areas to the actual total monitored area. Retrieve pre-established parameters relating aphid numbers to hotspot areas per unit area; Multiply the coverage rate of the hotspot area by the aphid population per unit area of ​​the hotspot area; Multiply the calculation result by the actual total monitored area to obtain the total number of aphids on the ground surface.

8. The method for detecting aphid numbers in a green manure orchard ecosystem using multi-dimensional sampling as described in claim 1, characterized in that: The specific steps for drilling soil samples in the dense root zone of the underground root system are as follows: A core root distribution area is delineated around the main trunk of the fruit tree. Using the boundary of the core root distribution area as a reference, a series of sampling points are set up at preset intervals along its outer contour line. Soil drilling equipment was used to drill vertically downwards at each sampling point to obtain soil column samples, and the drilling depth and geographical coordinates of each sampling point were recorded. The obtained continuous soil column samples were laterally cut at preset depth intervals to separate them into several independent layered soil samples. Multi-angle images of the surface of dispersed soil particles were acquired using a microscopic imaging device.

9. The method for detecting aphid numbers in a green manure orchard ecosystem using multi-dimensional sampling as described in claim 1, characterized in that: The specific details of identifying the quantity of aphid nymphs based on their morphological characteristics are as follows: The particle surface images of each layer of soil sample are retrieved sequentially. Based on the preset primary screening threshold for biological morphology, the outline objects identified in the images are judged, and objects that meet the threshold are judged as suspected targets. The morphological features of the suspected target are compared with the pre-stored set of typical morphological parameters of aphid nymphs. If the suspected target features match the parameter set definition combination, it is determined to be a nymph individual and is marked and counted. After traversing all images, the total number of nymphs in the sample is obtained by accumulating the marker counts. The standard cross-sectional area of ​​the soil column sample is obtained based on the drilling equipment parameters. The volume components of each independent layer sample are calculated based on the layer cutting depth. The total sampling volume is output by summing the volume components. The nymph density per unit volume of soil is obtained by dividing the total number of nymphs in the sample by the total sampling volume.

10. The method for detecting aphid numbers in a green manure orchard ecosystem using multi-dimensional sampling as described in claim 1, characterized in that: The specific details of calculating the total aphid population in this layer by combining the sampling volume and the total soil volume are as follows: The three-dimensional spatial boundary data of the root system layer is obtained, the root system layer depth value of each sample is extracted, and the arithmetic mean is calculated as the depth. At the same time, the horizontal projected area of ​​the surface layer is obtained, and the depth is multiplied by the horizontal projected area to obtain the total soil volume. The total number of aphids in the root zone is estimated by multiplying the nymph density base number with the total soil volume.

Images