CN-121999396-A - Degraded grassland unmanned aerial vehicle stellera chamaejasme plant positioning and population characteristic analysis method and system

CN121999396ACN 121999396 ACN121999396 ACN 121999396ACN-121999396-A

Abstract

The invention particularly relates to a method and a system for positioning and analyzing population characteristics of a degraded grassland unmanned aerial vehicle stellera chamaejasme plant, which solve the problem that a target algorithm is weak in the outline and characteristic detection capability of a small plant. The method comprises the steps of acquiring degraded grassland panchromatic image data through image acquisition equipment carried by an unmanned aerial vehicle, detecting and identifying radix euphorbiae Fischerianae plants in an image by utilizing a SAM v2 model, outputting radix euphorbiae Fischerianae plant data frames and confidence in the image, extracting center point coordinate information of each divided image based on a Python library, positioning radix euphorbiae Fischerianae plants in the image according to azimuth angles and distances, and distinguishing the dependence degree of a radix euphorbiae Fischerianae population space distribution pattern and the distribution pattern of the radix euphorbiae Fischerianae population under different scales based on K (r) and L (r) functions. Realizes the monitoring and real-time evaluation of the stellera chamaejasme plants in the landscape scale, and provides a new technical system and a research paradigm for evaluating the degraded grasslands.

Inventors

WANG JINNIU
LIU CHUNHONG
ZHANG NING
LI JIA
ZHENG ZHE

Assignees

中国科学院成都生物研究所

Dates

Publication Date: 20260508
Application Date: 20260206

Claims (9)

1. The method for positioning and analyzing the population characteristics of the degraded grassland unmanned aerial vehicle stellera chamaejasme plants is characterized by comprising the following steps: S1, acquiring full-color image data of degraded grassland through image acquisition equipment carried by an unmanned aerial vehicle; S2, preprocessing the acquired image data, detecting and identifying the radix euphorbiae Fischerianae plants in the image by using a SAM v2 model, and outputting a radix euphorbiae Fischerianae plant position data frame and confidence in the image; S3, extracting the coordinate information of the center point of each divided image based on a Python library according to a positioning system carried by the unmanned aerial vehicle, and positioning the stellera chamaejasme plants in the images according to azimuth angles and distances; S4, distinguishing the dependence degree of the spatial distribution pattern of the stellera chamaejasme seed group and the distribution pattern of the stellera chamaejasme seed group under different scales based on a K (r) function and an L (r) function according to the individuality and the position information of the stellera chamaejasme plant.
2. The method according to claim 1, wherein the step S1 comprises: S101, flying an unmanned aerial vehicle carrying an RGB full-color camera according to a preset unmanned aerial vehicle flight path, and shooting a degraded grassland to obtain low-altitude unmanned aerial vehicle initial full-color image data; s102, performing image feature point matching, point cloud generation, grid construction, texture mapping and orthophoto creation on the initial panchromatic image data, and outputting a region unmanned aerial vehicle synthesized image.
3. The method according to claim 1, wherein the step S2 comprises: S201, carrying out data slicing processing on a synthesized image of the regional unmanned aerial vehicle according to the SAM v2 model input requirement, and mapping the synthesized image to a feature space by using a convolution algorithm; s202, obtaining features matched with an image embedding space through marking points, marking frames and mask prompts; s203, up-sampling the Mask result in S202 by fusing the image features from the encoder with the context information of the target area captured by the prompt information to enable the Mask result to be matched with the original image size in S201, transforming the up-sampled features through a series of full connection layers, calculating Mask probability for each pixel, and evaluating the overlapping degree between the generated Mask and the real Mask.
4. The analysis method according to claim 1, wherein in the step S3, the longitude and latitude (X 0 ,Y 0 ) of the center point of the image are extracted, the azimuth angle of the detected stellera chamaejasme plant from the center point is θ, the distance is d, the earth radius is R, and the longitude and latitude (X 1 ,Y 1 ) of the stellera chamaejasme plant are obtained by the following steps: S301, converting longitude and latitude of a central point and radian: X 0 `= X 0 ×π/180; Y 0 ` = Y 0 ×π/180; θ`= θ×π/180; X 0 `、Y 0 'is longitude and latitude of the center point of the image in radian, and θ' is azimuth angle of the center point in radian; s302, calculating longitude and latitude of a target point: X 1 `=X 0 `+arctan2(sinθ`×sin (d/R)×cosY 0 `, cos(d/R)-sinX 0 `sinY 0 `); Y 1 `=arcsin(sinY 0 `×cos (d/R)+cosY 0 `×sin(d/R)×cosθ`); S303, converting the radian system of the target point into longitude and latitude: X 1 =X 1 `×180/π; Y 1 =Y 1 `×180/π。
5. The method according to claim 1, wherein in the step S4, the degree of dependence of the spatial layout of the stellera chamaejasme plant is calculated as follows: K(r)=λ -1 E (#(r ij ≤ r)) ; Wherein lambda represents the density of radix Euphorbiae Fischerianae plants per unit area in the degraded grassland, E (&) represents the expected number of radix Euphorbiae Fischerianae plants under a certain distance scale, # & gt represents the number, r ij represents the distance between any radix Euphorbiae Fischerianae plant i and other plants j in the degraded grassland, r represents the distance scale; the distribution type of the stellera chamaejasme plants under the scale is judged by adopting the estimated value of the L (r) function: L(r)= 。
6. the degraded grassland unmanned aerial vehicle stellera chamaejasme plant positioning and population characteristic analysis system is characterized by comprising the following components: The unmanned aerial vehicle low-altitude image acquisition module acquires degraded grassland orthopanchromatic image data through image acquisition equipment carried by the unmanned aerial vehicle, wherein the image data comprises R, G, B wave bands and spatial position information; The radix euphorbiae Fischerianae plant identification module is used for preprocessing the acquired image data, detecting and identifying radix euphorbiae Fischerianae plants in the image by utilizing a SAM v2 model, identifying the degenerated grass radix euphorbiae Fischerianae plants by image coding, prompt coding and mask decoding, and outputting the shape characteristics and the confidence coefficient of the degenerated grass radix euphorbiae Fischerianae plants; the stellera chamaejasme plant position solving module is used for extracting the coordinate information of the central point of each divided image based on exifread libraries in Python according to a positioning system carried by the unmanned aerial vehicle, and calculating the positioning of the stellera chamaejasme plant in the image according to the azimuth angle and the distance; the radix euphorbiae Fischerianae seed group characteristic analysis module is used for distinguishing the dependence degree of the radix euphorbiae Fischerianae seed group spatial distribution pattern and the distribution mode of the radix euphorbiae Fischerianae seed group under different scales based on K (r) and L (r) functions according to the individuality and the position information of radix euphorbiae Fischerianae plants.
7. A computer apparatus/device/system comprising a memory, a processor and a computer program stored on the memory, the processor executing the computer program to perform the steps of the method according to any of claims 1-5.
8. A computer-readable storage medium, characterized in that it has stored thereon a computer program/instruction which, when executed by a processor, implements the steps of the method according to any of claims 1-5.
9. A computer program product comprising computer programs/instructions which, when executed by a processor, implement the steps of the method according to any of claims 1-5.

Description

Degraded grassland unmanned aerial vehicle stellera chamaejasme plant positioning and population characteristic analysis method and system Technical Field The invention belongs to the technical field of low-altitude target identification, and particularly relates to a method and a system for positioning and analyzing population characteristics of a degraded grassland unmanned aerial vehicle stellera chamaejasme plant. Background Along with the rapid development of unmanned aerial vehicle technology, unmanned aerial vehicles are increasingly widely applied in the fields of agricultural estimated production, pest control and the like. Particularly in the low-altitude field, the unmanned aerial vehicle becomes an important tool for target positioning and identification due to flexibility and high efficiency. The degraded grassland unmanned aerial vehicle stellera chamaejasme plant identification and positioning and population analysis system can realize rapid identification and positioning of the ground stellera chamaejasme plant, analyze the stellera chamaejasme seed group characteristics, and provide accurate and timely information support for species distribution investigation and analysis, degraded grassland monitoring and evaluation and the like. In the prior art, unmanned aerial vehicles need to identify and position grassland vegetation when vegetation is investigated. The plants are particularly difficult to investigate due to small size, various characters, and images of complicated ground surfaces, illumination changes, shrub shielding and other factors, and the detection capability of the plants is relatively weak due to the limitation of a target detection algorithm. Under a complex surface environment, the system may not accurately capture the outline and characteristics of the miniature plants, so that vegetation investigation is incomplete and other conditions frequently occur. Therefore, a system for identifying and analyzing the population characteristics of the stellera chamaejasme plants in the degraded grasslands is provided. Disclosure of Invention In order to solve the problems in the background technology, the invention aims to provide a method and a system for positioning and analyzing population characteristics of a degraded grassland unmanned aerial vehicle stellera chamaejasme plant. In order to achieve the aim of the invention, the technical scheme adopted by the invention is that the method for positioning and analyzing the population characteristics of the degraded grassland unmanned aerial vehicle stellera chamaejasme plants comprises the following steps: S1, acquiring full-color image data of degraded grassland by using image acquisition equipment carried by an unmanned aerial vehicle, wherein the image data comprises R, G, B wave bands and spatial position information; S2, preprocessing the acquired image data, detecting and identifying the radix euphorbiae Fischerianae plants in the image by using a SAM v2 model, and outputting a radix euphorbiae Fischerianae plant position data frame and confidence in the image; S3, extracting the coordinate information of the center point of each divided image based on a Python library according to a positioning system carried by the unmanned aerial vehicle, and positioning the stellera chamaejasme plants in the images according to azimuth angles and distances; S4, distinguishing the dependence degree of the spatial distribution pattern of the stellera chamaejasme seed group and the distribution pattern of the stellera chamaejasme seed group under different scales based on a K (r) function and an L (r) function according to the individuality and the position information of the stellera chamaejasme plant. In a further embodiment, the step S1 includes: S101, flying an unmanned aerial vehicle carrying an RGB full-color camera according to a preset unmanned aerial vehicle flight path, and shooting a degraded grassland to obtain low-altitude unmanned aerial vehicle initial full-color image data; s102, performing image feature point matching, point cloud generation, grid construction, texture mapping and orthophoto creation on the initial panchromatic image data, and outputting a region unmanned aerial vehicle synthesized image. In a further embodiment, the step S2 includes: S201, carrying out data slicing processing on a synthesized image of the regional unmanned aerial vehicle according to the SAM v2 model input requirement, and mapping the synthesized image to a feature space by using a convolution algorithm; s202, obtaining features matched with an image embedding space through marking points, marking frames and mask prompts; The SAM v2 model extracts features and classifies targets of the images, judges whether the stellera chamaejasme plants exist in the images, and outputs individual shape features such as the size, the shape and the like of the stellera chamaejasme plants; s203, up-sampling the Mask result in S202 by fusing the image featu