A Tea Tree Disease Identification System Based on UAV Remote Sensing Data

Abstract

The present invention discloses a tea tree disease identification system based on unmanned aerial vehicle (UAV) remote sensing data. The pest and disease conditions of tea trees directly affect the yield and quality of tea leaves. Traditional tea tree disease monitoring has problems such as time-consuming, laborious, low efficiency, and high cost. Therefore, the present invention combines UAV remote sensing technology with tea tree disease control, uses UAVs to collect hyperspectral image data of the tea tree canopy, corrects the collected image data, and inversely calculates the chlorophyll content and leaf area index of the tea tree canopy based on the reflectance data measured by hyperspectral remote sensing. By comparing and analyzing the spectral characteristics of healthy and diseased tea leaves, the pest and disease conditions of the tea garden can be effectively predicted in advance. This can meet the high-frequency regular monitoring requirements of tea tree pest and disease information and improve the census efficiency of tea tree pests and diseases.

Description

Technical Field

[0001] The present invention relates to the field of tea tree maintenance, and particularly to a tea tree disease identification system based on unmanned aerial vehicle (UAV) remote sensing data. Background Art

[0002] During the growth and development process of tea trees, they are constantly faced with the hazards of pests and diseases. Pests and diseases can lead to poor tea quality and reduced yields, causing serious economic losses. At present, the common methods for preventing and controlling pests and diseases in tea gardens in China include spraying pesticides and trapping with insect nets, etc. These methods are inefficient, with limited control effects, and the chemical pesticides used can cause certain harm to the human body and the tea garden environment. Timely and effectively monitoring tea trees and providing early warnings are conducive to controlling the development of tea tree pests and diseases, minimizing economic losses to the greatest extent, and improving economic benefits. Currently, the main method is to use traditional manual census to detect pests and diseases in tea trees, which has a large labor intensity, consumes financial resources, and has a low census efficiency. If tea farmers use UAV remote sensing technology, they can accurately detect the tea garden environment, understand the growth status of tea trees, and combine advanced technologies such as big data and artificial intelligence to analyze the degree and cause of tea tree disasters, so as to formulate scientific and effective control plans, save the amount of pesticide spraying, and improve the comprehensive efficiency of the tea garden.

[0003] Hyperspectral remote sensing technology is one of the means for monitoring the spectral characteristics changes of crop pests and diseases internationally at present. It has the characteristics of high spectral resolution, integration of spectrum and image, many and continuous bands, and large amount of spectral information. Using UAVs as carriers and carrying sensors such as hyperspectral imagers to obtain surface remote sensing images, and processing the collected image information by computers. This combines the advantages of the UAV being portable, easy to operate, and the hyperspectral imager being able to collect continuous spectral information of ground objects, and can conveniently and quickly monitor large areas of tea mountains and prevent and control pests and diseases. The application of hyperspectral data makes information extraction more beneficial. Studying the spectral changes of crops after being damaged by pests and diseases, determining the sensitive bands and sensitive periods for monitoring different crops and pests and diseases, and how to fuse spectral data with computer vision for machine learning are the research hotspots and keys for current hyperspectral remote sensing used in crop pest and disease monitoring. In view of the above problems, a solution is proposed below. Summary of the Invention

[0004] The purpose of the present invention is to provide a tea tree disease identification system based on UAV remote sensing data, which has the advantages of being able to obtain more comprehensive overall and local pest control and spraying plans, and solving the technical problems of current waste of human resources and inability to carry out precise pesticide application operations.

[0005] The above technical purpose of the present invention is achieved through the following technical solutions:

[0006] A tea tree disease identification system based on UAV remote sensing data, including a data monitoring and receiving system, a UAV autonomous navigation system, and a data transmission and cloud computing platform,

[0007] The data monitoring and receiving system includes:

[0008] A monitoring and positioning module, which is used to locate the UAV collective to obtain the flight coordinates and slope elevation data;

[0009] A sensor remote sensing module, which is used to control and receive the monitoring data collected by the UAV. The monitoring data includes visible spectrum data, which is the spectral image of tea leaf surface diseases and generates high-precision spectral image data. The high-precision spectral image data includes spectral image data of multiple bands. The sensor module is used to provide visible spectrum data, and the sensor module is used to provide flight coordinates and slope elevation data;

[0010] An artificial algorithm module, which is used to correct the processing accuracy and contrast of the image, and the artificial algorithm is used for edge calculation of correcting the high-precision spectral image data;

[0011] A data receiving and storing module based on STM32F103, which is used to transmit and save the monitoring data from the peripheral device to the local memory through the synchronous DMA short pulse transmission protocol to generate local memory data;

[0012] The data transmission and cloud computing platform includes:

[0013] A data transmission unit DTU, which is a wireless terminal device that converts the local memory data into IP data and transmits it through a 4G mobile information system;

[0014] A tea tree disease degree index calculation module, which is preset with disease degree levels, obtains the tea tree disease degree value according to the transmitted high-precision visible spectrum data and spectral image data of multiple bands, and obtains the first pre-plan data for providing prevention and control suggestions;

[0015] A geographical segment integration module, which is used to integrate the tea garden pest and disease distribution vector map from the flight coordinates and slope elevation, and obtain the second pre-plan data;

[0016] A pest and disease calculation module, which is used to input the first pre-plan data and the second pre-plan data into the neural network spraying model to obtain pest and disease detection and spraying data;

[0017] Pest and disease inspection and control model, which is used to form a pest and disease inspection and control plan according to pest and disease inspection and control data. The pest and disease inspection and control plan includes an overall inspection and control plan for the tea garden and a key local inspection and control plan;

[0018] The UAV autonomous navigation system includes a flight control system, a UAV positioning system, and a vision processing system.

[0019] The sensor module acquires surface remote sensing image data and performs calibration processing on the image data. The calibration processing includes radiometric calibration of the collected hyperspectral image data, then denoising processing, and finally geometric calibration.

[0020] Preferably, the artificial algorithm adjusts the output of the filter according to the local variance of the hyperspectral image data, establishes models for each band of the image data, and uses the established models to achieve radiometric calibration of the hyperspectral image data; the Wiener filtering algorithm is used to denoise the hyperspectral image data after radiometric consistency calibration; a dual-channel is used to repair the missing values of the hyperspectral image data.

[0021] Preferably, the UAV positioning system includes an orientation target with known reflectivity changes. The orientation target is installed in the tea garden, and the sensor module collects hyperspectral image data of the calibration target.

[0022] Preferably, the neural network spraying model is used to input the levels of the amount of pesticides to be sprayed one by one for calculation, and its output vector is used to control the spraying amount of pesticides.

[0023] The pest and disease inspection and control model is used to form a pest and disease inspection and control plan according to pest and disease inspection and control data;

[0024] The pest and disease inspection and control plan includes an overall inspection and control plan for the tea garden (tea slope) and a key local inspection and control plan;

[0025] The monitoring data monitored by the sensor remote sensing module and the flight coordinates and slope elevation data provided by the positioning system are synchronously acquired by the STM32 single-chip microcomputer's dual ADC and saved internally through the DMA protocol. Since among all visible spectral bands, the visible spectral data that can reflect the growth information of tea crops is only at special band positions among all bands, the sensor module will first screen out the visible spectral data and transmit this data and the positioning data to the DTU through the modbus protocol to make the usage process more automated.

[0026] The above is the edge processing process of the system data monitoring and reception. The following details the specific data transmission and cloud processing process:

[0027] A wireless terminal device that converts data into IP data through a DTU and transmits it through a 4G mobile information system. The cloud corrects and denoises the collected content, optimizes and splices some of the corrected data, saves the available results, and constructs a complete vector map of the pest and disease area. The cloud computing platform filters out the corresponding pest and disease control plans from the library through the vector map.

[0028] The automatic cruise of the unmanned aerial vehicle (UAV) automatic navigation system is also stored in the cloud and the UAV body after being set. For the automatic cruise, the flight route and cruise time are variable files that can be modified at any time. Multiple different sub-files and alternative files can also be provided. The cruise operation process is recorded at least twice on the same target to obtain higher-accuracy hyperspectral data.

[0029] The tea tree disease identification system based on UAV remote sensing data provided by the present utility model includes a UAV and a cloud platform. The components carried by the UAV include:

[0030] An STM32 embedded main control module based on an ARM Cortex-M3 processor;

[0031] A visible light spectrum sensor that collects hyperspectral image data of the tea garden;

[0032] A storage module that saves the collected hyperspectral images and positioning information;

[0033] A 4G communication module that transmits the collected hyperspectral images and positioning information to the cloud platform and performs positioning through WIFI

[0034] The beneficial effects of the present invention are:

[0035] (1) Through the WIFI positioning method, the three-dimensional reconstruction of the tea garden designed for contour terraces can be carried out, which can be applied to the monitoring requirements of pests and diseases on slopes with different heights and frequencies, and improve the efficiency of fine census of the tea garden.

[0036] (2) The regular cruise monitoring of the tea garden effectively liberates productivity.

[0037] (3) Compared with extracting single tea tree-related growth information from each part, the vector map based on the structure of the tea garden slope can more comprehensively and purposefully implement pest and disease control.

[0038] (4) The cloud-edge collaboration method is adopted, which speeds up the processing ability of the overall system and improves the utilization degree. Description of the Drawings

[0039] Figure 1 It is a principle block diagram of the tea tree disease identification system based on UAV remote sensing data for the embodiment;

[0040] Figure 2This is the data transmission schematic diagram of the tea tree disease identification system based on UAV remote sensing data for the embodiment. Specific implementation manners

[0041] The following are only the preferred implementation manners of the present invention, and the protection scope is not limited to this embodiment. All technical solutions falling within the idea of the present invention shall belong to the protection scope of the present invention.

[0042] The present invention provides a tea tree disease identification system based on UAV remote sensing data, adopting a sub-system processing mode, including a UAV and a cloud platform. The UAV is equipped with: an STM32 embedded main control module based on an ARM Cortex-M3 processor; a visible light spectrum sensor for collecting hyperspectral image data of the tea garden; a storage module for saving the collected hyperspectral images and positioning information; a 4G communication module for transmitting the collected hyperspectral images and positioning information to the cloud platform and for positioning via WIFI

[0043] The above UAV system is used for the general survey of tea red spider mites in Darjeeling tea. The method includes:

[0044] (1) First, on the slope at the foot of the Himalayas, it is divided into four gradients and four known targets according to the ground laying; the positioning of the targets is obtained through signal transmission; the design of the four gradients is to guide the flight displacement of the UAV at different slope heights;

[0045] (2) Use the visible light spectrum sensor to obtain the corresponding visible hyperspectral image data of the tea tree; during the shooting process, the same location is shot at least twice, and the resolvability of the obtained data is judged by calculating the overlap rate;

[0046] (3) Correct and store the obtained hyperspectral image data through the DMA protocol; changes in weather elements (such as rain) will cause different colored light values in different shootings; all will be corrected using artificial algorithms. For the same location, the overlap values of multiple obtained pictures are taken, and some bands with different wave points are removed for consistency correction; the corrected data is stored with the positioning data through the DMA protocol;

[0047] (4) The stored data is simultaneously transmitted to the DTU through the modbus protocol, and the DTU transmits the collected data to the cloud platform for analysis and implementation through 4G mobile communication;

[0048] (5) The cloud platform obtains the pest control and killing plan through further correction and integration. At the same time, the STM32 main controller also controls the flight of the UAV, so as to realize a more accurate and efficient general survey of tea garden pest information.

[0049] The specific embodiments described above further elaborate on the technical problems to be solved, the technical solutions, and the beneficial effects of the present invention. It should be understood that the above are only specific embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principles of the present invention shall be included within the protection scope of the present invention.

Claims

1. A tea tree disease identification system based on unmanned aerial vehicle remote sensing data, characterized in that, It includes a data monitoring and receiving system, an unmanned aerial vehicle (UAV) autonomous navigation system, and a data transmission and cloud computing platform. The data monitoring and receiving system includes: A monitoring and positioning module, which is used to locate the UAV collective to obtain the flight coordinates and slope elevation data. A sensor remote sensing module, which is used to control and receive the monitoring data collected by the UAV. The monitoring data includes visible spectral data, which is the spectral image of tea leaf surface diseases and generates high-precision spectral image data. The high-precision spectral image data includes spectral image data of multiple bands. The remote sensing sensor module is used to provide visible spectral data, and the sensor module is used to provide flight coordinates and slope elevation data. An artificial algorithm module, which is used to correct the processing accuracy and contrast of the image. The artificial algorithm is used for edge calculation of the correction processing of high-precision spectral image data. A data receiving and storing module based on STM32F103, which is used to transmit and save the monitoring data from the peripheral device to the local memory through the synchronous DMA short pulse transmission protocol to generate local memory data. The data transmission and cloud computing platform includes: A data transmission unit DTU, which is a wireless terminal device that converts the local memory data into IP data and transmits it through a 4G mobile information system. A tea disease degree index calculation module, which is preset with disease degree levels. It obtains the tea disease degree value based on the transmitted high-precision visible spectral data and spectral image data of multiple bands, and obtains the first pre-plan data for providing prevention and control suggestions. A geographical segment integration module, which is used to integrate the UAV's flight coordinates and slope elevation data into a vector map of the distribution of tea garden pests and diseases, and obtain the second pre-plan data. A pest control and disease calculation module, which is used to input the first pre-plan data and the second pre-plan data into a neural network spraying model to obtain pest and disease detection and control data. A pest and disease detection and control model, which is used to form a pest and disease detection and control plan based on the pest and disease detection and control data. The pest and disease detection and control plan includes an overall detection and control plan for the tea garden and a key local detection and control plan. The UAV autonomous navigation system includes a flight control system, a UAV positioning system, and a vision processing system.

2. The tea tree disease recognition system based on unmanned aerial vehicle remote sensing data according to claim 1, characterized in that, The sensor module obtains surface remote sensing image data and corrects the image data. The correction processing includes radiometric correction of the collected hyperspectral image data, then denoising processing, and finally geometric correction.

3. The tea tree disease identification system based on UAV remote sensing data according to claim 2, characterized in that, The artificial algorithm adjusts the output of the filter based on the local variance of the hyperspectral image data, establishes models for each band of the image data, and uses the established models to achieve radiometric correction of the hyperspectral image data; uses the Wiener filtering algorithm to denoise the hyperspectral image data after radiometric consistency correction; and uses a dual-channel method to repair the missing values of the hyperspectral image data.

4. The tea tree disease identification system based on UAV remote sensing data according to claim 1, characterized in that, The UAV positioning system includes a directional target with a known reflectivity change. The directional target is installed in the tea garden, and the sensor module collects hyperspectral image data of the calibration target.

5. The tea tree disease identification system based on UAV remote sensing data according to claim 1, characterized in that, The neural network spraying model is used to input the levels of the amount of pesticides to be sprayed one by one for calculation, and its output vector is used to control the spraying amount of pesticides.

Images