Dragon fruit growth monitoring method, device, equipment and storage medium

Abstract

The present application relates to the field of fruit monitoring, and discloses a dragon fruit growth monitoring method, device, equipment and storage medium, which is used for monitoring the growth state of dragon fruit; the method comprises the following steps: acquiring position information and basic information of a dragon fruit tree, establishing an identity card of the dragon fruit tree and a dragon fruit file according to the position information and the basic information of the dragon fruit tree; planning a patrol route, and controlling a patrol robot to collect real-time images of the dragon fruit tree according to the patrol route; identifying the position information and the fruit hanging amount of the dragon fruit tree according to the collected real-time images, and detecting the fruit maturity of the dragon fruit tree; confirming the identity card of the dragon fruit tree according to the identified position information of the dragon fruit tree, and updating the dragon fruit file corresponding to the identity card of the dragon fruit tree according to the collected real-time images, the identified fruit hanging amount of the dragon fruit tree and the detected fruit maturity of the dragon fruit tree.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of fruit monitoring, and in particular to a dragon fruit growth monitoring method, device, equipment and storage medium. BACKGROUND

[0002] In recent years, the planting area of dragon fruit in China has rapidly increased. With the expansion of the planting scale, achieving automatic and intelligent management of orchards is an important link for production scaling. In the intelligent management of orchards, maturity recognition can help orchard managers more accurately grasp the growth and maturity of fruits, optimize fruit picking time, and thus improve the production efficiency of orchards. In orchard management, the stem strip with more fruit setting needs to be thinned, so as to improve the sweetness of dragon fruit, the unripe fruits need to be bagged to avoid damage during growth, which is beneficial to the uniform coloring of the fruit skin, and the dragon fruit needs to be picked in time after ripening. If the best picking time is missed, the nutrients inside the fruit may be lost, affecting the taste of the fruit. Since dragon fruit can have up to 14 batches of fruit per year, the current monitoring of the growth state of dragon fruit is mainly manual, and lacks systematic management.

[0003] Therefore, the prior art still needs to be improved and developed. SUMMARY

[0004] The present application provides a dragon fruit growth monitoring method, device, equipment and storage medium for monitoring the growth state of dragon fruit.

[0005] The first aspect of the present application provides a dragon fruit growth monitoring method, the dragon fruit growth monitoring method comprising: acquiring position information and basic information of a dragon fruit tree, establishing an identity card of the dragon fruit tree and a dragon fruit file according to the position information and the basic information of the dragon fruit tree; planning a patrol route, and controlling a patrol robot to collect real-time images of the dragon fruit tree according to the patrol route; identifying the position information and the fruit hanging amount of the dragon fruit tree according to the collected real-time images, and detecting the fruit maturity of the dragon fruit tree; confirming the identity card of the dragon fruit tree according to the identified position information of the dragon fruit tree, and updating the dragon fruit file corresponding to the identity card of the dragon fruit tree according to the collected real-time images, the identified fruit hanging amount of the dragon fruit tree, and the detected fruit maturity of the dragon fruit tree.

[0006] Preferably, the position information and basic information of the dragon fruit trees are acquired, the identity card of the dragon fruit trees and the dragon fruit archives are established according to the position information and basic information of the dragon fruit trees, and the method comprises the following steps: using a drone or a ground measuring device to measure and record the position information of the dragon fruit trees, wherein the position information comprises longitude and latitude; acquiring the basic information of the dragon fruit trees, wherein the basic information comprises tree age, tree crown size and health condition, and integrating the position information and the basic information of the dragon fruit trees to establish the identity card of the dragon fruit trees; and establishing the dragon fruit archives corresponding to the identity card of the dragon fruit trees in a preset dragon fruit growth database.

[0007] Preferably, the patrol route is planned, and the patrol robot is controlled to collect real-time images of the dragon fruit trees according to the patrol route, and the method comprises the following steps: using a laser radar to collect point cloud data in the area between the two ends of each row of dragon fruit trees, and using an RTK-GNSS device to collect spatial position data in the area of the two ends of each row of dragon fruit trees; constructing a three-dimensional point cloud map as a prior map according to the point cloud data; estimating the initial pose of the laser radar, and optimizing the initial pose of the laser radar based on the prior map to obtain the optimized pose of the laser radar; in a loose coupling manner, fusing the spatial position data collected by the RTK-GNSS device and the optimized pose of the laser radar to obtain a fusion pose, and mapping the feature points at the corresponding time into the prior map according to the fusion pose to construct an environment map; obtaining a patrol starting point and a patrol ending point, planning a patrol route in the environment map by using a path planning algorithm, and controlling the patrol robot to collect real-time images of the dragon fruit trees according to the patrol route.

[0008] Preferably, the three-dimensional point cloud map is constructed as a prior map according to the point cloud data, and the method comprises the following steps: according to the scanning mode of the laser radar, the point cloud data is distinguished according to the vertical direction and the horizontal direction, and the edge points and the plane points are extracted from the distinguished point cloud data as global feature points; according to the global feature points, the point cloud data collected at different times in the same frame is compensated for motion deformation to obtain single-frame point cloud data without deformation; the single-frame point cloud data without deformation is registered to a global coordinate system and voxel decimation is performed to obtain a three-dimensional point cloud map, and the three-dimensional point cloud map is taken as a prior map.

[0009] Preferably, the initial pose of the laser radar is estimated, and the initial pose of the laser radar is optimized based on the prior map to obtain an optimized pose of the laser radar, comprising: loading the prior map, dividing the grid array in the prior map, and storing the global feature points into the corresponding grid; obtaining the IMU data of the laser radar, and determining the motion initial value of the laser radar according to the IMU data; obtaining the point cloud data of the current frame of the laser radar, extracting the current frame feature points from the point cloud data of the current frame, and determining the initial pose of the laser radar in the prior map according to the motion initial value of the laser radar and the current frame feature points; selecting the global feature points in the grid around the initial pose to form a local subgraph; performing feature matching on the current frame feature points and the global feature points in the local subgraph, and optimizing the initial pose of the laser radar according to the matching result to obtain the optimized pose of the laser radar.

[0010] Preferably, the position information and fruit hanging amount of the pitaya tree are recognized according to the collected real-time image, and the maturity of the fruits of the pitaya tree is detected, comprising: pre-processing the collected real-time image to obtain a pre-processed image; marking the position of the pitaya tree in the pre-processed image using a target detection algorithm to obtain a marked image; converting the position information of the pitaya tree in the marked image into actual position information according to a preset camera parameter; counting the number of fruits of the pitaya tree in the marked image to obtain the fruit hanging amount of the pitaya tree; performing maturity detection on each fruit in the marked image, and outputting the maturity information of each fruit and the overall maturity distribution.

[0011] Preferably, the maturity of each fruit in the marked image is detected, and the maturity information of each fruit and the overall maturity distribution are outputted, comprising: extracting the fruit area of the marked image using an image segmentation-based algorithm to obtain a plurality of fruit images; extracting the color features of the plurality of fruit images respectively, and inputting the extracted color features into a pre-trained SVM classifier to obtain the classification results outputted by the SVM classifier, the classification results including unripe fruits, semi-ripe fruits and ripe fruits; classifying the plurality of fruit images according to the classification results outputted by the SVM classifier, and counting the number and proportion of fruit images in each category.

[0012] The second aspect of the present application provides a pitaya growth monitoring device, comprising: a establishing module, configured to acquire position information and basic information of pitaya trees, and establish an identity card and a pitaya file of the pitaya trees according to the position information and the basic information of the pitaya trees; a planning module, configured to plan a patrol route, and control a patrol robot to collect real-time images of the pitaya trees according to the patrol route; an identification and detection module, configured to identify position information and fruit hanging amount of the pitaya trees according to the collected real-time images, and detect fruit maturity of the pitaya trees; and an updating module, configured to confirm the identity card of the pitaya trees according to the identified position information of the pitaya trees, and update the pitaya file corresponding to the identity card of the pitaya trees according to the collected real-time images, the identified fruit hanging amount of the pitaya trees and the detected fruit maturity of the pitaya trees.

[0013] The third aspect of the present application provides a pitaya growth monitoring device, comprising: a memory and at least one processor, the memory stores computer readable instructions, and the memory and the at least one processor are interconnected through a circuit; the at least one processor invokes the computer readable instructions in the memory, so that the pitaya growth monitoring device performs each step of the pitaya growth monitoring method as described above.

[0014] The fourth aspect of the present application provides a computer readable storage medium, which stores computer readable instructions, when running on a computer, so that the computer performs each step of the pitaya growth monitoring method as described above.

[0015] In the technical solution provided by the present application, by collecting position information and basic information of pitaya trees, establishing an identity card and a file, collecting real-time images by a patrol robot, and identifying the position of the pitaya trees, the fruit hanging amount and the fruit maturity in the collected real-time images by image recognition technology, and finally updating the pitaya file according to the identification result and the detection result, the growth state of the pitaya trees is quickly and comprehensively monitored, the monitoring efficiency is greatly improved, the manual intervention is reduced, and the management cost is reduced. BRIEF DESCRIPTION OF DRAWINGS

[0016] Figure 1 The flowchart of the pitaya growth monitoring method provided for the embodiment of the present application;

[0017] Figure 2 The structural schematic diagram of the pitaya growth monitoring device provided for the embodiment of the present application;

[0018] Figure 3 The structural schematic diagram of the pitaya growth monitoring device provided for the embodiment of the present application; DETAILED DESCRIPTION

[0019] The terms "first", "second", "third", "fourth" and the like in the description and in the claims of the present application, and above-described drawings, if any, are used to distinguish between similar objects and not necessarily for describing a particular sequential or chronological order. It is to be understood that the use of the terms so construed herein can be interchanged, under appropriate circumstances, and that the embodiments described herein can be carried out in other sequences than the one illustrated or described herein. Furthermore, the terms "comprising" or "including" and any of their derivatives, are intended to cover non-exclusive inclusions, such that a process, method, system, product, or apparatus that comprises a list of steps or units not necessarily limited to those explicitly stated, and can include other steps or units not expressly listed or inherent to such process, method, product, or apparatus.

[0020] For the purpose of facilitating understanding, the specific flow of the embodiments of the present application is described below, please refer to Figure 1 The first embodiment of the pitaya growth monitoring method in the embodiments of the present application comprises:

[0021] S101, acquiring the position information and basic information of the pitaya tree, and establishing the identity card of the pitaya tree and the pitaya file according to the position information and basic information of the pitaya tree.

[0022] S102, planning a patrol route, and controlling the patrol robot to collect real-time images of the pitaya tree according to the patrol route.

[0023] S103, identifying the position information and fruit hanging amount of the pitaya tree according to the collected real-time images, and detecting the fruit maturity of the pitaya tree.

[0024] S104, confirming the identity card of the pitaya tree according to the identified position information of the pitaya tree, and updating the pitaya file corresponding to the identity card of the pitaya tree according to the collected real-time images, the identified fruit hanging amount of the pitaya tree, and the detected fruit maturity of the pitaya tree.

[0025] It can be understood that the execution subject of the present application can be a pitaya growth monitoring device, and can also be a terminal or a server, which is not limited here. The embodiments of the present application take the server as the execution subject as an example for description.

[0026] In this embodiment, in step S101, the position information and basic information of the pitaya tree are obtained, the identity card of the pitaya tree is established according to the position information of the pitaya tree, and the pitaya file of the pitaya tree is established according to the identity card. Specifically, the position information of the pitaya tree is measured and recorded using a drone or a ground measuring device, and the position information includes latitude and longitude; the basic information of the pitaya tree is obtained, the basic information includes tree age, crown size and health status, and the position information and basic information of the pitaya tree are integrated to establish the identity card of the pitaya tree; and the pitaya file corresponding to the identity card of the pitaya tree is established in the preset pitaya growth database.

[0027] In this embodiment, a high-precision GPS system or a ground measuring device (such as an RTK GPS receiver) is used on a drone to accurately position each pitaya tree in the pitaya orchard. The latitude and longitude coordinates of each tree are recorded to ensure the accuracy and reliability of the data, providing a basis for subsequent identity card establishment.

[0028] In this embodiment, the basic information of the pitaya tree can be collected in advance through on-site investigation and recording, including tree age (determined by annual ring observation or planting record), crown size (estimated diameter or area of the crown using measuring tools such as a tape measure or a laser range finder), and health status (observation of leaf color, pest and disease conditions, etc.). Then upload the basic information to the system, so that the basic information of the pitaya tree can be obtained in the system.

[0029] Integrate the position information and basic information of the pitaya tree to form an identity card containing key information such as tree name, position information (latitude and longitude), tree age, crown size, and health status.

[0030] In the preset pitaya growth database, a corresponding file is created for each pitaya tree. The information in the identity card is entered into the file to ensure the completeness and traceability of the data. The pitaya growth database supports data updating and query functions to update the growth status and yield information of the pitaya tree in real time as it grows.

[0031] In this embodiment, the pitaya file can also be updated and maintained regularly to ensure the timeliness and accuracy of the data.

[0032] In practical applications, the information in the database can be used for data analysis and mining to find problems and rules in the growth process of pitaya, and according to the analysis results, targeted management measures and optimization schemes can be developed, such as adjusting the amount of fertilizer, pruning the crown, preventing and treating pests and diseases, etc.

[0033] In this embodiment, the position information of pitaya trees is accurately obtained by a UAV or a ground measuring device, and an identity card is established in combination with basic information, and then a special file is created for each pitaya tree in the growth database, realizing intelligent and fine pitaya planting management, and facilitating tracking, analysis and optimization of the growth conditions of pitaya.

[0034] In this embodiment, in step S102, a patrol route is planned, and the patrol robot is controlled to collect real-time images of pitaya trees according to the patrol route, including: using a laser radar to collect point cloud data in the area between the two ends of each row of pitaya trees, and using an RTK-GNSS device to collect spatial position data in the area of the two ends of each row of pitaya trees; constructing a three-dimensional point cloud map as a prior map according to the point cloud data; estimating the initial pose of the laser radar, and optimizing the initial pose of the laser radar based on the prior map to obtain the optimized pose of the laser radar; in a loosely coupled manner, fusing the spatial position data collected by the RTK-GNSS device and the optimized pose of the laser radar to obtain a fused pose, and mapping the feature points corresponding to the moment into the prior map according to the fused pose to construct an environment map; obtaining a patrol starting point and a patrol ending point, planning a patrol route in the environment map by using a path planning algorithm, and controlling the patrol robot to collect real-time images of pitaya trees according to the patrol route.

[0035] In this embodiment, the three-dimensional point cloud map is constructed as a prior map according to the point cloud data, including: according to the scanning mode of the laser radar, distinguishing the point cloud data according to the vertical direction and the horizontal direction, and extracting edge points and plane points from the distinguished point cloud data as global feature points; according to the global feature points, performing motion deformation compensation on the point cloud data collected at different moments in the same frame to obtain non-deformed single-frame point cloud data; registering the non-deformed single-frame point cloud data to a global coordinate system and performing voxel thinning to obtain a three-dimensional point cloud map, and taking the three-dimensional point cloud map as the prior map.

[0036] In this embodiment, the laser radar is installed on a mobile platform (such as a vehicle, a robot or a UAV) in the pitaya orchard, and it is ensured that it can stably obtain point cloud data. Data collection is performed according to a predetermined trajectory in the pitaya orchard, ensuring that the entire pitaya orchard area is covered, and special attention is paid to the position and distribution of pitaya trees.

[0037] In this embodiment, the scanning mode of the laser radar usually includes vertical (pitch angle) and horizontal (yaw angle) scanning. By analyzing the data format of the laser radar, the pitch angle and yaw angle information of each point cloud can be obtained, and according to these information, the point cloud data is distinguished according to the vertical direction and the horizontal direction, preparing for the subsequent feature point extraction.

[0038] In this embodiment, the edge points are usually located on the contour or edge of the object, and are sensitive to rotation and shape change. The planar points are located on the flat surface, and are sensitive to translation and surface change. Using the feature point extraction method in the LOAM algorithm, the edge points and planar points are distinguished according to the curvature or other geometric characteristics of the point cloud.

[0039] In this embodiment, during the scanning of the laser radar, due to the motion of the mobile platform, the point cloud data collected at different times in the same frame may be deformed by motion. Using the feature point matching and pose estimation method (such as the ICP algorithm or its improved version), the motion trajectory and pose change of the mobile platform are estimated according to the extracted edge points and planar points, and then the motion compensation is performed on the point cloud data collected at different times in the same frame according to the estimated pose change, so as to eliminate the motion deformation.

[0040] In this embodiment, the ICP algorithm or its improved version is used to register the non-deformed single-frame point cloud data with the global map. During the registration process, the pose of the single-frame point cloud data is constantly adjusted until the matching degree with the global map reaches the optimal.

[0041] Further, in order to reduce the redundancy of the point cloud data and improve the sparsity of the map, the voxel decimation method is used to down-sample the points in each voxel, and only the representative points (such as the center point, the centroid, etc.) in each voxel are retained to construct a sparse three-dimensional point cloud map.

[0042] In this embodiment, the initial pose of the laser radar is estimated, and the initial pose of the laser radar is optimized based on the prior map to obtain the optimized pose of the laser radar, including: loading the prior map, dividing the grid array in the prior map, and storing the global feature points into the corresponding grid; obtaining the IMU data of the laser radar, and determining the motion initial value of the laser radar according to the IMU data; obtaining the point cloud data of the current frame of the laser radar, extracting the current frame feature points from the point cloud data of the current frame, and determining the initial pose of the laser radar in the prior map according to the motion initial value and the current frame feature points; selecting the global feature points in the grid around the initial pose to form a local sub-map; performing feature matching on the current frame feature points and the global feature points in the local sub-map, and optimizing the initial pose according to the matching result to obtain the optimized pose.

[0043] In this embodiment, in a loose coupling manner, the spatial position data collected by the RTK-GNSS device and the optimized pose of the laser radar are fused, and the data of the RTK-GNSS and the data of the laser radar are time-synchronized before the fusion, so as to ensure that they correspond to the same time.

[0044] In this embodiment, a fusion algorithm such as weighted average, Kalman filter, etc. is selected to fuse the data of RTK-GNSS and lidar, and a fused pose is obtained. Then, according to the fused pose, the feature points at the corresponding moment are mapped into the prior map, and the above steps are repeatedly repeated to map the new feature points into the prior map, and a complete environment map is gradually constructed.

[0045] In this embodiment, A*, Dijkstra or other path planning algorithms can be used to plan an optimal path from the patrol starting point to the ending point in the environment map as the patrol route.

[0046] In this embodiment, when the patrol robot collects real-time images of pitaya trees according to the patrol route, a relocalization algorithm based on the prior map is designed based on the three-dimensional SLAM algorithm to improve the accuracy and stability of positioning, and the estimated value of the 3D SLAM algorithm is corrected through GNSS, so that the global positioning accuracy of the patrol robot meets the requirements of orchard autonomous navigation.

[0047] In other embodiments, the patrol robot can also be put into remote control mode, and control data is sent to the robot driver according to the instructions, so as to realize the patrol and photographing of pitaya.

[0048] In this embodiment, in step S103, the position information and fruit hanging amount of the pitaya tree are recognized according to the collected real-time images, and the maturity of the fruits of the pitaya tree is detected, which specifically includes: pre-processing the collected real-time images to obtain pre-processed images; using a target detection algorithm to mark the position of the pitaya tree in the pre-processed images to obtain a marked image; converting the position information of the pitaya tree in the marked image into actual position information according to the preset camera parameters; counting the number of fruits of the pitaya tree in the marked image to obtain the fruit hanging amount of the pitaya tree; and detecting the maturity of each fruit in the marked image, and outputting the maturity information of each fruit and the overall maturity distribution.

[0049] In this embodiment, the pre-processing of the collected real-time images aims to improve the image quality, reduce noise and interference, and provide clear input for the subsequent target detection algorithm. The pre-processing includes image denoising, contrast enhancement, size adjustment, etc.

[0050] In this embodiment, deep learning or computer vision technology (such as YOLO, Faster R-CNN, etc. target detection algorithm) is used to recognize pitaya trees in the image. The algorithm will output the position (usually in the form of a bounding box) and confidence of each detected pitaya tree. These position information will be used for subsequent marking and conversion.

[0051] In the present embodiment, by means of the intrinsic parameters (such as focal length, optical center position, etc.) and extrinsic parameters (pose and position of the camera relative to the world coordinate system) of the camera, the pixel coordinates in the marker image can be converted into actual three-dimensional space coordinates, thereby obtaining the exact position of the pitaya tree in the orchard.

[0052] In the present embodiment, by counting the number of fruits within the bounding box or using other segmentation techniques, the fruit load of each tree can be counted.

[0053] In the present embodiment, for each fruit in the marker image, maturity detection is performed, and the maturity information of each fruit and the maturity distribution of the pitaya tree as a whole are output, specifically including: using an image segmentation-based algorithm to extract the fruit region of the marker image to obtain a plurality of fruit images; extracting color features of the plurality of fruit images respectively, and inputting the extracted color features into a pre-trained SVM classifier to obtain a classification result output by the SVM classifier, the classification result including unripe fruits, semi-ripe fruits and ripe fruits; classifying the plurality of fruit images according to the classification result output by the SVM classifier, and counting the number and proportion of each category.

[0054] In the present embodiment, an image segmentation-based algorithm is applied to extract the fruit region, which can distinguish fruits from backgrounds, branches and leaves, etc., thereby obtaining a plurality of independent fruit images. The image segmentation-based algorithm includes threshold segmentation, region growing, edge detection, level set method or deep learning method (such as semantic segmentation network).

[0055] After segmentation, some post-processing steps can be performed, such as morphological operations (dilation, erosion, opening operation, closing operation) to remove noise, fill small holes or separate adhered fruits.

[0056] Finally, the fruit region is extracted from the segmented image, and each fruit image is assigned a unique number for subsequent processing.

[0057] In the present embodiment, since the color of pitaya changes from green (unripe) to red (ripe), color features can be extracted from each fruit image based on the HSV color space, specifically including average color value, color histogram, color moment, etc.

[0058] In the present embodiment, the pre-trained SVM classifier is trained based on samples including unripe fruits, semi-ripe fruits and ripe fruits. The SVM classifier will output a classification result for each fruit image according to the input color features, i.e., whether the fruit is unripe, semi-ripe or ripe.

[0059] In this implementation, according to the output results of the SVM classifier, the multiple fruit images are classified into corresponding categories (unripe fruit, semi-ripe fruit, ripe fruit). Then, the number of fruits in each category is calculated, and according to the number of fruits in each category, the proportion of them in the total number of fruits is calculated to obtain the maturity distribution of the whole pitaya tree.

[0060] In this embodiment, in step S104, according to the recognized position information, the identity card of the pitaya tree is accurately matched. The collected real-time image, the recognized fruit hanging amount, and the detected fruit maturity information are integrated and updated into the pitaya file corresponding to the identity card. At the same time, the time stamp at the time of updating is recorded for subsequent tracking and analysis.

[0061] For example, the position information of the pitaya tree is longitude 113.3°, latitude 23.1°, the basic information includes tree name pitaya tree A, tree age 5 years, variety red crystal, planting date March 15, 2018.

[0062] The pitaya tree identity card is as follows:

[0063] Tree name: pitaya tree A

[0064] Position: longitude 113.3°, latitude 23.1°

[0065] Tree age: 5 years

[0066] Variety: red crystal

[0067] Planting date: March 15, 2018.

[0068] The corresponding pitaya file initially only contains basic information, such as annual yield records, pest and disease conditions, etc. Now, according to the latest data collected by the patrol robot, the file of this pitaya tree is updated.

[0069] The updated pitaya file is as follows:

[0070] Tree name: pitaya tree A

[0071] Position: longitude 113.3°, latitude 23.1°

[0072] Tree age: 5 years

[0073] Variety: red crystal

[0074] Planting date: March 15, 2018

[0075] Growth monitoring record (newly added / updated)

[0076] Date: September 10, 2024

[0077] Real-time image: Attached are high-definition pictures of the dragon fruit tree A taken by the patrol robot today, which clearly show the overall growth condition and fruit distribution of the tree.

[0078] Fruit hanging amount: The current fruit hanging amount is 120.

[0079] Fruit maturity detection:

[0080] Ripe fruits: 30, bright red in color, with a full texture, expected to be picked within this week.

[0081] Semi-ripe fruits: 60, with a color transition from green to red, some areas still green, expected to reach the optimal picking period next week.

[0082] Unripe fruits: 30, mainly green in color, need to continue to grow.

[0083] Note: Today's observation found that the tree leaves are lush, no obvious signs of pests and diseases, and the overall growth condition is good.

[0084] It can be understood that the position information can also be recorded in the form of Orchard 3rd Zone, 5th Row, 2nd Column, etc.

[0085] The embodiment provides a dragon fruit growth monitoring method, which collects position information and basic information of a dragon fruit tree, establishes an identity card and a file, collects real-time images by using a patrol robot, and identifies the position of the dragon fruit tree, the fruit hanging amount, and the maturity of the fruits in the collected real-time images by using image recognition technology; finally, the dragon fruit file is updated according to the identification result and the detection result, realizing rapid and comprehensive monitoring of the growth state of the dragon fruit tree, greatly improving the monitoring efficiency, reducing manual intervention, and reducing the management cost.

[0086] The above describes the dragon fruit growth monitoring method in the embodiment of the application, and the device in the embodiment of the application is described below, please refer to Figure 2 The implementation of the dragon fruit growth monitoring device in the embodiment of the application includes:

[0087] The establishing module 201 is used for acquiring the position information and the basic information of the dragon fruit tree, and establishing the identity card of the dragon fruit tree and the dragon fruit file according to the position information and the basic information of the dragon fruit tree;

[0088] The planning module 202 is used for planning a patrol route, and controlling the patrol robot to collect real-time images of the dragon fruit tree according to the patrol route;

[0089] The identification and detection module 203 is used for identifying the position information and the fruit hanging amount of the dragon fruit tree according to the collected real-time images, and detecting the maturity of the fruits of the dragon fruit tree;

[0090] The updating module 204 confirms the identity card of the pitaya tree according to the identified position information of the pitaya tree, and updates the pitaya file corresponding to the identity card of the pitaya tree according to the collected real-time image, the identified hanging amount of the pitaya tree, and the detected ripeness of the fruits of the pitaya tree.

[0091] In the embodiment, the position information and the basic information of the pitaya tree are collected to establish the identity card and the file, the real-time image is collected by the patrol robot, the position of the pitaya tree, the fruit hanging amount, and the ripeness of the fruits in the collected real-time image are identified by the image recognition technology, and finally, the pitaya file is updated according to the identification result and the detection result, so that the rapid and comprehensive monitoring of the growth state of the pitaya tree is realized, the monitoring efficiency is greatly improved, the manual intervention is reduced, and the management cost is reduced.

[0092] Figure 2 The structure of the illustrated pitaya growth monitoring device does not constitute a limitation on the pitaya growth monitoring device, and the steps of the pitaya growth monitoring method provided in each method embodiment can be implemented.

[0093] The above Figure 2 The pitaya growth monitoring device in the embodiment of the application is described in detail from the perspective of a modular functional entity, and the pitaya growth monitoring device in the embodiment of the application is described in detail from the perspective of hardware processing.

[0094] Figure 3 Fig. 3 is a structural schematic diagram of a pitaya growth monitoring device provided in the embodiment of the application. The device 300 can have great differences due to different configurations or performances, and can include one or more processors (central processing units, CPUs) 310 (for example, one or more processors) and a memory 320, and one or more storage media 330 (for example, one or more mass storage devices) storing application programs 333 or data 332. The memory 320 and the storage media 330 can be temporary storage or persistent storage. The programs stored in the storage media 330 can include one or more modules (not shown in the figure), and each module can include a series of instruction operations in the device 300. Furthermore, the processor 310 can be configured to communicate with the storage media 330 to execute a series of instruction operations in the storage media on the device 300.

[0095] The device 300 can also include one or more power supplies 340, one or more wired or wireless network interfaces 350, one or more input and output interfaces 360, and / or one or more operating systems 331, such as Windows Serve, Mac OS X, Unix, Linux, FreeBSD, and the like.

[0096] The present application also provides a computer readable storage medium, which can be a non-volatile computer readable storage medium or a volatile computer readable storage medium. The computer readable storage medium stores instructions, which, when executed on a computer, cause the computer to perform the steps of the method for monitoring growth of pitaya.

[0097] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working process of the system or device, unit described above can refer to the corresponding process in the foregoing method embodiments, which will not be described here.

[0098] The integrated unit, if implemented in the form of a software function unit and sold or used as an independent product, can be stored in a computer readable storage medium. Based on such understanding, the technical solutions of the present application essentially or the part that contributes to the prior art or the whole or part of the technical solutions can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes a number of instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present application. The foregoing storage medium includes: a U disk, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, and various program code storage media.

[0099] The above-described embodiments are merely used to illustrate the technical solutions of the present application, rather than limit the present application; even though the present application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that: they can still modify the technical solutions recorded in the foregoing embodiments, or make equivalent replacements to some technical features; and these modifications or replacements do not cause the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A pitaya growth monitoring method, characterized in that, The pitaya growth monitoring method comprises: Obtaining the location information and basic information of the pitaya tree, and establishing an identity card and a pitaya file for the pitaya tree based on the location information and basic information of the pitaya tree; Plan patrol routes and control patrol robots to collect real-time images of dragon fruit trees along the patrol routes; Identify the location information and fruit quantity of the dragon fruit tree based on the collected real-time images, and detect the maturity of the fruit of the dragon fruit tree; Confirm the identity card of the dragon fruit tree according to the position information of the identified dragon fruit tree, and update the dragon fruit file corresponding to the identity card of the dragon fruit tree according to the real-time image collected, the hanging amount of the identified dragon fruit tree, and the fruit maturity of the dragon fruit tree detected; The method comprises planning a patrol route and controlling a patrol robot to collect real-time images of pitaya trees according to the patrol route, comprising: using a laser radar to collect point cloud data in an area between two ends of each row of pitaya trees, and using an RTK-GNSS device to collect spatial position data in an area at two ends of each row of pitaya trees; constructing a three-dimensional point cloud map as a priori map based on the point cloud data; estimating an initial position of the laser radar, and optimizing the initial position of the laser radar based on the priori map to obtain an optimized position of the laser radar; adopting a loose coupling method to fuse the spatial position data collected by the RTK-GNSS device with the optimized position of the laser radar to obtain a fused position, and mapping the feature points at corresponding moments to the priori map according to the fused position to construct an environmental map; obtaining a patrol starting point and a patrol end point, planning a patrol route in the environmental map using a path planning algorithm, and controlling the patrol robot to collect real-time images of pitaya trees according to the patrol route; The method of estimating the initial pose of the laser radar and optimizing the initial pose of the laser radar based on the prior map to obtain the optimized pose of the laser radar includes: loading the prior map, dividing the grid array in the prior map, and storing the global feature points of the point cloud data in the corresponding grids; obtaining the IMU data of the laser radar, and determining the initial motion value of the laser radar based on the IMU data; obtaining the point cloud data of the current frame of the laser radar, extracting the current frame feature points from the point cloud data of the current frame, and determining the initial pose of the laser radar in the prior map based on the initial motion value of the laser radar and the current frame feature points; selecting the global feature points in the grid around the initial pose to form a local subgraph; performing feature matching between the current frame feature points and the global feature points in the local subgraph, and optimizing the initial pose of the laser radar based on the matching results to obtain the optimized pose of the laser radar.

2. The method for monitoring growth of pitaya according to claim 1, wherein The method of obtaining the position information and basic information of the pitaya tree and establishing an identity card and a pitaya file of the pitaya tree according to the position information and basic information of the pitaya tree comprises: Use a drone or ground measurement equipment to measure and record the location information of the dragon fruit tree, wherein the location information includes longitude and latitude; Obtain basic information about the pitaya tree, including tree age, crown size, and health status, and integrate the location information and basic information of the pitaya tree to create an identity card for the pitaya tree; A pitaya file corresponding to the identity card of the pitaya tree is established in a preset pitaya growth database.

3. The method for monitoring growth of pitaya according to claim 1, wherein The step of constructing a three-dimensional point cloud map as a priori map based on the point cloud data includes: According to the scanning mode of the laser radar, the point cloud data is distinguished in the vertical direction and the horizontal direction, and edge points and plane points are extracted from the distinguished point cloud data as global feature points; Based on the global feature points, motion deformation compensation is performed on the point cloud data collected at different times in the same frame to obtain single-frame point cloud data without deformation; The undeformed single-frame point cloud data is registered to the global coordinate system and voxel thinning is performed to obtain a three-dimensional point cloud map, which is used as a priori map.

4. The method for monitoring growth of pitaya according to claim 1, wherein The method of identifying the position information and the amount of fruit on the pitaya tree according to the collected real-time image, and detecting the fruit maturity of the pitaya tree, comprises: Preprocessing the collected real-time image to obtain a preprocessed image; Marking the position of the pitaya tree in the preprocessed image using a target detection algorithm to obtain a marked image; According to preset camera parameters, the position information of the pitaya tree in the marked image is converted into actual position information; Counting the number of fruits of the pitaya tree in the marked image to obtain the fruit yield of the pitaya tree; For each fruit in the marked image, maturity detection is performed, and maturity information of each fruit and overall maturity distribution are output.

5. The method for monitoring the growth of pitaya according to claim 4, wherein The process of performing maturity detection on each fruit in the marked image and outputting maturity information of each fruit and overall maturity distribution includes: Extracting the fruit region of the marked image using an algorithm based on image segmentation to obtain multiple fruit images; Extracting color features of the plurality of fruit images respectively, and inputting the extracted color features into a pre-trained SVM classifier to obtain classification results output by the SVM classifier, the classification results including immature fruit, semi-mature fruit and mature fruit; The plurality of fruit images are classified according to the classification results output by the SVM classifier, and the number and proportion of fruit images in each category are counted.

6. A pitaya growth monitoring device, characterized in that: include: Establish a module for obtaining the location information and basic information of the pitaya tree, and establish an identity card and a pitaya file of the pitaya tree according to the location information and basic information of the pitaya tree; The planning module is used to plan the patrol route and control the patrol robot to collect real-time images of pitaya trees along the patrol route, specifically including: using the laser radar to collect point cloud data in the area between the two ends of each row of pitaya trees, and using the RTK-GNSS device to collect spatial position data in the area at the two ends of each row of pitaya trees; constructing a three-dimensional point cloud map as a priori map based on the point cloud data; estimating the initial position of the laser radar, and optimizing the initial position of the laser radar based on the priori map to obtain the optimized position of the laser radar; adopting a loose coupling method to fuse the spatial position data collected by the RTK-GNSS device with the optimized position of the laser radar to obtain a fused position, and mapping the feature points at the corresponding time to the priori map according to the fused position to construct an environmental map; obtaining the patrol starting point and patrol end point, using the path planning algorithm to plan a patrol route in the environmental map, and controlling the patrol robot to follow the patrol route The method comprises the following steps: collecting a real-time image of a pitaya tree by line; estimating an initial pose of a laser radar, and optimizing the initial pose of the laser radar based on the prior map to obtain an optimized pose of the laser radar, comprising: loading the prior map, dividing the grid array in the prior map, and storing the global feature points of the point cloud data in corresponding grids; obtaining IMU data of the laser radar, and determining an initial motion value of the laser radar according to the IMU data; obtaining point cloud data of a current frame of the laser radar, extracting feature points of the current frame from the point cloud data of the current frame, and determining the initial pose of the laser radar in the prior map according to the initial motion value and the feature points of the current frame; selecting global feature points in the grid around the initial pose to form a local subgraph; performing feature matching between the feature points of the current frame and the global feature points in the local subgraph, and optimizing the initial pose of the laser radar according to the matching result to obtain an optimized pose of the laser radar; An identification and detection module is used to identify the location information and fruit quantity of the pitaya tree based on the collected real-time image, and detect the maturity of the fruit of the pitaya tree; The updating module confirms the identity card of the pitaya tree according to the identified location information of the pitaya tree, and updates the pitaya file corresponding to the identity card of the pitaya tree according to the collected real-time image, the identified hanging amount of the pitaya tree and the detected fruit maturity of the pitaya tree.

7. A pitaya growth monitoring device, characterized in that: comprising a memory and at least one processor, wherein the memory has computer-readable instructions stored therein; The at least one processor calls the computer-readable instructions in the memory to execute each step of the dragon fruit growth monitoring method according to any one of claims 1 to 5.

8. A computer-readable storage medium having computer-readable instructions stored thereon, characterized in that: When the computer-readable instructions are executed by a processor, the steps of the dragon fruit growth monitoring method according to any one of claims 1 to 5 are implemented.

Images