Methods, apparatus, computer equipment and computer-readable storage media for identifying infected trees

Abstract

This application provides a method, apparatus, computer device, and computer-readable storage medium for identifying infected trees. The method includes: acquiring a set of remote sensing images collected for an area to be identified; inputting the remote sensing image set into a trained infected tree identification model for infected tree identification to obtain a target remote sensing image containing infected tree location information; acquiring the spatial coordinate information corresponding to the infected tree location information, and filtering and merging the infected tree location information according to the spatial coordinate information and the image acquisition time of the target remote sensing image to obtain the location of infected trees in the area to be identified. Using this application can effectively improve the accuracy of infected tree identification.

Description

Technical Field

[0001] This application relates to the field of monitoring trees infected with pine wilt disease, specifically to a method, device, computer equipment, and computer-readable storage medium for identifying infected trees. Background Technology

[0002] Pine wilt disease is a devastating disease of pine trees caused by the pine wilt nematode. It is currently the most serious forestry quarantine pest that damages my country's forests. The disease spreads rapidly and has a strong ability to adapt. Therefore, quickly finding and removing infected trees (commonly known as infected trees) is crucial for protecting the ecological health of forest areas.

[0003] Currently, the industry's method for locating pine wilt disease-infected trees typically involves using drones to fly at low altitudes and capture remote sensing images. These images are then stitched together to obtain a composite image of the monitored area. Deep learning methods are then used to perform target detection on the stitched image to locate the infected trees. However, to ensure no area is missed during the capturing process, images from adjacent capture points have a high overlap rate. This leads to problems such as image misalignment, distortion, and gaps in overlapping areas during image stitching, hindering the identification of infected trees.

[0004] Therefore, existing methods for identifying infected trees suffer from low accuracy. Summary of the Invention

[0005] Therefore, it is necessary to provide a method, device, computer equipment, and computer-readable storage medium for identifying diseased trees in response to the above-mentioned technical problems, so as to improve the accuracy of identifying pine wilt disease-infected trees.

[0006] Firstly, this application provides a method for identifying infected trees, including:

[0007] Acquire a set of remote sensing images for the area to be identified;

[0008] The remote sensing image set is input into the trained diseased tree identification model to identify diseased trees and obtain the target remote sensing image containing the location information of diseased trees.

[0009] The spatial coordinates of the infected trees are obtained. Based on the spatial coordinates and the image acquisition time of the target remote sensing image, the infected tree location information is filtered and merged to obtain the location of the infected trees in the area to be identified.

[0010] In some embodiments of this application, the region to be identified includes at least two flight path regions to be identified. Acquiring a set of remote sensing images for the region to be identified includes: acquiring a first remote sensing image for a first flight path region and preprocessing the first remote sensing image to obtain a first remote sensing image set; and acquiring a second remote sensing image for a second flight path region and preprocessing the second remote sensing image to obtain a second remote sensing image set; wherein the first flight path region and the second flight path region are adjacent flight path regions whose corresponding image overlap rate meets a preset condition; and determining the first remote sensing image set and the second remote sensing image set as a remote sensing image set.

[0011] In some embodiments of this application, acquiring a first remote sensing image for a first flight path area and preprocessing the first remote sensing image to obtain a first remote sensing image set includes: acquiring a first remote sensing image for a first flight path area; performing frame extraction processing on the first remote sensing image to obtain at least two frames of first ground remote sensing images; and performing sub-image cropping processing on each first ground remote sensing image according to a preset cutting size and overlap size and a preset cropping direction to obtain a first remote sensing image set; wherein the cropping direction includes a horizontal cropping direction and / or a vertical cropping direction.

[0012] In some embodiments of this application, the remote sensing image set is input into a trained diseased tree identification model for diseased tree identification to obtain a target remote sensing image containing diseased tree location information. This includes: inputting the remote sensing image set into the trained diseased tree identification model for diseased tree identification, outputting candidate remote sensing images marked with target diseased tree points, and image coordinate information of the target diseased tree points; determining the origin coordinate information of the candidate remote sensing images, and performing coordinate restoration processing on the image coordinate information based on the origin coordinate information to obtain the diseased tree location information of the target diseased tree points; and determining the candidate remote sensing image containing the diseased tree location information as the target remote sensing image.

[0013] In some embodiments of this application, the target remote sensing image is at least two frames of target remote sensing images, and the infected tree location information includes first infected tree location information and second infected tree location information belonging to different target remote sensing images; wherein, obtaining the spatial coordinate information corresponding to the infected tree location information, and filtering and merging the infected tree location information according to the spatial coordinate information and the image acquisition time of the target remote sensing image to obtain the infected tree location in the area to be identified, includes: obtaining and according to the image acquisition time of the target remote sensing image, obtaining the first spatial coordinate information corresponding to the first infected tree location information and the second spatial coordinate information corresponding to the second infected tree location information to obtain spatial coordinate information; calculating the distance value between the first spatial coordinate information and the second spatial coordinate information to filter out the infected tree location information with a distance value less than a preset distance threshold as the target location combination; extracting the location image features of each of the target location combinations, and merging the infected tree location information based on the location image features to analyze and obtain the infected tree location in the area to be identified.

[0014] In some embodiments of this application, the target remote sensing image is at least two frames of target remote sensing images containing GPS coordinate information; wherein, obtaining and based on the image acquisition time of the target remote sensing image, obtaining the first spatial coordinate information corresponding to the first infected tree location information and the second spatial coordinate information corresponding to the second infected tree location information to obtain spatial coordinate information includes: obtaining and based on the image acquisition time of the target remote sensing image, sequentially arranging each target remote sensing image to obtain a remote sensing image sequence; extracting the target remote sensing images of two adjacent frames in the remote sensing image sequence, respectively as the first image and the second image, and using the infected tree location information contained in the first image as the first infected tree location information and the infected tree location information contained in the second image as the second infected tree location information; based on the GPS coordinate information, obtaining the first spatial coordinate information corresponding to the first infected tree location information and the second spatial coordinate information corresponding to the second infected tree location information to obtain spatial coordinate information.

[0015] In some embodiments of this application, the point image features of each of the target point combination are extracted, and the diseased tree point information is merged based on the point image features to analyze and obtain the location of the diseased tree in the area to be identified. This includes: extracting the point image features of each of the diseased tree point information contained in the target point combination, and using them as the first point image feature and the second point image feature respectively; calculating the similarity value between the first point image feature and the second point image feature, so as to select the target point combination with a similarity value greater than a preset similarity threshold for point merging processing to obtain the target point information; obtaining the spatial coordinate information corresponding to the target point information, and the spatial coordinate information corresponding to the unmerged diseased tree point information, as the location of the diseased tree in the area to be identified.

[0016] Secondly, this application provides a diseased wood identification device, comprising:

[0017] The image acquisition module is used to acquire a set of remote sensing images of the area to be identified;

[0018] The diseased tree identification module is used to input the remote sensing image set into the trained diseased tree identification model to identify diseased trees and obtain target remote sensing images containing diseased tree location information.

[0019] The location analysis module is used to obtain the spatial coordinate information corresponding to the location of infected trees. Based on the spatial coordinate information and the image acquisition time of the target remote sensing image, the location information of infected trees is filtered and merged to obtain the location of infected trees in the area to be identified.

[0020] Thirdly, this application also provides a computer device, comprising:

[0021] One or more processors;

[0022] The memory; and one or more applications, wherein the one or more applications are stored in the memory and configured to be executed by the processor to implement the above-described method for identifying infected trees.

[0023] Fourthly, this application also provides a computer-readable storage medium having a computer program stored thereon, the computer program being loaded by a processor to perform the steps in the above-described method for identifying infected animals.

[0024] Fifthly, embodiments of this application provide a computer program product or computer program that includes computer instructions stored in a computer-readable storage medium. A processor of a computer device reads the computer instructions from the computer-readable storage medium and executes the computer instructions, causing the computer device to perform the method provided in the first aspect described above.

[0025] The aforementioned method, apparatus, computer equipment, and computer-readable storage medium for identifying infected trees involve a server acquiring a set of remote sensing images of the area to be identified. This set of images is then input into a trained infected tree identification model to obtain a target remote sensing image containing infected tree location information. Next, the spatial coordinates corresponding to these locations are acquired. Based on the image acquisition time and spatial coordinates of the target remote sensing image, the infected tree location information is filtered and merged to ultimately determine the location of the infected trees in the area to be identified. Therefore, the infected tree identification scheme proposed in this application eliminates the need for image stitching, saving time and avoiding image feature damage caused by stitching, thus effectively improving the accuracy of infected tree identification. Attached Figure Description

[0026] To more clearly illustrate the technical solutions in the embodiments of this application, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0027] Figure 1 This is an application scenario diagram of the diseased tree identification method provided in the embodiments of this application;

[0028] Figure 2 This is a flowchart illustrating the method for identifying infected trees provided in an embodiment of this application;

[0029] Figure 3 This is a flowchart illustrating the sub-image cropping steps provided in an embodiment of this application;

[0030] Figure 4 This is a flowchart illustrating the point merging steps provided in an embodiment of this application;

[0031] Figure 5 This is a flowchart illustrating the model inference steps provided in an embodiment of this application;

[0032] Figure 6 This is a schematic diagram of the structure of the diseased tree identification device provided in the embodiments of this application;

[0033] Figure 7 This is a schematic diagram of the structure of the computer device in the embodiments of this application. Detailed Implementation

[0034] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0035] In the description of this application, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of indicated technical features. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of the stated features. In the description of this application, "multiple" means two or more, unless otherwise explicitly specified.

[0036] In the description of this application, the term "for example" is used to mean "used as an example, illustration, or description." Any embodiment described as "for example" in this application is not necessarily to be construed as being more preferred or advantageous than other embodiments. The following description is provided to enable any person skilled in the art to make and use the invention. Details are set forth in the following description for purposes of explanation. It should be understood that those skilled in the art will recognize that the invention can be made without using these specific details. In other instances, well-known structures and processes will not be described in detail to avoid obscuring the description of the invention with unnecessary detail. Therefore, the invention is not intended to be limited to the embodiments shown, but is consistent with the broadest scope of the principles and features disclosed in this application.

[0037] In this application's embodiments, the epidemic identification method mainly involves computer vision (CV) technology within artificial intelligence (AI). Artificial intelligence utilizes digital computers or machines controlled by digital computers to simulate, extend, and expand human intelligence, perceiving the environment, acquiring knowledge, and using that knowledge to obtain optimal results—theories, methods, technologies, and application systems. In other words, artificial intelligence is a comprehensive technology within computer science that attempts to understand the essence of intelligence and produce a new kind of intelligent machine capable of reacting in a manner similar to human intelligence.

[0038] Computer vision is a science that studies how to enable machines to "see." More specifically, it refers to machine vision, which uses cameras and computers to replace human eyes in recognizing, tracking, and measuring targets, and then performs image processing to create images more suitable for human observation or transmission to instruments for detection. As a scientific discipline, computer vision studies related theories and technologies, attempting to build artificial intelligence systems capable of extracting information from images or multidimensional data. Computer vision technologies typically include image processing, image recognition, image semantic understanding, image retrieval, optical character recognition (OCR), video processing, video semantic understanding, video content / behavior recognition, 3D object reconstruction, 3D technology, virtual reality, augmented reality, simultaneous localization and mapping (SLAM), and also common biometric recognition technologies such as face recognition and fingerprint recognition. In this application, for the image to be detected, CV mainly implements image detection within Image Semantic Understanding (ISU), detecting target objects in the image and outputting the detection results. It is understood that the target object can be any object determined by actual business needs, such as people, vehicles, packages, or other people or things. However, in this embodiment, the target object can refer to trees infected with pine wilt disease. Of course, the diseased tree identification method proposed in this embodiment can also be used to identify trees with other plant diseases.

[0039] This application provides a method, apparatus, computer device, and computer-readable storage medium for identifying infected trees, which will be described in detail below.

[0040] See Figure 1 , Figure 1This is a schematic diagram illustrating a scenario for the diseased tree identification method provided in this application, which can be applied to a diseased tree identification system. The diseased tree identification system includes a terminal 102 and a server 104. The terminal 102 can be a device that includes both receiving and transmitting hardware, i.e., a device with receiving and transmitting hardware capable of performing bidirectional communication over a bidirectional communication link. Such a device can include cellular or other communication devices, having a single-line display, a multi-line display, or no display. Specifically, the terminal 102 can be a desktop terminal or a mobile terminal; it can also be a mobile phone, tablet computer, laptop computer, monocular camera, multi-view camera, or drone. The server 104 can be a standalone server or a server network or server cluster, including but not limited to computers, network hosts, single network servers, multiple network server sets, or cloud servers composed of multiple servers. The cloud server consists of a large number of computers or network servers based on cloud computing. Furthermore, the terminal 102 and the server 104 establish a communication connection through a network, which can be any of a wide area network (WAN), local area network (LAN), or metropolitan area network (MAN).

[0041] Those skilled in the art will understand that Figure 1 The application environment shown is merely one applicable scenario for the solution in this application and does not constitute a limitation on the application scenario of the solution in this application. Other application environments may include more than one. Figure 1 The number of computer devices shown is more or less, for example Figure 1 Only one server, 104, is shown in the image. It is understood that this epidemic identification system may also include one or more other servers; specific details are not specified here. Additionally, as... Figure 1 As shown, the epidemic identification system may also include a memory for storing data, such as remote sensing images collected by drones.

[0042] It should be noted that, Figure 1 The schematic diagram of the epidemic tree identification system shown is merely an example. The epidemic tree identification system and scenario described in this embodiment of the invention are for the purpose of more clearly illustrating the technical solutions of this embodiment of the invention, and do not constitute a limitation on the technical solutions provided by this embodiment of the invention. As those skilled in the art will know, with the evolution of the epidemic tree identification system and the emergence of new business scenarios, the technical solutions provided by this embodiment of the invention are also applicable to similar technical problems.

[0043] See Figure 2 This application provides a method for identifying infected trees. This embodiment mainly applies this method to the above-mentioned... Figure 1 Taking server 104 as an example, the method includes steps S201 to S203, as follows:

[0044] S201, Acquire the set of remote sensing images collected for the area to be identified.

[0045] The area to be identified can be a preset mountain forest, agricultural forest or forest area where it is necessary to identify whether there are infected trees. The shape of the area can be any shape, such as a circle, square, rectangle or any polygon. The specific implementation of this application is not limited.

[0046] The remote sensing image set can refer to a collection of at least two remote sensing images. These images can be those from which the location of the infected tree needs to be identified, including but not limited to pictures and video frames within a video. The video can be in formats such as short videos and long videos. Short videos can be less than 10 minutes long, and long videos can be more than 10 minutes long. Of course, the definition of video length is not limited to "10 minutes," and those skilled in the art can set other criteria for defining short and long videos according to actual business needs. Specific embodiments of this application do not limit this definition.

[0047] In specific implementation, to improve the accuracy of diseased tree identification, this application proposes combining artificial intelligence technology to identify diseased trees from remote sensing image sets collected for the area to be identified. Specifically, a target detection algorithm is used to directly identify diseased tree points in the remote sensing images. The identified diseased tree points are then merged using an association algorithm. Finally, the merged diseased tree points and the remaining unmerged diseased tree points are output as the result. Since this method bypasses the remote sensing image stitching process, it significantly saves the time required for image stitching and avoids the problem of reduced identification accuracy due to image feature damage to the target area during stitching. This results in a significant improvement in both time consumption and the quality of the identification results.

[0048] Furthermore, the remote sensing image set required for epidemic tree identification can be acquired by terminal 102 and sent to server 104, or it can be acquired by other devices and transmitted to server 104 via terminal 102. The acquisition methods of the remote sensing image set include, but are not limited to, one of the following: 1. In a normal network structure, server 104 receives the remote sensing image set from terminal 102 or other cloud devices with established network connections; 2. In a pre-built blockchain network, server 104 can synchronously acquire the remote sensing image set from other terminal nodes or server nodes. This blockchain network can be a public chain, a private chain, etc.; 3. In a pre-built tree structure, server 104 can request the remote sensing image set from an upper-level server or obtain the remote sensing image set by polling from a lower-level server. Specifically, the embodiments of this application do not specifically limit the acquisition method of the remote sensing image set, and can be determined according to the actual business scenario or needs. Terminal 102 can be an unmanned aerial vehicle carrying image acquisition equipment, such as a drone.

[0049] In one embodiment, the area to be identified includes at least two flight path areas to be identified. Step S201 includes: acquiring a first remote sensing image for a first flight path area and preprocessing the first remote sensing image to obtain a first remote sensing image set; acquiring a second remote sensing image for a second flight path area and preprocessing the second remote sensing image to obtain a second remote sensing image set; wherein the first flight path area and the second flight path area are adjacent flight path areas whose corresponding image overlap rate meets a preset condition; and determining the first remote sensing image set and the second remote sensing image set as a remote sensing image set.

[0050] Among them, the flight path area is a sub-region of the area to be identified that is divided for the flight path of the UAV. The following will use the first flight path area and the second flight path area as examples for explanation.

[0051] The image overlap rate refers to the percentage similarity in pixel content between remote sensing images collected by terminal 102 (such as a drone) along the first flight path area and remote sensing images collected along the second flight path area. The corresponding preset conditions can be set according to actual application needs. For example, setting it to "60%" means that the image overlap rate of the selected adjacent flight path areas is not less than "60%" to ensure that no area is missed.

[0052] In specific implementation, server 104 can perform frame extraction processing on the first remote sensing image / second remote sensing image using image processing tools, such as "OpenCV" or "ffmpeg". The frame extraction frequency can be once per second or multiple times per second; the number of frames extracted can be one frame or multiple frames at a time, which is not limited in this application. After acquiring the first remote sensing image and the second remote sensing image and performing video image frame extraction processing, server 104 will obtain more than one first video frame as the first remote sensing image set and more than one second video frame as the second remote sensing image set.

[0053] Therefore, by acquiring the first remote sensing image set and the second remote sensing image set, the server 104 can obtain the remote sensing image set required to complete the current epidemic identification task.

[0054] It should be noted that in this embodiment, the area to be identified is divided into a first route area and a second route area in order to improve the accuracy of diseased tree identification. The improvement principle is mainly reflected in the subsequent association algorithm that is more difficult to achieve in terms of spatial dimension. That is, the area to be identified is divided into multiple sub-regions for diseased tree identification, and then the diseased tree location is analyzed and merged using a target association algorithm. This can improve the accuracy of diseased tree identification, thereby improving the accuracy of diseased tree identification. The specific implementation steps will be explained in detail below.

[0055] In one embodiment, acquiring a first remote sensing image for a first flight path area and preprocessing the first remote sensing image to obtain a first remote sensing image set includes: acquiring the first remote sensing image for a first flight path area; performing frame extraction processing on the first remote sensing image to obtain at least two frames of first ground remote sensing images; and performing sub-image cropping processing on each first ground remote sensing image according to a preset cutting size and overlap size and a preset cropping direction to obtain a first remote sensing image set; wherein the cropping direction includes a horizontal cropping direction and / or a vertical cropping direction.

[0056] The cutting size can be the size of the sub-image extracted from the original image. For example, if the original image is the first ground remote sensing image obtained by extracting frames from the first remote sensing image and the original image size is "6504*6336", then the cutting size can be "1024*1024", which means that the original image can be cut into several sub-images of size "1024*1024".

[0057] The overlap size can be the allowed overlap pixel size when cropping the original image. For example, if the overlap size is set to "X=200", it means that when cropping in the horizontal direction, the overlap area between the sub-images is "200" pixels. If the overlap size is set to "Y=100", it means that when cropping in the vertical direction, the overlap area between the sub-images is "100" pixels.

[0058] The screenshot direction includes the horizontal screenshot direction and / or the vertical screenshot direction. The horizontal screenshot direction can be regarded as the "X-axis" direction of the screen coordinate system, and the vertical screenshot direction can be regarded as the "Y-axis" direction of the screen coordinate system. The origin of the screen coordinate system can be set at the upper left corner.

[0059] For specific implementation details, please refer to [link / reference]. Figure 3 The server 104 acquires the first remote sensing image collected by the terminal 102 for the first flight path area, and performs video image frame extraction processing on the first remote sensing image to obtain multiple frames of the first ground remote sensing image. Then, it can perform screenshot processing on each frame of the first ground remote sensing image. For example, according to preset cutting and overlap sizes, sub-images can be extracted along the screenshot direction, thus extracting multiple sub-images from each frame of the first ground remote sensing image and packaging them to obtain the first remote sensing image set. It is understood that the screenshot method described in this embodiment is a sliding window type. The screenshot direction of the sliding window type can be horizontal first and then vertical, or vertical first and then horizontal, as long as the cutting and overlap sizes are consistent in both directions.

[0060] Furthermore, after the server 104 processes and obtains the first remote sensing image set, it can acquire the second remote sensing image set in the same way; specific details will not be elaborated in this application embodiment. After obtaining the first and second remote sensing image sets, the image sets can be input into a trained epidemic identification model (such as...). Figure 3The model (as shown in "Model 1") is used to detect infected trees so that the model can output a remote sensing image containing the location information of infected trees based on the infected tree detection capabilities it has learned.

[0061] S202, input the remote sensing image set into the trained diseased tree identification model to identify diseased trees and obtain the target remote sensing image containing the location information of diseased trees.

[0062] The trained target detection model can be any model with target detection capabilities, such as Faster R-CNN, SSD, or YOLO, and the model type is not limited in this application embodiment.

[0063] Among them, the location information of infected trees can refer to the image coordinate information of the infected tree points detected in the remote sensing image, such as "(x, y)".

[0064] In specific implementation, before the server 104 inputs the remote sensing image set into the trained diseased tree identification model, it needs to train the diseased tree identification model. The model training steps may include: constructing an initial diseased tree identification model (such as Faster R-CNN, SSD, or YOLO model); acquiring a ground image set and dividing the ground image set into a training set and a test set; the ground image set includes multiple ground images with labeled diseased tree locations; using the training set to perform preliminary training on the initial diseased tree identification model to obtain a preliminary trained diseased tree identification model; and using the test set to test and adjust the preliminary trained diseased tree identification model to obtain a trained diseased tree identification model.

[0065] Furthermore, after server 104 inputs the remote sensing image set into the trained infected tree identification model, the model will infer the meaning of each image in the remote sensing image set and output remote sensing images that are marked with the target object—"infected tree spot" (e.g., indicated by model number, or outlined with a rectangle). These are the target remote sensing images, while remote sensing images that do not detect "infected tree spots" are non-target remote sensing images. It is understandable that the trained infected tree identification model can not only identify and mark "infected tree spots," but also output their image coordinate information. The purpose of the image coordinate information is to determine the location of the infected tree in the real scene, which will be explained in detail below.

[0066] In one embodiment, step S202 includes: inputting a remote sensing image set into a trained diseased tree identification model for diseased tree identification, outputting candidate remote sensing images marked with target diseased tree points, and image coordinate information of the target diseased tree points; determining the origin coordinate information of the candidate remote sensing images, and performing coordinate restoration processing on the image coordinate information based on the origin coordinate information to obtain the diseased tree point location information of the target diseased tree points; and determining the candidate remote sensing image containing the diseased tree point location information as the target remote sensing image.

[0067] Among them, the origin coordinate information can refer to the image coordinate information of the origin at the upper left corner of the corresponding candidate remote sensing image. The origin coordinate information can be represented as "(x0, y0)" and the image coordinate information can be represented as "(x, y)".

[0068] In the specific implementation, the previous embodiment mentioned that the remote sensing image marked with the target object - "epidemic point" can be used as the target remote sensing image. However, in this embodiment, it needs to be used as a candidate remote sensing image. The reason is that this embodiment is based on the sub-image cropping mentioned in the previous steps. If there is no sub-image cropping step, then there is no need to apply the coordinate restoration scheme proposed in this embodiment.

[0069] For details, please refer to Figure 3 When server 104 cuts a sub-image from the first or second ground remote sensing image, the coordinates of the top-left corner of the sub-image (i.e., the origin coordinates) are "(x0, y0)". Then, the image coordinates of a target infected tree point obtained using the infected tree detection algorithm are "(x, y)". Therefore, it is necessary to analyze the coordinates of this target infected tree point on the original image as "(x0+x, y0+y)" to obtain the new infected tree point location information. Only then can the remote sensing image with updated infected tree point information be used as the target remote sensing image. In other words, if no sub-image is cut, the target remote sensing image contains the infected tree point location information "(x, y)"; but if a sub-image is cut, the target remote sensing image contains the infected tree point location information "(x0+x, y0+y)".

[0070] S203, obtain the spatial coordinate information corresponding to the location information of the infected trees, and filter and merge the location information of the infected trees according to the spatial coordinate information and the image acquisition time of the target remote sensing image to obtain the location of the infected trees in the area to be identified.

[0071] Among them, the location information of the infected tree is the image coordinate information corresponding to the screen coordinate system, while the spatial coordinate information is the spatial coordinate information corresponding to the world coordinate system, which can be represented by latitude and longitude.

[0072] In specific implementation, after the server 104 obtains the location information of the infected trees in the remote sensing image set based on the scheme described in the above embodiment, it can further obtain the spatial coordinate information of each infected tree location information contained in the image based on the GPS information contained in each target remote sensing image, so as to perform coordinate filtering and merging processing based on the spatial coordinate information, thereby analyzing the real location of the infected trees in the area to be identified.

[0073] In one embodiment, the target remote sensing image is at least two frames of target remote sensing images, and the infected tree location information includes first infected tree location information and second infected tree location information belonging to different target remote sensing images. Step S203 includes: acquiring and according to the image acquisition time of the target remote sensing image, acquiring the first spatial coordinate information corresponding to the first infected tree location information and the second spatial coordinate information corresponding to the second infected tree location information to obtain spatial coordinate information; calculating the distance value between the first spatial coordinate information and the second spatial coordinate information to filter out infected tree location information with a distance value less than a preset distance threshold as target location combination; extracting the location image features of each target location combination, and merging the infected tree location information based on the location image features to analyze and obtain the location of infected trees in the area to be identified.

[0074] In the specific implementation, in order to obtain the location of the infected tree in the area to be identified, the server 104 can perform pairwise comparative analysis on multiple target remote sensing images obtained from the previous steps. For example, if the target remote sensing images include four images: A, B, C, and D, then A and B need to be analyzed first, then B and C, and finally C and D. The analysis method for more images is similar.

[0075] Furthermore, before analyzing the two target remote sensing images A and B, it is necessary to obtain the first spatial coordinate information "(w1, y1)" of the first target remote sensing image "A". ′ y1 ′ And obtain the second spatial coordinate information "(x2, y2)" of the second epidemic point location information "(x2, y2)" contained in the target remote sensing image "B". ′ y2 ′ This is used as the spatial coordinate information mentioned above. Then, the distance between the first and second spatial coordinate information is calculated. This distance value "d" can be calculated using the formula for the distance between two points, as shown below:

[0076]

[0077] Furthermore, server 104 can compare the distance value "d" with a preset distance threshold. If the distance value "d" is less than the preset distance threshold, the corresponding infected tree location information "(x1, y1)" and "(x2, y2)" are retained as the target location combination, which serves as the basis for subsequent infected tree location merging analysis. If the distance value "d" is greater than or equal to the preset distance threshold, the corresponding spatial coordinate information "(x1, y1)" and "(x2, y2)" are retained as the target location combination, which serves as the basis for subsequent infected tree location merging analysis. ′ y1 ′ ) and (x2 ′ y2 ′")" refers to the final location of the infected tree. The steps for obtaining spatial coordinate information and obtaining the location of the infected tree involved in this embodiment will be described in detail below.

[0078] In one embodiment, the target remote sensing image is at least two frames of target remote sensing images containing GPS coordinate information; wherein, acquiring and obtaining the first spatial coordinate information corresponding to the first infected tree location information and the second spatial coordinate information corresponding to the second infected tree location information based on the image acquisition time of the target remote sensing image, to obtain spatial coordinate information, includes: acquiring and arranging each target remote sensing image sequentially according to the image acquisition time of the target remote sensing image to obtain a remote sensing image sequence; extracting the target remote sensing images of two adjacent frames in the remote sensing image sequence, respectively as the first image and the second image, and using the infected tree location information contained in the first image as the first infected tree location information and the infected tree location information contained in the second image as the second infected tree location information; and acquiring the first spatial coordinate information corresponding to the first infected tree location information and the second spatial coordinate information corresponding to the second infected tree location information based on the GPS coordinate information, to obtain spatial coordinate information.

[0079] The GPS coordinate information can be coordinate information expressed in latitude and longitude based on the Global Positioning System. In other words, the UAV can not only collect remote sensing images, but also collect the corresponding GPS coordinate information of the images through the Global Positioning System at the same time.

[0080] In practice, the flight altitude of the drone during low-altitude flight is about 1,000 meters. This is because, while maximizing the shooting area, the ground area corresponding to each pixel in the image is kept within 10 centimeters. Thus, the remote sensing image taken by the drone during low-altitude flight contains not only RGB information but also GPS coordinate information. By determining the GPS coordinates corresponding to the origin coordinates, the actual spatial coordinates can be determined from the location information.

[0081] Furthermore, the previous embodiment only proposed that multiple target remote sensing images should be compared and analyzed in pairs. For example, if the target remote sensing images include four images: A, B, C, and D, then A and B should be analyzed first, followed by B and C, and finally C and D. However, the grouping criteria were not clearly defined. Therefore, this embodiment proposes that the images can be grouped according to their acquisition time. For example, the acquisition time of each target remote sensing image can be determined, and then the target remote sensing images can be arranged in ascending order according to their acquisition time to obtain a remote sensing image sequence. Then, the target remote sensing images of two adjacent frames in the sequence can be extracted as a group, and the spatial coordinate information corresponding to the location information of the target points contained in the two consecutive target remote sensing images can be obtained. This yields the first spatial coordinate information and the second spatial coordinate information recorded by different target remote sensing images.

[0082] It should be noted that if a remote sensing image of a target contains multiple target points for infected trees, that is, if there is multiple information on the location of infected trees, then when performing distance analysis, it is necessary to perform permutation and combination analysis on all the information on the location of infected trees in order to select the precise combination of target points.

[0083] In one embodiment, the point image features of each of the target point combination are extracted, and the location of the infected trees is merged based on the point image features to analyze and obtain the location of the infected trees in the area to be identified. This includes: extracting the point image features of each of the infected tree location information contained in the target point combination, and using them as the first point image feature and the second point image feature, respectively; calculating the similarity value between the first point image feature and the second point image feature, so as to select the target point combination with a similarity value greater than a preset similarity threshold for point merging processing to obtain the target point information; and obtaining the spatial coordinate information corresponding to the target point information and the spatial coordinate information corresponding to the unmerged infected tree location information, as the location of the infected trees in the area to be identified.

[0084] In specific implementation, the point selection scheme based on time dimension analysis of distance values ​​described in the above embodiment is actually a target association algorithm based on time dimension. This embodiment is a continuation of this algorithm, which can extract image features from the target point combination, then compare the image feature vectors belonging to different target remote sensing images, that is, perform feature similarity analysis, and finally merge the epidemic tree point information with high similarity to avoid the same epidemic tree location being output repeatedly.

[0085] Furthermore, server 104 can use the feature extraction model "embedding" to extract the point image features of the two images respectively. Specifically, the point image features of a region centered on the target infected tree with a certain side length (e.g., "30m*30m") are transformed into a 512-dimensional feature vector after the model integrates a series of appearance information such as the color, shape, texture, size, and location distribution of the infected tree and surrounding trees. These features are respectively used as the first point image feature "X" belonging to target remote sensing image "A" and the second point image feature "Y" belonging to target remote sensing image "B". Then, the cosine similarity between the two point image features is calculated to obtain the similarity value. Finally, target point combinations with similarity values ​​greater than a preset similarity threshold are selected for point merging processing, thus combining the two into one image, as shown below. Figure 4 The target location information is shown. Therefore, the spatial coordinates corresponding to the target location information, and the spatial coordinates corresponding to the unmerged infected tree location information, constitute the current location of the infected tree to be identified.

[0086] It should be noted that the feature extraction model "embedding" uses "ResNet18" as the backbone network; the inference process can be found in [link to documentation]. Figure 5 The cosine similarity formula is as follows:

[0087]

[0088] It should also be noted that the above-mentioned screening and merging of infected tree location information should be performed on a single flight route area. If the area to be identified is divided into multiple flight route areas, the screening and merging of infected tree location information should be performed on each flight route area separately. Then, the infected tree location information retained between two adjacent flight route areas should be analyzed in terms of spatial dimension. The analysis method is also the distance analysis and feature similarity analysis described in the above embodiment. Finally, the spatial coordinate information of the merged target location information and the unmerged infected tree location information is used as the location of the infected tree in the area to be identified.

[0089] The diseased tree identification method in the above embodiments involves the server acquiring a set of remote sensing images of the area to be identified and inputting this set into a trained diseased tree identification model for identification. This yields a target remote sensing image containing diseased tree location information. Then, the server acquires the spatial coordinate information corresponding to the diseased tree location information. Based on the image acquisition time and spatial coordinate information of the target remote sensing image, the server filters and merges the diseased tree location information to finally obtain the location of the diseased trees in the area to be identified. Therefore, the diseased tree identification scheme proposed in this application eliminates the need for image stitching, saving time and avoiding problems such as image feature damage caused by stitching, effectively improving the accuracy of diseased tree identification.

[0090] It should be understood that, although Figure 2 The steps in the flowchart are shown sequentially as indicated by the arrows, but these steps are not necessarily executed in the order indicated by the arrows. Unless otherwise specified herein, there is no strict order in which these steps are executed, and they can be performed in other orders. Figure 2 At least some of the steps in the process may include multiple sub-steps or multiple stages. These sub-steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these sub-steps or stages is not necessarily sequential, but can be executed in turn or alternately with other steps or at least some of the sub-steps or stages of other steps.

[0091] To better implement the diseased wood identification method provided in the embodiments of this application, based on the diseased wood identification method proposed in the embodiments of this application, this application also provides a diseased wood identification device, such as... Figure 6 As shown, the infected wood identification device 600 includes:

[0092] Image acquisition module 610 is used to acquire a set of remote sensing images collected for the area to be identified;

[0093] The diseased tree identification module 620 is used to input the remote sensing image set into the trained diseased tree identification model to identify diseased trees and obtain a target remote sensing image containing diseased tree location information.

[0094] The location analysis module 630 is used to obtain the spatial coordinate information corresponding to the location information of infected trees, and to filter and merge the location information of infected trees based on the spatial coordinate information and the image acquisition time of the target remote sensing image, so as to obtain the location of infected trees in the area to be identified.

[0095] In one embodiment, the area to be identified includes at least two flight path areas to be identified. The image acquisition module 610 is further configured to acquire a first remote sensing image for the first flight path area and preprocess the first remote sensing image to obtain a first remote sensing image set; and acquire a second remote sensing image for the second flight path area and preprocess the second remote sensing image to obtain a second remote sensing image set; wherein the first flight path area and the second flight path area are adjacent flight path areas whose corresponding image overlap rate meets a preset condition; and determine the first remote sensing image set and the second remote sensing image set as a remote sensing image set.

[0096] In one embodiment, the image acquisition module 610 is further configured to acquire a first remote sensing image collected for the first flight path area; perform frame extraction processing on the first remote sensing image to obtain at least two frames of the first ground remote sensing image; and perform sub-image cropping processing on each of the first ground remote sensing images according to a preset cutting size and overlap size and a preset screenshot direction to obtain a first remote sensing image set; wherein the screenshot direction includes a horizontal screenshot direction and / or a vertical screenshot direction.

[0097] In one embodiment, the diseased tree identification module 620 is further configured to input a remote sensing image set into a trained diseased tree identification model for diseased tree identification, output candidate remote sensing images marked with target diseased tree points, and image coordinate information of the target diseased tree points; determine the origin coordinate information of the candidate remote sensing images, and perform coordinate restoration processing on the image coordinate information based on the origin coordinate information to obtain the diseased tree point location information of the target diseased tree points; and determine the candidate remote sensing image containing the diseased tree point location information as the target remote sensing image.

[0098] In one embodiment, the target remote sensing image is at least two frames of target remote sensing images. The infected tree location information includes first infected tree location information and second infected tree location information belonging to different target remote sensing images. The location analysis module 630 is further used to acquire and, based on the image acquisition time of the target remote sensing image, acquire the first spatial coordinate information corresponding to the first infected tree location information and the second spatial coordinate information corresponding to the second infected tree location information to obtain spatial coordinate information; calculate the distance value between the first spatial coordinate information and the second spatial coordinate information to filter out infected tree location information with a distance value less than a preset distance threshold as target location combination; extract the location image features of each target location combination, and, based on the location image features, merge the infected tree location information to analyze and obtain the location of infected trees in the area to be identified.

[0099] In one embodiment, the target remote sensing image is at least two frames of target remote sensing images containing GPS coordinate information. The position analysis module 630 is further configured to acquire and arrange the target remote sensing images sequentially according to the image acquisition time of the target remote sensing images to obtain a remote sensing image sequence; extract the target remote sensing images of two adjacent frames in the remote sensing image sequence as the first image and the second image, respectively, and use the epidemic tree location information contained in the first image as the first epidemic tree location information, and the epidemic tree location information contained in the second image as the second epidemic tree location information; based on the GPS coordinate information, acquire the first spatial coordinate information corresponding to the first epidemic tree location information and the second spatial coordinate information corresponding to the second epidemic tree location information to obtain spatial coordinate information.

[0100] In one embodiment, the location analysis module 630 is further configured to extract the location image features of each of the infected tree location information contained in the target location combination, and use them as the first location image feature and the second location image feature, respectively; calculate the similarity value between the first location image feature and the second location image feature, so as to filter out the target location combination with the similarity value greater than the preset similarity threshold for location merging processing to obtain the target location information; and obtain the spatial coordinate information corresponding to the target location information and the spatial coordinate information corresponding to the unmerged infected tree location information, as the location of the infected tree in the area to be identified.

[0101] In the above embodiments, the server acquires a set of remote sensing images of the area to be identified and inputs the set of images into a trained diseased tree identification model for diseased tree identification. This yields a target remote sensing image containing diseased tree location information. Then, the server acquires the spatial coordinate information corresponding to the diseased tree location information. Based on the image acquisition time and spatial coordinate information of the target remote sensing image, the server filters and merges the diseased tree location information to finally obtain the location of the diseased trees in the area to be identified. Therefore, the diseased tree identification scheme proposed in this application eliminates the need for image stitching, saving time and avoiding problems such as image feature damage caused by stitching, effectively improving the accuracy of diseased tree identification.

[0102] It should be noted that the specific limitations regarding the diseased wood identification device can be found in the limitations of the diseased wood identification method described above, and will not be repeated here. Each module in the aforementioned diseased wood identification device can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in the processor of the electronic device in hardware form or independently of it, or stored in the memory of the electronic device in software form, so that the processor can call and execute the corresponding operations of each module.

[0103] In some embodiments of this application, the epidemic identification device 600 can be implemented as a computer program, which can be implemented in, for example... Figure 7 The device runs on the computer shown. The computer's memory can store the various program modules that make up the virus identification device 600, for example, Figure 6 The image acquisition module 610, the infected tree identification module 620, and the location analysis module 630 shown; the computer program composed of each program module causes the processor to execute the steps in the infected tree identification methods of the various embodiments of this application described in this specification. For example, Figure 7 The computer equipment shown can be used as follows Figure 6 The image acquisition module 610 of the diseased tree identification device 600 shown executes step S201. The computer device can execute step S202 via the diseased tree identification module 620. The computer device can execute step S203 via the location analysis module 630. The computer device includes a processor, memory, and a network interface connected via a system bus. The processor provides computing and control capabilities. The memory includes a non-volatile storage medium and internal memory. The non-volatile storage medium stores the operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage medium. The network interface of the computer device is used for communication with external computer devices via a network connection. When the computer program is executed by the processor, it implements a diseased tree identification method.

[0104] As will be understood by those skilled in the art, Figure 7 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0105] In some embodiments of this application, a computer device is provided, including one or more processors; a memory; and one or more application programs, wherein the one or more application programs are stored in the memory and configured to be executed by the processors as steps of the above-described diseased wood identification method. The steps of this diseased wood identification method may be steps from the diseased wood identification methods of the various embodiments described above.

[0106] In some embodiments of this application, a computer-readable storage medium is provided, storing a computer program. The computer program is loaded by a processor, causing the processor to execute the steps of the above-described diseased wood identification method. The steps of the diseased wood identification method here can be the steps in the diseased wood identification methods of the various embodiments described above.

[0107] Those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. Any references to memory, storage, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, or optical storage, etc. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM), etc.

[0108] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0109] The above provides a detailed description of a method, apparatus, computer device, and computer-readable storage medium for identifying infected trees according to embodiments of this application. Specific examples have been used to illustrate the principles and implementation methods of the present invention. The descriptions of the above embodiments are only for the purpose of helping to understand the method and core ideas of the present invention. At the same time, those skilled in the art will recognize that there will be changes in the specific implementation methods and application scope based on the ideas of the present invention. Therefore, the content of this specification should not be construed as a limitation of the present invention.

Claims

1. A method for identifying a vaccine wood, characterized by, The method comprises the following steps: acquiring a remote sensing image set collected for a to-be-identified region, the to-be-identified region comprising at least two to-be-identified flight line regions, acquiring a first remote sensing image collected for a first flight line region, and pre-processing the first remote sensing image to obtain a first remote sensing image set, the pre-processing method comprising the following steps: frame extraction processing is performed on the first remote sensing image to obtain at least two first ground remote sensing images; based on a preset cutting size and an overlap size, sub-image cropping processing is performed on each of the first ground remote sensing images in a preset cropping direction to obtain the first remote sensing image set; wherein the cropping direction comprises a horizontal cropping direction and / or a vertical cropping direction; the remote sensing image set is input into a trained tree disease identification model for tree disease identification to obtain a target remote sensing image comprising tree disease point position information; spatial coordinate information corresponding to the tree disease point position information is acquired, and the tree disease point position information is filtered and combined according to the spatial coordinate information and the image acquisition time of the target remote sensing image to obtain the tree disease position in the to-be-identified region; the acquiring of the remote sensing image set collected for the to-be-identified region further comprises the following steps: a second remote sensing image collected for a second flight line region is acquired, and the second remote sensing image is pre-processed to obtain a second remote sensing image set; wherein the first flight line region and the second flight line region are adjacent flight line regions corresponding to an image overlap rate satisfying a preset condition; the first remote sensing image set and the second remote sensing image set are determined as the remote sensing image set; the target remote sensing image is at least two target remote sensing images, and the tree disease point position information comprises first tree disease point position information and second tree disease point position information belonging to different target remote sensing images; wherein the acquiring of the spatial coordinate information corresponding to the tree disease point position information, and the filtering and combining of the tree disease point position information according to the spatial coordinate information and the image acquisition time of the target remote sensing image to obtain the tree disease position in the to-be-identified region comprises the following steps: the first spatial coordinate information corresponding to the first tree disease point position information and the second spatial coordinate information corresponding to the second tree disease point position information are acquired according to the image acquisition time of the target remote sensing image to obtain the spatial coordinate information; a distance value between the first spatial coordinate information and the second spatial coordinate information is calculated to filter out tree disease point position information with a distance value less than a preset distance threshold as a target point combination; point image features of the target point combination are extracted, and the tree disease point position information is combined based on the point image features to analyze and obtain the tree disease position in the to-be-identified region.

2. The method of claim 1, wherein, the inputting of the remote sensing image set into the trained tree disease identification model for tree disease identification to obtain a target remote sensing image comprising tree disease point position information comprises the following steps: the remote sensing image set is input into the trained tree disease identification model for tree disease identification to output a candidate remote sensing image marked with a target tree disease point and image coordinate information of the target tree disease point; Determine the origin coordinate information of the candidate remote sensing image, so as to perform coordinate restoration processing on the image coordinate information based on the origin coordinate information, and obtain the epidemic tree point position information of the target epidemic tree point; Determine the candidate remote sensing image containing the epidemic tree point position information as the target remote sensing image.

3. The method of claim 1, wherein, The target remote sensing image is at least two target remote sensing images containing GPS coordinate information; The obtaining and obtaining the first spatial coordinate information corresponding to the first epidemic tree point position information and the second spatial coordinate information corresponding to the second epidemic tree point position information based on the image acquisition time of the target remote sensing image, to obtain the spatial coordinate information, includes: Obtain and arrange the target remote sensing image in sequence based on the image acquisition time of the target remote sensing image, to obtain a remote sensing image sequence; Extract the target remote sensing image of the adjacent two frames in the remote sensing image sequence as the first image and the second image respectively, and take the epidemic tree point position information contained in the first image as the first epidemic tree point position information, and take the epidemic tree point position information contained in the second image as the second epidemic tree point position information; Based on the GPS coordinate information, obtain the first spatial coordinate information corresponding to the first epidemic tree point position information and the second spatial coordinate information corresponding to the second epidemic tree point position information, to obtain the spatial coordinate information.

4. The method of claim 1, wherein, The extraction of the target point position combination of the point position image feature, and the merging processing of the epidemic tree point position information based on the point position image feature, to analyze the epidemic tree position in the to-be-identified region, includes: Extract the point position image feature of the epidemic tree point position information contained in the target point position combination as the first point position image feature and the second point position image feature respectively; Calculate the similarity value between the first point position image feature and the second point position image feature, to filter out the target point position combination with the similarity value greater than the preset similarity threshold for point position merging processing, to obtain the target point position information; Obtain the spatial coordinate information corresponding to the target point position information and the spatial coordinate information corresponding to the unmerged epidemic tree point position information as the epidemic tree position in the to-be-identified region.

5. A device for identifying a vaccine wood, characterized by, It includes: An image acquisition module is used to acquire a remote sensing image set collected for a to-be-identified region, and the to-be-identified region includes at least two to-be-identified flight line regions; The image acquisition module is also used to acquire a first remote sensing image collected for a first flight line region, and to pre-process the first remote sensing image to obtain a first remote sensing image set; and to acquire a second remote sensing image collected for a second flight line region, and to pre-process the second remote sensing image to obtain a second remote sensing image set; wherein the first flight line region and the second flight line region are adjacent flight line regions with a corresponding image overlap rate meeting a preset condition; the first remote sensing image set and the second remote sensing image set are determined as the remote sensing image set; The image acquisition module is further configured to acquire a first remote sensing image collected for a first air route region; perform frame extraction processing on the first remote sensing image to obtain at least two first ground remote sensing images; perform sub-image cropping processing on each of the first ground remote sensing images according to a preset cropping direction based on a preset cutting size and an overlap size, to obtain the first remote sensing image set; and the cropping direction includes a horizontal cropping direction and / or a vertical cropping direction. The pest wood identification module is configured to input the remote sensing image set into a trained pest wood identification model to perform pest wood identification, to obtain a target remote sensing image containing pest wood location information, the target remote sensing image being at least two target remote sensing images, and the pest wood location information including first pest wood location information and second pest wood location information belonging to different target remote sensing images. The position analysis module is configured to acquire spatial coordinate information corresponding to the pest wood location information, to perform screening and merging processing on the pest wood location information according to the spatial coordinate information and image acquisition time of the target remote sensing image, and to obtain pest wood positions in the region to be identified. The position analysis module is further configured to acquire first spatial coordinate information corresponding to the first pest wood location information and second spatial coordinate information corresponding to the second pest wood location information according to the image acquisition time of the target remote sensing image, to obtain the spatial coordinate information; calculate a distance value between the first spatial coordinate information and the second spatial coordinate information, to screen out pest wood location information with a distance value less than a preset distance threshold as a target location combination; extract location image features of the target location combination, and perform merging processing on the pest wood location information based on the location image features, to analyze and obtain pest wood positions in the region to be identified.

6. A computer device, comprising: comprise: one or more processors; a memory; and one or more application programs, wherein the one or more application programs are stored in the memory and configured to be executed by the processor to implement the pest wood identification method of any one of claims 1 to 4.

7. A computer readable storage medium characterized by a computer program is stored thereon, and the computer program is loaded by a processor to perform the steps in the pest wood identification method of any one of claims 1 to 4.

Images