Wheat heading period identification method, device, electronic device and storage medium

By acquiring and integrating high-level semantic information and wheat ear image features in the wheat planting area, the problem of low accuracy in the ear recognition in the prior art is solved, and high accuracy recognition is achieved in complex environments.

CN113139552BActive Publication Date: 2024-11-12AEROSPACE NEWSKY TECHNOLOGY CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202110476746.X
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2021-04-29
Publication Date
2024-11-12
Estimated Expiration
2041-04-29

AI Technical Summary

Technical Problem

The accuracy of wheat earing period in the prior art is low, and is greatly affected by the background and complex lighting environment factors. It is difficult to accurately determine whether wheat enters the earing period in complex environments.

Method used

By obtaining the monitoring images and high-level semantic information of the wheat planting area, the wheat ear image characteristics are extracted, and the high-level semantic information and wheat ear image characteristics are fused, the target features are generated to determine the wheat ear recognition results, and pre-processing and decorrelation transformation are used to , threshold segmentation and feature fusion technologies to improve identification accuracy.

Benefits of technology

It improves the accuracy and robustness of wheat heading period identification, enhances the reliability of identification results, and can effectively judge whether wheat enters heading period in complex environments.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN113139552B_ABST
    Figure CN113139552B_ABST
Patent Text Reader

Abstract

The present application provides a method, device, electronic device and storage medium for wheat heading period recognition, the method comprising: obtaining monitoring images of wheat planting areas and high-level semantic information of the monitoring images; extracting wheat ear image features in the monitoring images; fusing the high-level semantic information of the monitoring images and the wheat ear image features to obtain target features; and determining the wheat heading period recognition result according to the target features. The method provided by the above scheme improves the accuracy of the recognition result by determining the recognition result according to the target features that fuse the high-level semantic information of the monitoring images and the wheat ear image features.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present application relates to the field of image processing technology, and in particular to a method, device, electronic device and storage medium for identifying the heading period of wheat. Background Art

[0002] The production rate of wheat is affected by factors such as regional climate conditions. In order to ensure the healthy growth of wheat, growers need to monitor its growth period and take corresponding cultivation measures at the appropriate time. At present, in order to improve the monitoring efficiency, how to automatically monitor the growth period of wheat has become a key research topic.

[0003] In the existing technology, usually only one algorithm such as target detection and image classification is used to determine whether wheat has entered the heading period. Therefore, under the influence of complex environmental factors such as background and lighting, the recognition accuracy is low. Summary of the invention

[0004] The present application provides a wheat heading period identification method, device, electronic device and storage medium to solve the defects of the existing automatic identification technology such as low accuracy.

[0005] The first aspect of the present application provides a method for identifying the heading period of wheat, comprising:

[0006] Acquire monitoring images of a wheat planting area and high-level semantic information of the monitoring images;

[0007] Extracting wheat ear image features in the monitoring image;

[0008] fusing the high-level semantic information of the monitoring image and the wheat ear image features to obtain target features;

[0009] The wheat heading period recognition result is determined according to the target characteristics.

[0010] Optionally, before extracting the wheat ear image features in the monitoring image, the method further includes:

[0011] Preprocessing the monitoring image to obtain a preprocessed monitoring image;

[0012] Performing decorrelation transformation processing on the preprocessed monitoring image to obtain a decorrelation transformed monitoring image;

[0013] Threshold segmentation is performed on the surveillance image after the decorrelation transformation to obtain a surveillance image without background.

[0014] Optionally, before fusing the high-level semantic information of the monitoring image with the wheat ear image features, the method further includes:

[0015] The high-level semantic information and wheat ear image features of the monitoring image are input into a preset buffer layer, and the buffer layer is used to buffer the high-level semantic information and wheat ear image features to weaken the difference between the high-level semantic information and the wheat ear image features.

[0016] Optionally, the fusing of the high-level semantic information of the monitoring image and the wheat ear image features includes:

[0017] Performing vector conversion on the high-level semantic information and wheat ear image features of the monitoring image respectively to obtain corresponding high-level semantic information vectors and wheat ear image feature vectors;

[0018] The high-level semantic information vector and the wheat ear image feature vector are fused.

[0019] Optionally, extracting wheat ear image features in the surveillance image includes:

[0020] According to the growth direction of wheat, the monitoring image is divided into a plurality of feature domains;

[0021] Extract image features of each feature domain;

[0022] The wheat ear image features in the monitoring image are determined according to the image features of each feature domain.

[0023] Optionally, obtaining the monitoring image of the wheat planting area includes:

[0024] Obtaining original monitoring images of wheat-growing areas;

[0025] The original monitoring image is cropped according to a preset cropping size to obtain a plurality of monitoring images.

[0026] Optionally, determining the wheat heading period recognition result according to the target feature includes:

[0027] Determine the heading period recognition result of each monitoring image according to the target features corresponding to each monitoring image;

[0028] According to the heading period recognition result of each monitoring image, it is determined whether the wheat in the wheat planting area is in the heading period.

[0029] A second aspect of the present application provides a wheat heading period identification device, comprising:

[0030] An acquisition module, used to acquire monitoring images of the wheat planting area and high-level semantic information of the monitoring images;

[0031] A feature extraction module, used to extract wheat ear image features in the monitoring image;

[0032] A feature fusion module, used for fusing the high-level semantic information of the monitoring image and the wheat ear image features to obtain target features;

[0033] The recognition module is used to determine the wheat heading period recognition result according to the target characteristics.

[0034] Optionally, the acquisition module is further used to:

[0035] Preprocessing the monitoring image to obtain a preprocessed monitoring image;

[0036] Performing decorrelation transformation processing on the preprocessed monitoring image to obtain a decorrelation transformed monitoring image;

[0037] Threshold segmentation is performed on the surveillance image after the decorrelation transformation to obtain a surveillance image without background.

[0038] Optionally, the feature fusion is also used for:

[0039] The high-level semantic information and wheat ear image features of the monitoring image are input into a preset buffer layer, and the buffer layer is used to buffer the high-level semantic information and wheat ear image features to weaken the difference between the high-level semantic information and the wheat ear image features.

[0040] Optionally, the feature fusion module is specifically used to:

[0041] Performing vector conversion on the high-level semantic information and wheat ear image features of the monitoring image respectively to obtain corresponding high-level semantic information vectors and wheat ear image feature vectors;

[0042] The high-level semantic information vector and the wheat ear image feature vector are fused.

[0043] Optionally, the feature extraction module is specifically used to:

[0044] According to the growth direction of wheat, the monitoring image is divided into a plurality of feature domains;

[0045] Extract image features of each feature domain;

[0046] The wheat ear image features in the monitoring image are determined according to the image features of each feature domain.

[0047] Optionally, the acquisition module is specifically used to:

[0048] Obtaining original monitoring images of wheat-growing areas;

[0049] The original monitoring image is cropped according to a preset cropping size to obtain a plurality of monitoring images.

[0050] Optionally, the identification module is specifically used to:

[0051] Determine the heading period recognition result of each monitoring image according to the target features corresponding to each monitoring image;

[0052] According to the heading period recognition result of each monitoring image, it is determined whether the wheat in the wheat planting area is in the heading period.

[0053] A third aspect of the present application provides an electronic device, comprising: at least one processor and a memory;

[0054] The memory stores computer-executable instructions;

[0055] The at least one processor executes the computer-executable instructions stored in the memory, so that the at least one processor performs the method described in the first aspect and various possible designs of the first aspect.

[0056] A fourth aspect of the present application provides a computer-readable storage medium, wherein the computer-readable storage medium stores computer-executable instructions. When a processor executes the computer-executable instructions, the method described in the first aspect and various possible designs of the first aspect are implemented.

[0057] The technical solution of this application has the following advantages:

[0058] The wheat heading period recognition method, device, electronic device and storage medium provided by the present application obtain monitoring images of wheat planting areas and high-level semantic information of monitoring images; extract wheat ear image features in monitoring images; fuse high-level semantic information of monitoring images and wheat ear image features to obtain target features; and determine the wheat heading period recognition result based on the target features. The method provided by the above scheme improves the accuracy of the recognition result by determining the recognition result based on the target features that fuse the high-level semantic information of the monitoring image and the wheat ear image features. BRIEF DESCRIPTION OF THE DRAWINGS

[0059] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present application, and a person skilled in the art can also obtain other drawings based on these drawings.

[0060] Figure 1 This is a schematic diagram of the structure of the wheat heading period identification system based on the embodiment of the present application;

[0061] Figure 2 A schematic diagram of a process for identifying the heading period of wheat provided in an embodiment of the present application;

[0062] Figure 3 A schematic diagram of a flow chart of another wheat heading period identification method provided in an embodiment of the present application;

[0063] Figure 4 A schematic diagram of a wheat heading period recognition model provided in an embodiment of the present application;

[0064] Figure 5 A schematic diagram of the overall process of an exemplary wheat heading period identification method provided in an embodiment of the present application;

[0065] Figure 6 A schematic diagram of the structure of a wheat heading period identification device provided in an embodiment of the present application;

[0066] Figure 7 A schematic diagram of the structure of an electronic device provided in an embodiment of the present application.

[0067] The above drawings have shown clear embodiments of the present application, which will be described in more detail below. These drawings and text descriptions are not intended to limit the scope of the present disclosure in any way, but to illustrate the concepts of the present application to those skilled in the art by referring to specific embodiments. DETAILED DESCRIPTION

[0068] In order to make the purpose, technical solution and advantages of the embodiments of the present application clearer, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0069] In addition, the terms "first", "second", etc. are used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of technical features indicated. In the description of the following embodiments, the meaning of "multiple" is more than two, unless otherwise clearly and specifically defined.

[0070] In the prior art, a single algorithm such as target detection and image classification is usually used to determine whether wheat has entered the heading period, and no special image preprocessing module and feature extraction module are designed for this problem; therefore, under the influence of complex environmental factors such as background and lighting, the recognition accuracy is low.

[0071] In response to the above problems, the wheat heading period recognition method, device, electronic device and storage medium provided in the embodiments of the present application obtain monitoring images of wheat planting areas and high-level semantic information of the monitoring images; extract wheat ear image features in the monitoring images; fuse the high-level semantic information of the monitoring images and the wheat ear image features to obtain target features; and determine the wheat heading period recognition result based on the target features. The method provided by the above scheme improves the accuracy of the recognition result by determining the recognition result based on the target features that fuse the high-level semantic information of the monitoring images and the wheat ear image features.

[0072] The following specific embodiments may be combined with each other, and the same or similar concepts or processes may not be described in detail in some embodiments. The embodiments of the present invention will be described below in conjunction with the accompanying drawings.

[0073] First, the structure of the wheat heading period recognition system on which this application is based is described:

[0074] The wheat heading period identification method, device, electronic device and storage medium provided in the embodiments of the present application are suitable for automatically identifying whether the wheat has entered the heading period during its growth process. Figure 1 The figure is a schematic diagram of the structure of the wheat heading period recognition system based on the embodiment of the present application, which mainly includes a camera and a wheat heading period recognition device for identifying the wheat heading period. Specifically, the camera is used to collect monitoring images of the wheat planting area, and the collected monitoring images are sent to the wheat heading period recognition device for it to identify whether the wheat has entered the heading period.

[0075] The embodiment of the present application provides a wheat heading period identification method for identifying whether wheat has entered the heading period. The execution subject of the embodiment of the present application is an electronic device, such as a server, a desktop computer, a laptop computer, a tablet computer, and other electronic devices that can be used to identify the wheat heading period.

[0076] like Figure 2 FIG. 1 is a flow chart of a wheat heading period identification method provided in an embodiment of the present application, the method comprising:

[0077] Step 201, obtaining a monitoring image of a wheat planting area and high-level semantic information of the monitoring image.

[0078] It should be noted that the high-level semantic information can be obtained based on the AlexNet algorithm, also known as neural network features (CNN features). The specific extraction process can refer to the prior art and is not limited in the embodiments of the present application.

[0079] Step 202: extract wheat ear image features in the monitoring image.

[0080] The wheat ear image features may include at least texture features and color features, etc., and a preset wheat heading period recognition model may be used to extract the wheat ear image features.

[0081] Specifically, the interference of background factors such as wheat flag leaves is removed by giving an HSV spatial threshold, so as to obtain a wheat ear image without background interference and extract the features of the wheat ear image.

[0082] Step 203: fuse the high-level semantic information of the monitoring image with the wheat ear image features to obtain the target features.

[0083] It should be noted that the extraction of high-level semantic information is currently affected by factors such as the similarity between the wheat ear features and the background color, and the wheat ear features are easily blocked, especially the image features in the early stage of heading are not significant enough, resulting in a large error in the single heading period recognition method based on high-level semantic information. The embodiment of the present application proposes to obtain the target features by feature fusion, and then automatically recognize the heading period, which solves the above problems.

[0084] Specifically, by fusing high-level semantic information and wheat ear image features, the obtained target features combine the advantages of both, have stronger robustness, and are conducive to improving the accuracy of recognition results.

[0085] Specifically, feature fusion can be performed by serial feature fusion or parallel feature fusion. The appropriate fusion method can be selected according to the actual situation, and the embodiment of the present application does not limit it.

[0086] Specifically, in one embodiment, the high-level semantic information and wheat ear image features of the monitoring image can be vectorized separately to obtain corresponding high-level semantic information vectors and wheat ear image feature vectors; and the high-level semantic information vectors and wheat ear image feature vectors can be fused.

[0087] Specifically, before feature fusion, high-level semantic information (in the form of feature graph) is expanded into a one-dimensional feature vector α, the wheat ear image feature vector is vectorized into β, and the target feature (fusion vector) H = iα + jβ.

[0088] Among them, i and j respectively represent the fusion weights corresponding to the two features, which can be set according to actual conditions and are not limited in the embodiments of the present application.

[0089] Step 204, determining the wheat heading period recognition result according to the target features.

[0090] It should be noted that before wheat heading period recognition, it is necessary to construct a corresponding wheat heading period recognition model, and use a large number of samples (image samples) to train the model so that the model gradually determines the feature range (distinguishing features) corresponding to the wheat heading period, so that the trained model can accurately determine whether the wheat is currently in the heading period or the booting period. Among them, the training principle of the model corresponds to the principle in the wheat heading period recognition method provided in the above embodiment, and will not be repeated here.

[0091] Specifically, the trained wheat heading period recognition model determines the growth period corresponding to the wheat in the monitoring image according to the target feature and the determined heading period feature range when the target feature is obtained. For example, if it is determined to be in the heading period, 1 is output; if it is not in the heading period, that is, in the booting period, 0 is output.

[0092] On the basis of the above-mentioned embodiment, since the wheat planting area usually covers a large outdoor area, the camera is affected by the outdoor environment, and it is difficult to ensure the image quality of the monitoring image, which is not conducive to ensuring the accuracy of the recognition result.

[0093] In response to the above problems, Figure 3 A flowchart of another wheat heading period recognition method provided in an embodiment of the present application is provided. As an implementable manner, in one embodiment, before extracting wheat ear image features in a monitoring image, the method further includes:

[0094] Step 301, preprocessing the monitoring image to obtain a preprocessed monitoring image;

[0095] Step 302, performing decorrelation transformation processing on the preprocessed monitoring image to obtain a surveillance image after decorrelation transformation;

[0096] Step 303: Perform threshold segmentation processing on the surveillance image after decorrelation transformation to obtain a surveillance image without background.

[0097] Specifically, an illumination adaptive algorithm (Retinex algorithm) may be used to pre-process the surveillance image to enhance the adaptability of the surveillance image to illumination changes and reduce the impact of different illumination intensities at different times on surveillance image recognition.

[0098] It should be noted that the color of an object is determined by its ability to reflect long-wave (red), medium-wave (green), and short-wave (blue) light, rather than by the absolute value of the intensity of the reflected light. The color of an object is not affected by illumination non-uniformity and is consistent. The Retinex algorithm is designed based on color consistency. The image processed by this algorithm is equivalent to color normalization, which reduces the impact of color changes under different illumination intensities on color-based feature extraction. Moreover, unlike traditional linear and nonlinear methods that can only enhance a certain type of image features, the Retinex algorithm can achieve a balance in dynamic range compression, edge enhancement, and color constancy, and achieve adaptive enhancement of various types of images.

[0099] Furthermore, after completing the image preprocessing operation, the image enhanced by the illumination adaptive algorithm is decorrelated and stretched to amplify the slight color difference between the wheat ear and the flag leaf, laying a good foundation for texture feature extraction based on color feature threshold.

[0100] It should be noted that due to the high similarity between the various bands in the surveillance image, the hue change of the surveillance image is limited and the saturation is low; if the grayscale value range is directly stretched for each band, the value range of the image brightness is increased, while the image color saturation is still low. Decorrelation stretching can effectively amplify the part of the information with low correlation, improve the saturation of the surveillance image, and retain the color characteristics.

[0101] It should be further explained that, for the above-mentioned threshold segmentation, the threshold segmentation can be specifically performed according to the color threshold. The determined color threshold ensures the integrity of the wheat ear as much as possible, removes background information such as wheat flag leaves, that is, extracts the foreground information of the monitoring image, and the specific value of the color threshold is not limited in this embodiment of the application.

[0102] On the basis of the above embodiment, since the high-level semantic information and wheat ear image features are extracted based on two different methods, there are certain differences between the two. The wheat heading period recognition model provided by the above embodiment can be constructed based on a neural network algorithm, and the feature fusion work needs to be implemented before the fully connected layer of the model. If the difference between the two is large, it will directly affect the feature fusion effect.

[0103] In response to the above problems, Figure 4 As shown, it is an architecture diagram of the wheat heading period recognition model provided in an embodiment of the present application. As an implementable method, in one embodiment, before fusing the high-level semantic information of the monitoring image and the wheat ear image features, the high-level semantic information and the wheat ear image features of the monitoring image can be input into a preset buffer layer, and the buffer layer is used to buffer the high-level semantic information and the wheat ear image features to weaken the difference between the high-level semantic information and the wheat ear image features.

[0104] The working principle of the buffer layer may refer to the prior art and is not limited in the embodiments of the present application.

[0105] Specifically, high-level semantic information and wheat ear image features first enter the buffer layer and then enter the fusion layer after buffering.

[0106] Furthermore, in order to improve the effect of feature fusion, a deactivation function with a probability of 0.5 can be used between the buffer layer and the fusion layer.

[0107] For example, Figure 5 The figure is a schematic diagram of the overall process of an exemplary wheat heading period identification method provided in the embodiment of the present application. Figure 5 The method shown is Figure 2 An exemplary implementation of the method shown is shown, and the principles of the two are the same and will not be described in detail.

[0108] Specifically, in one embodiment, in order to improve the reliability of the extracted image features, the monitoring image can be divided into multiple feature domains according to the growth direction of wheat; the image features of each feature domain are extracted; and the image features of wheat ears in the monitoring image are determined based on the image features of each feature domain.

[0109] For example, if the longitudinal direction of the monitoring image corresponds to the wheat growth direction, the monitoring image can be divided into three strips (feature domains) in the longitudinal direction. Texture features are extracted from the three strips respectively, and then a feature representation method based on the local maximum incidence rate is adopted to ensure the sensitivity of feature information while improving its robustness, and finally obtain wheat ear image features with illumination self-adaptation ability and wheat ear feature sensitivity.

[0110] On the basis of the above embodiment, since the camera usually collects surveillance images of wheat-growing areas, it usually collects a large-scale planting area, that is, a surveillance image contains a large area of ​​wheat-growing areas and a large number of wheat ears. After image compression, the image will be distorted, which is not conducive to ensuring the reliability of subsequent recognition results.

[0111] Therefore, as an implementable method, in one embodiment, the original monitoring image of the wheat planting area can be obtained; the original monitoring image is cropped according to a preset cropping size to obtain a plurality of monitoring images.

[0112] For example, taking the size of the original surveillance image as 2560*1920, the cropping size can be set to 512*512 to divide the original surveillance image into multiple square surveillance images. Even if such a surveillance image is compressed (such as compressed to 256*256), it will not affect the shape of patterns such as wheat ears, laying the foundation for improving recognition accuracy.

[0113] Furthermore, in one embodiment, based on the target features corresponding to each monitoring image, the heading period recognition result of each monitoring image is determined; based on the heading period recognition result of each monitoring image, it is determined whether the wheat in the wheat planting area is in the heading period.

[0114] Specifically, the overall recognition result may be determined according to the proportion of the monitoring images at the heading stage and the monitoring images at the booting stage.

[0115] For example, if the original monitoring image is divided into 10 monitoring images, among which, the heading period recognition results of 7 monitoring images are true (indicating that the wheat has entered the heading period), and the heading period recognition results of 3 monitoring images are false (indicating that the wheat is in the heading stage), then it is determined that the wheat in the current wheat planting area has entered the heading period.

[0116] The wheat heading period recognition method provided in the embodiment of the present application obtains the monitoring image of the wheat planting area and the high-level semantic information of the monitoring image; extracts the wheat ear image features in the monitoring image; fuses the high-level semantic information of the monitoring image and the wheat ear image features to obtain the target features; and determines the wheat heading period recognition result according to the target features. The method provided by the above scheme improves the accuracy of the recognition result by determining the recognition result according to the target features that fuse the high-level semantic information of the monitoring image and the wheat ear image features. In addition, the high-level semantic information and the wheat ear image features are fused in the image recognition process, the feature representation of the wheat ear is strengthened, and it has stronger robustness and improves the recognition accuracy. By further adding the wheat ear image features on the basis of the original high-level semantic information, it is equivalent to adding an attention mechanism to the model, so that the machine learning model can more effectively capture the features that are conducive to distinguishing the booting period and the heading period during the classification process, so that the model converges faster and the recognition accuracy is higher.

[0117] An embodiment of the present application provides a wheat heading period identification device, which is used to execute the wheat heading period identification method provided in the above embodiment.

[0118] like Figure 6 , which is a schematic diagram of the structure of a wheat heading period identification device provided in an embodiment of the present application. The wheat heading period identification device 60 comprises an acquisition module 601 , a feature extraction module 602 , a feature fusion module 603 and a recognition module 604 .

[0119] Among them, the acquisition module is used to obtain the monitoring images of the wheat planting area and the high-level semantic information of the monitoring images; the feature extraction module is used to extract the wheat ear image features in the monitoring images; the feature fusion module is used to fuse the high-level semantic information of the monitoring images and the wheat ear image features to obtain the target features; the recognition module is used to determine the wheat heading period recognition results based on the target features.

[0120] Specifically, in one embodiment, the acquisition module is further used to:

[0121] Preprocessing the monitoring image to obtain a preprocessed monitoring image;

[0122] Performing decorrelation transformation processing on the preprocessed monitoring image to obtain a decorrelation transformed monitoring image;

[0123] Threshold segmentation is performed on the surveillance image after decorrelation transformation to obtain a surveillance image without background.

[0124] Specifically, in one embodiment, feature fusion is also used for:

[0125] The high-level semantic information and wheat ear image features of the monitoring image are input into a preset buffer layer, and the buffer layer is used to buffer the high-level semantic information and wheat ear image features to weaken the difference between the high-level semantic information and the wheat ear image features.

[0126] Specifically, in one embodiment, the feature fusion module is specifically used to:

[0127] Performing vector conversion on the high-level semantic information and wheat ear image features of the monitoring image respectively to obtain the corresponding high-level semantic information vector and wheat ear image feature vector;

[0128] Fusion of high-level semantic information vector and wheat ear image feature vector.

[0129] Specifically, in one embodiment, the feature extraction module is specifically used to:

[0130] According to the wheat growth direction, the monitoring image is divided into multiple feature domains;

[0131] Extract image features of each feature domain;

[0132] According to the image features of each feature domain, the wheat ear image features in the monitoring image are determined.

[0133] Specifically, in one embodiment, the acquisition module is specifically used to:

[0134] Obtaining original monitoring images of wheat-growing areas;

[0135] The original monitoring image is cropped according to a preset cropping size to obtain multiple monitoring images.

[0136] Specifically, in one embodiment, the identification module is specifically used to:

[0137] Determine the heading period recognition result of each monitoring image according to the target features corresponding to each monitoring image;

[0138] According to the heading period recognition results of each monitoring image, it is determined whether the wheat in the wheat planting area is in the heading period.

[0139] Regarding the wheat heading period identification device in this embodiment, the specific manner in which each module performs operations has been described in detail in the embodiment of the method, and will not be elaborated here.

[0140] The wheat heading period identification device provided in the embodiment of the present application is used to execute the wheat heading period identification method provided in the above embodiment. Its implementation method and principle are the same and will not be repeated here.

[0141] An embodiment of the present application provides an electronic device for executing the wheat heading period identification method provided in the above embodiment.

[0142] like Figure 7 , which is a schematic diagram of the structure of an electronic device provided in an embodiment of the present application. The electronic device 70 includes: at least one processor 71 and a memory 72;

[0143] The memory stores computer-executable instructions; at least one processor executes the computer-executable instructions stored in the memory, so that the at least one processor executes the wheat heading period identification method provided in the above embodiment.

[0144] An electronic device provided in an embodiment of the present application is used to execute the wheat heading period identification method provided in the above embodiment. Its implementation method and principle are the same and will not be repeated here.

[0145] An embodiment of the present application provides a computer-readable storage medium, in which computer-executable instructions are stored. When a processor executes the computer-executable instructions, the wheat heading period identification method provided in any of the above embodiments is implemented.

[0146] The storage medium containing computer executable instructions in the embodiment of the present application can be used to store computer executable instructions of the wheat heading period identification method provided in the aforementioned embodiment. The implementation method and principle are the same and will not be repeated here.

[0147] Those skilled in the art can clearly understand that for the convenience and simplicity of description, only the division of the above-mentioned functional modules is used as an example for illustration. In actual applications, the above-mentioned functions can be assigned to different functional modules as needed, that is, the internal structure of the device is divided into different functional modules to complete all or part of the functions described above. The specific working process of the device described above can refer to the corresponding process in the aforementioned method embodiment, and will not be repeated here.

[0148] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit it. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or replace some or all of the technical features therein with equivalents. However, these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present application.

Claims

1. A method for identifying the heading period of wheat, It is characterized in that include: Acquire monitoring images of a wheat planting area and high-level semantic information of the monitoring images; Extracting wheat ear image features in the monitoring image; wherein the wheat ear image features include at least texture features and color features; fusing the high-level semantic information of the monitoring image and the wheat ear image features to obtain target features; Determining a wheat heading period recognition result according to the target characteristics; The step of fusing the high-level semantic information of the monitoring image and the wheat ear image features includes: Performing vector conversion on the high-level semantic information and wheat ear image features of the monitoring image respectively to obtain corresponding high-level semantic information vectors and wheat ear image feature vectors; fusing the high-level semantic information vector and the wheat ear image feature vector; Before fusing the high-level semantic information of the monitoring image with the wheat ear image features, the method further includes: Inputting the high-level semantic information and wheat ear image features of the monitoring image into a preset buffer layer, and using the buffer layer to buffer the high-level semantic information and wheat ear image features to weaken the difference between the high-level semantic information and the wheat ear image features; The step of extracting wheat ear image features from the monitoring image comprises: According to the growth direction of wheat, the monitoring image is divided into a plurality of feature domains; Extract image features of each feature domain; The wheat ear image features in the monitoring image are determined according to the image features of each feature domain.

2. The method according to claim 1, It is characterized in that Before extracting the wheat ear image features in the monitoring image, the method further includes: Preprocessing the monitoring image to obtain a preprocessed monitoring image; Performing decorrelation transformation processing on the preprocessed monitoring image to obtain a decorrelation transformed monitoring image; Threshold segmentation is performed on the surveillance image after the decorrelation transformation to obtain a surveillance image without background.

3. The method according to claim 1, It is characterized in that The step of obtaining the monitoring image of the wheat planting area comprises: Obtaining original monitoring images of wheat-growing areas; The original monitoring image is cropped according to a preset cropping size to obtain a plurality of monitoring images.

4. The method according to claim 3, It is characterized in that Determining the wheat heading period recognition result according to the target feature includes: Determine the heading period recognition result of each monitoring image according to the target features corresponding to each monitoring image; According to the heading period recognition result of each monitoring image, it is determined whether the wheat in the wheat planting area is in the heading period.

5. A wheat heading period identification device, It is characterized in that include: An acquisition module, used to acquire monitoring images of the wheat planting area and high-level semantic information of the monitoring images; A feature extraction module, used to extract wheat ear image features in the monitoring image; wherein the wheat ear image features at least include texture features and color features; A feature fusion module, used for fusing the high-level semantic information of the monitoring image and the wheat ear image features to obtain target features; An identification module, used to determine the wheat heading period identification result according to the target characteristics; The feature fusion module is specifically used for: Performing vector conversion on the high-level semantic information and wheat ear image features of the monitoring image respectively to obtain corresponding high-level semantic information vectors and wheat ear image feature vectors; fusing the high-level semantic information vector and the wheat ear image feature vector; The feature fusion is also used for: Inputting the high-level semantic information and wheat ear image features of the monitoring image into a preset buffer layer, and using the buffer layer to buffer the high-level semantic information and wheat ear image features to weaken the difference between the high-level semantic information and the wheat ear image features; The feature extraction module is specifically used for: According to the growth direction of wheat, the monitoring image is divided into a plurality of feature domains; Extract image features of each feature domain; The wheat ear image features in the monitoring image are determined according to the image features of each feature domain.

6. An electronic device, It is characterized in that include: at least one processor and memory; The memory stores computer-executable instructions; The at least one processor executes the computer-executable instructions stored in the memory, so that the at least one processor performs the method according to any one of claims 1 to 4.

7. A computer-readable storage medium, It is characterized in that The computer-readable storage medium stores computer-executable instructions, and when the processor executes the computer-executable instructions, the method according to any one of claims 1 to 4 is implemented.