Plant root growth status analysis system and method

Abstract

The present invention relates to the field of artificial intelligence, and specifically to a system and method for analyzing the growth status of plant roots. The system comprises inputting image blocks into an encoder to obtain an encoded output for each image block, clustering the encoded outputs of the image blocks to obtain multiple classifications; inputting a class embedding and the encoded output of each image block into a decoder to obtain a decoder output corresponding to each image block and a decoder output for each class embedding; performing feature fusion on the decoder output and the corresponding class embedding based on the correspondence between the clustering results and the class embeddings, and outputting a root segmentation result after upsampling; and comparing the development of the same root system in two adjacent images in the sequence to obtain the growth status of the plant root system. The present invention improves the accuracy of root recognition and the precision of root analysis by introducing multiple class embeddings.

Description

Technical Field

[0001] The present invention relates to the field of artificial intelligence, and in particular to a system and method for analyzing plant root growth conditions. Background Art

[0002] Studying the distribution, depth, and density of root systems provides insight into how plants absorb water and nutrients under varying soil conditions. This information is crucial for developing more precise irrigation and fertilization plans, as different crops and soil types require different management strategies. For example, crops with deep root systems may require less frequent but deeper irrigation to encourage downward root growth, while shallow-rooted crops may require more frequent but shallow irrigation. Similarly, based on the distribution of root density, fertilization can be more targeted, ensuring that nutrients are supplied directly to areas with dense root systems, thereby reducing fertilizer waste and improving overall crop health and yield.

[0003] Methods for analyzing plant root systems primarily include direct observation, root traps and window methods, and underground imaging. Underground imaging, which does not require root destruction and allows for continuous observation of root growth, is gaining increasing popularity. However, root analysis primarily relies on manual labor, particularly the annotation of roots within images. This involves annotating images from different time periods to analyze root growth. Summary of the Invention

[0004] In order to solve the above problems, the present invention provides a method for analyzing plant root growth conditions based on multiple images, the method comprising the following steps:

[0005] Obtaining a sequence of plant root images taken at the same location but at different times, segmenting each image in the sequence of root images to obtain image blocks of each image, inputting the image blocks into an encoder to obtain an encoded output of each image block, and clustering the encoded outputs of the image blocks to obtain multiple classifications;

[0006] The class embedding and the encoded output of each image block are input into the decoder to obtain the decoder output corresponding to each image block and the decoder output of each class embedding. Based on the correspondence between the clustering results and the class embeddings, the decoder output and the corresponding class embedding are feature fused and the root segmentation result is output after upsampling. The number of clusters is the same as the number of class embeddings, and the number of class embeddings is a hyperparameter.

[0007] The growth status of the plant root system is obtained by comparing the development status of the same root system in two adjacent images in the sequence.

[0008] Preferably, clustering the encoding outputs of the image blocks is specifically performed as follows:

[0009] Cluster the image blocks according to their similarity. The number of clusters is the number of class embeddings. For each clustering result, retain the image block closest to the cluster center.

[0010] The coding output corresponding to the retained image block is used as an initial clustering center, and the coding output of the image block is clustered according to the similarity of the image block and the similarity of the coding output corresponding to the image block.

[0011] Preferably, clustering the coded outputs of the image blocks according to the similarity of the image blocks and the similarity of the coded outputs corresponding to the image blocks is specifically:

[0012] A first weight and a second weight are set, and the distance from the remaining image blocks to each cluster center is calculated using the first weight, the second weight, the similarity of the image blocks, and the similarity of the encoded outputs corresponding to the image blocks, and the encoded outputs of the image blocks are assigned to the cluster closest to the cluster center; wherein the sum of the first weight and the second weight is 1.

[0013] Preferably, according to the correspondence between the clustering results and the class embeddings, the output of the decoder and the corresponding class embedding are subjected to feature fusion, specifically:

[0014] Calculate the average value of the encoded outputs of all image blocks in each clustering result, and establish a corresponding relationship between the clustering result and the class embedding based on the average value;

[0015] Based on the correspondence, the class embedding corresponding to each decoder output is found; and the class embedding and the corresponding decoder output are feature fused.

[0016] Preferably, the comparing the development of the same root system in two adjacent images in the sequence to obtain the growth status of the plant root system is specifically as follows:

[0017] The segmentation results of each image are retained, and the root system changes between the two adjacent images are obtained based on the difference calculation results of the two adjacent images;

[0018] The segmentation result of the first image in the sequence is used as a benchmark, and the growth status of the root system is dynamically displayed on the benchmark.

[0019] In addition, the present invention also provides a plant root growth status analysis system based on multiple images, which includes the following modules:

[0020] an encoding clustering module for acquiring a sequence of plant root images taken at the same location but at different times, segmenting each image in the sequence of root images to obtain image blocks of each image, inputting the image blocks into an encoder to obtain an encoded output of each image block, and clustering the encoded outputs of the image blocks to obtain multiple classifications;

[0021] The image segmentation module is used to input the class embedding and the encoded output of each image block into the decoder to obtain the decoder output corresponding to each image block and the decoder output of each class embedding. Based on the correspondence between the clustering results and the class embeddings, the decoder output and the corresponding class embedding are feature fused and the root system segmentation result is output after upsampling. The number of clusters is the same as the number of class embeddings, and the number of class embeddings is a hyperparameter.

[0022] The root growth status analysis module is used to compare the development status of the same root system in two adjacent images in the sequence to obtain the growth status of the plant root system.

[0023] Preferably, clustering the encoding outputs of the image blocks is specifically performed as follows:

[0024] Cluster the image blocks according to their similarity. The number of clusters is the number of class embeddings. For each clustering result, retain the image block closest to the cluster center.

[0025] The coding output corresponding to the retained image block is used as an initial clustering center, and the coding output of the image block is clustered according to the similarity of the image block and the similarity of the coding output corresponding to the image block.

[0026] Preferably, clustering the coded outputs of the image blocks according to the similarity of the image blocks and the similarity of the coded outputs corresponding to the image blocks is specifically:

[0027] A first weight and a second weight are set, and the distance from the remaining image blocks to each cluster center is calculated using the first weight, the second weight, the similarity of the image blocks, and the similarity of the encoded outputs corresponding to the image blocks, and the encoded outputs of the image blocks are assigned to the cluster closest to the cluster center; wherein the sum of the first weight and the second weight is 1.

[0028] Preferably, according to the correspondence between the clustering results and the class embeddings, the output of the decoder and the corresponding class embedding are subjected to feature fusion, specifically:

[0029] Calculate the average value of the encoded outputs of all image blocks in each clustering result, and establish a corresponding relationship between the clustering result and the class embedding based on the average value;

[0030] Based on the correspondence, the class embedding corresponding to each decoder output is found; and the class embedding and the corresponding decoder output are feature fused.

[0031] Preferably, the comparing the development of the same root system in two adjacent images in the sequence to obtain the growth status of the plant root system is specifically as follows:

[0032] The segmentation results of each image are retained, and the root system changes between the two adjacent images are obtained based on the difference calculation results of the two adjacent images;

[0033] The segmentation result of the first image in the sequence is used as a benchmark, and the growth status of the root system is dynamically displayed on the benchmark.

[0034] In addition, the present invention also provides a computer program product, which implements the above method when executed by a processor.

[0035] In order to improve the accuracy of plant root segmentation, the present invention inputs image blocks into an encoder to obtain the encoded output of each image block, clusters the encoded output of the image blocks to obtain multiple classifications; inputs the class embedding and the encoded output of each image block into a decoder to obtain the decoder output corresponding to each image block and the decoder output of each class embedding; based on the correspondence between the clustering results and the class embeddings, the decoder output and the corresponding class embedding are feature fused, and the root segmentation result is output after upsampling.

[0036] The present invention has the following advantages: first, different class embeddings are used for different outputs according to the output of the decoder, avoiding the use of the same class embedding in the Segmenter model; second, multiple class embeddings are set, and different decoder outputs use different class embeddings, avoiding the problem of less learning content in class embeddings, improving the accuracy of segmentation, and greatly improving the growth analysis of plant roots. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] Figure 1 This is a flow chart of Example 1;

[0038] Figure 2 Root system images taken at different times at the same location for underground imaging;

[0039] Figure 3 This is the encoder structure diagram;

[0040] Figure 4 This is the decoder structure diagram;

[0041] Figure 5 Schematic diagram of the extracted root structure;

[0042] Figure 6 This is a structural diagram of Example 2. DETAILED DESCRIPTION

[0043] In this document, relational terms such as first and second are used solely to distinguish one entity or operation from another, and do not necessarily require or imply any actual relationship or order between these entities or operations. Furthermore, the terms "comprise," "include," or any other variations thereof are intended to encompass non-exclusive inclusion, such that a process, method, article, or device comprising a series of elements includes not only those elements but also other elements not explicitly listed, or elements inherent to such process, method, article, or device. Without further restriction, an element defined by the phrase "comprising a..." does not preclude the presence of other identical elements in the process, method, article, or device comprising the element. In the present invention, if the collection of personal privacy data involves, for example, facial features, mobile phone usage information, etc., the individual's permission will be obtained in advance, including but not limited to verbal reminders, posters, mobile phone reminders, etc.; if there is a conflict with laws and regulations, production or use will be within the scope permitted by laws and regulations.

[0044] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of them. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making any creative efforts shall fall within the scope of protection of the present invention.

[0045] In the first embodiment, the present invention provides a method for analyzing plant root growth conditions based on multiple images, such as Figure 1 As shown, the method includes the following steps:

[0046] S1, obtaining a sequence of plant root images taken at the same location but at different times, segmenting each image in the sequence of root images to obtain image blocks of each image, inputting the image blocks into an encoder to obtain an encoded output of each image block, and clustering the encoded outputs of the image blocks to obtain multiple classifications;

[0047] Using a plant root growth monitoring system, such as the CI-600 system, to capture a sequence of plant root images, you can obtain plant root growth at the same location at different time points, i.e., images at different moments in time. A plant root image sequence consists of images taken at different time points, arranged in time.

[0048] After acquiring the image, the image is preprocessed, and the preprocessing includes but is not limited to denoising, image enhancement, etc. In a specific embodiment, the preprocessing also includes removing the capillary roots in the image. The preprocessed image is segmented into multiple image blocks, and each image block is flattened to obtain the vector corresponding to the image block. The vector corresponding to the image block is input to the encoder, wherein the encoder preferably uses a Transformer Encoder, and the structure of the encoder is as follows Figure 3 When using the Segmenter model, a class embedding cls is involved. The class embedding cls is a learnable vector. Each class corresponds to a class embedding cls. However, the class embedding cls is a vector of fixed size and its learnable content is limited. Based on this, the present invention sets multiple class embedding cls for the same class.

[0049] After an image block is input to an encoder, the encoder will output multiple outputs, each of which corresponds to the input image block. After obtaining the encoded output of the image block, the encoded output of the image block is clustered to obtain multiple classifications. There are many ways to cluster, such as clustering based on the vector output of the encoder. In a more specific embodiment, the clustering of the encoded output of the image block is specifically as follows:

[0050] Cluster the image blocks according to their similarity. The number of clusters is the number of class embeddings. For each clustering result, retain the image block closest to the cluster center.

[0051] The number of class embeddings is the same as the number of clusters, meaning each cluster corresponds to a class embedding cls. When clustering, image blocks are first clustered. Similar image blocks generally contain the same or similar information, such as roots. When clustering, the image blocks are first clustered, and the image block closest to the cluster center is retained. For example, after clustering image blocks based on similarity, a total of five clusters are clustered. From the first cluster, the image block closest to the cluster center is selected.

[0052] In an alternative embodiment, the image blocks are clustered according to the similarity of the encoder outputs corresponding to the image blocks, or the image blocks are clustered according to the similarity of the image blocks and the similarity of the encoder outputs corresponding to the image blocks.

[0053] The coding output corresponding to the retained image block is used as an initial clustering center, and the coding output of the image block is clustered according to the similarity of the image block and the similarity of the coding output corresponding to the image block.

[0054] After obtaining the initial cluster centers, the coded outputs of the image blocks are clustered into multiple categories according to the similarity of the image blocks and the similarity of the coded outputs corresponding to the image blocks.

[0055] In a more specific embodiment, clustering the encoding outputs of the image blocks according to the similarity of the image blocks and the similarity of the encoding outputs corresponding to the image blocks is specifically as follows:

[0056] A first weight and a second weight are set, and the distance from the remaining image blocks to each cluster center is calculated using the first weight, the second weight, the similarity of the image blocks, and the similarity of the encoded outputs corresponding to the image blocks, and the encoded outputs of the image blocks are assigned to the cluster closest to the cluster center; wherein the sum of the first weight and the second weight is 1.

[0057] In this embodiment, the similarity of image blocks and the similarity of encoder output vectors are considered simultaneously, and different weights are set: a first weight and a second weight. The sum of the products of the weights and the similarities is used as the distance. For example, if the similarity between image blocks 1 and 2 is 0.7, and the similarity between the vector corresponding to the encoder output of image block 1 and the vector corresponding to the encoder output of image block 2 is 0.6, the first weight is 0.6, and the second weight is 0.4, then the distance between the two image blocks is 0.66. During clustering, each time a new image block is added to a cluster, the center of the corresponding cluster is updated.

[0058] S2: Input the class embedding and the encoded output of each image block into the decoder to obtain the decoder output corresponding to each image block and the decoder output of each class embedding. Based on the correspondence between the clustering results and the class embeddings, the decoder output and the corresponding class embedding are feature fused and the root segmentation result is output after upsampling. The number of clusters is the same as the number of class embeddings, and the number of class embeddings is a hyperparameter.

[0059] After obtaining the result of the encoder, the encoder result is input into the decoder. Preferably, the decoder uses Mask Transformer, which also has multiple class embeddings cls inputs and outputs multiple class embeddings cls after passing through Mask Transformer.

[0060] In a more specific embodiment, according to the correspondence between the clustering results and the class embeddings, the output of the decoder and the corresponding class embedding are subjected to feature fusion, specifically:

[0061] Calculate the average value of the encoded outputs of all image blocks in each clustering result, and establish a corresponding relationship between the clustering result and the class embedding based on the average value;

[0062] When using the model, for the average value of the encoded output corresponding to all image blocks in each cluster, the corresponding relationship between the clustering results and the class embedding is established based on the average value. For example, clustering result A corresponds to class embedding cls1, and clustering result B corresponds to class embedding cls3.

[0063] The average is calculated by summing the bit-wise encoding output vectors for all image blocks in each cluster, then calculating the average for each bit in the vector. For example, if the encoding output vectors for the first image block are [6, 2, 126] and [2, 8, 20], the average is [4, 5, 73]. Since the relationship between class embeddings and averages is recorded during training, when using the model for segmentation, the class embeddings corresponding to each cluster can be found based on this recorded relationship.

[0064] Based on the correspondence, the class embedding corresponding to each decoder output is found; and the class embedding and the corresponding decoder output are feature fused.

[0065] After obtaining the corresponding relationship, the class embedding and the decoder output are fused, such as Figure 4 As shown, if the image block corresponding to an output vector of the decoder belongs to cluster B, the corresponding class embedding is cls3. When feature fusion is performed, the decoder output vector corresponding to this image block will be feature fused with cls3. Since the decoder outputs of image blocks of different clusters correspond to different class embeddings, and class embeddings are obtained through learning, setting multiple class embeddings allows more knowledge to be learned, and the segmentation is more accurate when performing segmentation. It should be noted that there is only one category in the present invention: the root system, but this category has multiple class embeddings. This is different from the prior art in which different class embeddings are set according to different segmentation types. For example, in the prior art, if it is necessary to segment trees, sidewalks, and people in the image, three class embeddings Tree, Sidewalk, and Person will be set. The present invention only needs to segment one category: the root system, and the class embedding set is also one, but there are multiple ones. Figure 5 The present invention shows Figure 2 The result of segmentation on the right side.

[0066] S3, comparing the development conditions of the same root system in two adjacent images in the sequence to obtain the growth condition of the plant root system.

[0067] In the same sequence, the growth of the root system will be shown in the image. By segmenting the root system image using the above segmentation method, and then comparing the segmentation results of two adjacent images, the growth of the root system can be obtained. In a more specific embodiment, the growth status of the plant root system is obtained by comparing the development of the same root system in two adjacent images in the sequence, specifically:

[0068] The segmentation results of each image are retained, and the root system changes between the two adjacent images are obtained based on the difference calculation results of the two adjacent images;

[0069] The segmentation result of the first image in the sequence is used as a benchmark, and the growth status of the root system is dynamically displayed on the benchmark.

[0070] Before using the model, it needs to be trained. During the training process, the cluster relationship corresponding to each class embedding is stored, such as the average value of the encoded output corresponding to all image blocks in the cluster. During training, this average value is continuously updated. When using the model, the average value closest to the average value corresponding to the class embedding is found, or the correspondence between class embeddings and clusters is established through other methods, which are not specifically limited in this invention.

[0071] In the second embodiment, the present invention also provides a plant root growth status analysis system based on multiple images, such as Figure 6 As shown, the system includes the following modules:

[0072] an encoding clustering module for acquiring a sequence of plant root images taken at the same location but at different times, segmenting each image in the sequence of root images to obtain image blocks of each image, inputting the image blocks into an encoder to obtain an encoded output of each image block, and clustering the encoded outputs of the image blocks to obtain multiple classifications;

[0073] The image segmentation module is used to input the class embedding and the encoded output of each image block into the decoder to obtain the decoder output corresponding to each image block and the decoder output of each class embedding. Based on the correspondence between the clustering results and the class embeddings, the decoder output and the corresponding class embedding are feature fused and the root system segmentation result is output after upsampling. The number of clusters is the same as the number of class embeddings, and the number of class embeddings is a hyperparameter.

[0074] The root growth status analysis module is used to compare the development status of the same root system in two adjacent images in the sequence to obtain the growth status of the plant root system.

[0075] Preferably, clustering the encoding outputs of the image blocks is specifically performed as follows:

[0076] Cluster the image blocks according to their similarity. The number of clusters is the number of class embeddings. For each clustering result, retain the image block closest to the cluster center.

[0077] The coding output corresponding to the retained image block is used as an initial clustering center, and the coding output of the image block is clustered according to the similarity of the image block and the similarity of the coding output corresponding to the image block.

[0078] Preferably, clustering the coded outputs of the image blocks according to the similarity of the image blocks and the similarity of the coded outputs corresponding to the image blocks is specifically:

[0079] A first weight and a second weight are set, and the distance from the remaining image blocks to each cluster center is calculated using the first weight, the second weight, the similarity of the image blocks, and the similarity of the encoded outputs corresponding to the image blocks, and the encoded outputs of the image blocks are assigned to the cluster closest to the cluster center; wherein the sum of the first weight and the second weight is 1.

[0080] Preferably, according to the correspondence between the clustering results and the class embeddings, the output of the decoder and the corresponding class embedding are subjected to feature fusion, specifically:

[0081] Calculate the average value of the encoded outputs of all image blocks in each clustering result, and establish a corresponding relationship between the clustering result and the class embedding based on the average value;

[0082] Based on the correspondence, the class embedding corresponding to each decoder output is found; and the class embedding and the corresponding decoder output are feature fused.

[0083] Preferably, the comparing the development of the same root system in two adjacent images in the sequence to obtain the growth status of the plant root system is specifically as follows:

[0084] The segmentation results of each image are retained, and the root system changes between the two adjacent images are obtained based on the difference calculation results of the two adjacent images;

[0085] The segmentation result of the first image in the sequence is used as a benchmark, and the growth status of the root system is dynamically displayed on the benchmark.

[0086] In a third embodiment, the present invention further provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the method described above is implemented.

[0087] Through the description of the above embodiments, those skilled in the art will clearly understand that each embodiment can be implemented by adding the necessary general hardware platform, or of course, by combining hardware and software. Based on this understanding, the essence of the above technical solution or the portion that contributes to the prior art can be embodied in the form of a computer product. The present invention can take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0088] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it, and other embodiments may also be used. Although the present invention has been described in detail with reference to the aforementioned embodiments, ordinary technicians in this field should understand that they can still modify the technical solutions described in the aforementioned embodiments, or replace some 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 spirit and scope of the technical solutions of the various embodiments of the present invention.

Claims

1. A plant root growth analysis system, characterized in that: The system includes the following modules: an encoding clustering module for acquiring a sequence of plant root images taken at the same location but at different times, segmenting each image in the sequence of root images to obtain image blocks of each image, inputting the image blocks into an encoder to obtain an encoded output of each image block, and clustering the encoded outputs of the image blocks to obtain multiple classifications; An image segmentation module, which is used to input the class embedding and the encoding output of each image block into a decoder to obtain the decoder output corresponding to each image block and the decoder output of each class embedding; Based on the correspondence between the clustering results and the class embeddings, the decoder output and the corresponding class embeddings are feature fused, and after upsampling, the root system segmentation results are output. The number of clusters is the same as the number of class embeddings, and the number of class embeddings is a hyperparameter. a root growth status analysis module, configured to compare the development of the same root system in two adjacent images in the sequence to obtain the growth status of the plant root system; The encoding outputs of the image blocks are clustered, specifically: the image blocks are clustered according to the similarity of the image blocks, and the number of clusters is the number of class embeddings; for each clustering result, an image block closest to the cluster center is retained; the encoding output corresponding to the retained image block is used as the initial cluster center, and the encoding output of the image block is clustered according to the similarity of the image blocks and the similarity of the encoding outputs corresponding to the image blocks; According to the correspondence between the clustering results and the class embeddings, the decoder output and the corresponding class embedding are feature fused, specifically: the average value of the encoded outputs of all image blocks in each clustering result is calculated, and the correspondence between the clustering results and the class embeddings is established based on the average value; based on the correspondence, the class embedding corresponding to each decoder output is found; and the class embedding and the corresponding decoder output are feature fused.

2. The system according to claim 1, wherein The encoding outputs of the image blocks are clustered according to the similarity of the image blocks and the similarity of the encoding outputs corresponding to the image blocks, specifically: A first weight and a second weight are set, and the distance from the remaining image blocks to each cluster center is calculated using the first weight, the second weight, the similarity of the image blocks, and the similarity of the encoded outputs corresponding to the image blocks, and the encoded outputs of the image blocks are assigned to the cluster closest to the cluster center; wherein the sum of the first weight and the second weight is 1.

3. A method for analyzing plant root growth conditions, characterized in that: The method comprises the following steps: Obtaining a sequence of plant root images taken at the same location but at different times, segmenting each image in the sequence of root images to obtain image blocks of each image, inputting the image blocks into an encoder to obtain an encoded output of each image block, and clustering the encoded outputs of the image blocks to obtain multiple classifications; The class embedding and the encoded output of each image block are input into the decoder to obtain the decoder output corresponding to each image block and the decoder output of each class embedding. Based on the correspondence between the clustering results and the class embeddings, the decoder output and the corresponding class embedding are feature fused and the root segmentation result is output after upsampling. The number of clusters is the same as the number of class embeddings, and the number of class embeddings is a hyperparameter. Comparing the development of the same root system in two adjacent images in the sequence to obtain the growth status of the plant root system; The encoding outputs of the image blocks are clustered, specifically: the image blocks are clustered according to the similarity of the image blocks, and the number of clusters is the number of class embeddings; for each clustering result, an image block closest to the cluster center is retained; the encoding output corresponding to the retained image block is used as the initial cluster center, and the encoding output of the image block is clustered according to the similarity of the image blocks and the similarity of the encoding outputs corresponding to the image blocks; According to the correspondence between the clustering results and the class embeddings, the decoder output and the corresponding class embedding are feature fused, specifically: the average value of the encoded outputs of all image blocks in each clustering result is calculated, and the correspondence between the clustering results and the class embeddings is established based on the average value; based on the correspondence, the class embedding corresponding to each decoder output is found; and the class embedding and the corresponding decoder output are feature fused.

4. The method according to claim 3, wherein The encoding outputs of the image blocks are clustered according to the similarity of the image blocks and the similarity of the encoding outputs corresponding to the image blocks, specifically: A first weight and a second weight are set, and the distance from the remaining image blocks to each cluster center is calculated using the first weight, the second weight, the similarity of the image blocks, and the similarity of the encoded outputs corresponding to the image blocks, and the encoded outputs of the image blocks are assigned to the cluster closest to the cluster center; wherein the sum of the first weight and the second weight is 1.

5. The method according to claim 3, wherein The step of comparing the development of the same root system in two adjacent images in the sequence to obtain the growth status of the plant root system is specifically as follows: The segmentation results of each image are retained, and the root system changes between the two adjacent images are obtained based on the difference calculation results of the two adjacent images; The segmentation result of the first image in the sequence is used as a benchmark, and the growth status of the root system is dynamically displayed on the benchmark.

6. A computer storage device having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the computer program implements the method according to any one of claims 3 to 5.

Images