Corn growth monitoring method and system combining remote sensing technology and image recognition

Abstract

The invention discloses a corn growth monitoring method and system combining a remote sensing technology and image recognition, and relates to the technical field of corn growth monitoring. An unmanned aerial vehicle is carried with a visible light sensor and a multispectral sensor, and obtains a color RGB image and a multiband image of a to-be-recognized area; the color RGB image and the multi-band image are preprocessed; dividing sub-regions according to the preprocessed color RGB image; identifying a corn variety type in each sub-region, and selecting a corresponding growth cycle judgment model; and inputting the preprocessed multi-band image of each sub-region into the corresponding growth period judgment model, and determining the growth period of the corn in each sub-region. By combining the remote sensing technology and the image recognition technology, the problem of diversity of corn varieties in the region is solved, and the corn growth monitoring efficiency and accuracy are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of corn growth monitoring, and more particularly to a corn growth monitoring method and system combining remote sensing technology with image recognition. Background Art

[0002] The rapid development of UAV remote sensing technology in the agricultural field has made it a key tool for the rapid and accurate extraction of crop information. With high-precision sensors such as digital cameras, multispectral cameras and thermal infrared cameras integrated on the UAV platform, this technology can capture high-resolution images of farmland during low-altitude flight, greatly enhancing the ability to identify detailed features of surface crops, thus providing a new solution for real-time monitoring and precise management of crop growth.

[0003] For corn, a widely planted food crop around the world, its growth status is of vital importance to the stability of local economic and social development. In view of the diversity of corn varieties and the differences in growth environments, the scientific research community has conducted a lot of in-depth research on corn growth monitoring. For example, the research of Su Wei and others revealed that the remote sensing data collected by drones during ultra-low-altitude flights can most accurately extract the number of corn ridges during the corn jointing period; and Shao Guomin and others successfully achieved high-precision monitoring of corn leaf area index throughout the growth period through multispectral sensors carried by drones, combined with multivariate linear regression models and multiple vegetation indices.

[0004] When monitoring the growth of corn in a specific large area, a challenge that cannot be ignored is the diversity of corn varieties in the area. This diversity often makes it difficult for traditional monitoring methods to accurately distinguish the growth characteristics of different varieties, thus introducing judgment errors. Therefore, in view of the diversity of corn varieties in the region, how to accurately judge the growth of corn in each sub-region is a problem that technicians in this field urgently need to solve. Summary of the invention

[0005] In view of this, the present invention provides a corn growth monitoring method and system combining remote sensing technology and image recognition.

[0006] In order to achieve the above object, the present invention provides the following technical solutions:

[0007] A corn growth monitoring method combining remote sensing technology and image recognition comprises the following steps:

[0008] Step 1: The drone is equipped with a visible light sensor and a multispectral sensor to obtain a color RGB image and a multi-band image of the area to be identified respectively;

[0009] Step 2: preprocessing the color RGB image and the multi-band image;

[0010] Step 3: Divide the sub-regions according to the preprocessed color RGB image;

[0011] Step 4: Identify the corn variety type in each sub-region and select the corresponding growth cycle judgment model;

[0012] Step 5: Input the preprocessed multi-band image of each sub-region into the corresponding growth cycle judgment model to determine the growth period of corn in each sub-region.

[0013] Optionally, the growth period of corn is divided into five periods: sowing period, jointing period, trumpet period, tasseling period, and milky period.

[0014] Optionally, in step 2, the method for performing image preprocessing is:

[0015] Step 2.1, performing atmospheric correction on the color RGB image and the multi-band image to eliminate the influence of atmospheric absorption and scattering on the image quality;

[0016] Step 2.2, use filtering method to remove noise in the image and improve the signal-to-noise ratio and clarity of the image;

[0017] Step 2.3: Enhance the color and adjust the contrast of the image to highlight the features of corn in the image.

[0018] Optionally, in step 3, the method for dividing the sub-regions is:

[0019] Step 3.1, using the color information in the preprocessed color RGB image, perform color segmentation by using the K-means clustering algorithm to distinguish the areas of different colors in the image;

[0020] Step 3.2: Based on color segmentation, further boundary detection is performed, using the Canny edge detector to identify the edges of the regions in the image.

[0021] Step 3.3: Based on the results of color segmentation and boundary detection, perform region segmentation. Specifically, start from a set of seed points according to the region growing algorithm and merge adjacent pixels or regions with similar attributes into a larger sub-region.

[0022] Step 3.4: Optimize and adjust the initially divided sub-regions, including removing isolated noise regions, merging overly fragmented sub-regions, and smoothing sub-region boundaries.

[0023] Optionally, in step 4, the method for identifying the type of corn variety in each sub-region is:

[0024] Step 4.1, extracting key features from the preprocessed color RGB image of each sub-region, including at least color features, texture features, and shape features, and combining the extracted features into a feature vector for representing the corn features of each sub-region;

[0025] Step 4.2, pre-establishing a database containing characteristics of multiple corn varieties, wherein the database contains typical characteristic data of each corn variety;

[0026] Step 4.3, matching the feature vector of each sub-region with the data in the corn variety feature database;

[0027] Step 4.4: Use random forest to classify the feature vector to identify the type of corn variety in each sub-region.

[0028] Optionally, in step 4.3, data matching is performed by calculating the similarity between the feature vector of the sub-region and the data in the corn variety feature library.

[0029] Optionally, each corn variety type corresponds to a growth cycle judgment model, and the method for constructing the growth cycle judgment model is:

[0030] Data collection: Collect multi-band image data of corn fields of different corn varieties and different growth periods, and classify them according to corn variety types to obtain several image data sets;

[0031] Data annotation: Annotate the collected multi-band image data to identify the growth period of corn in each image;

[0032] Feature extraction: Extracting features related to the growth cycle from multi-band images, including at least the Normalized Difference Vegetation Index (NDVI);

[0033] Model construction: Select the convolutional neural network CNN and build an initial model based on the number of input features and output labels;

[0034] Model training: Use several labeled image data sets to train the same number of initial models to obtain several growth cycle judgment models corresponding to different corn varieties.

[0035] A corn growth monitoring system combining remote sensing technology and image recognition, comprising:

[0036] Image acquisition module, used for drones equipped with visible light sensors and multispectral sensors, to obtain color RGB images and multi-band images of the area to be identified respectively;

[0037] A preprocessing module, used for preprocessing the color RGB image and the multi-band image;

[0038] A sub-region division module is used to divide the sub-regions according to the pre-processed color RGB image;

[0039] Variety recognition module, used to identify the corn variety type in each sub-region and select the corresponding growth cycle judgment model;

[0040] The growth cycle judgment module is used to input the pre-processed multi-band image of each sub-region into the corresponding growth cycle judgment model to determine the growth period of corn in each sub-region.

[0041] It can be seen from the above technical solutions that the present invention provides a corn growth monitoring method and system combining remote sensing technology and image recognition, which has the following beneficial effects compared with the prior art:

[0042] Improve monitoring accuracy: By combining data from visible light sensors and multispectral sensors, the present invention can capture richer farmland information, including color, texture, spectral characteristics, etc., so as to more accurately identify corn varieties and judge their growth period. This method overcomes the judgment errors caused by the diversity of corn varieties in traditional methods and improves the accuracy and reliability of monitoring.

[0043] Diversity monitoring: By dividing the fields into sub-areas and analyzing the growth of corn in each sub-area independently, the present invention realizes diversity monitoring of farmland. Different sub-areas may have different corn growth conditions due to differences in varieties, soil and other conditions. This method can provide a customized judgment model for these differences, greatly improving the monitoring accuracy.

[0044] In summary, the present invention provides an efficient, accurate and automated corn growth monitoring method by combining remote sensing technology with image recognition technology, which solves the problem of diversity of corn varieties in the region and improves the efficiency and accuracy of corn growth monitoring. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] In order to more clearly illustrate the embodiments of the present invention 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 only embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on the provided drawings without paying creative work.

[0046] Figure 1 It is a schematic diagram of the method steps of the present invention;

[0047] Figure 2 It is a schematic diagram of the system module of the present invention. DETAILED DESCRIPTION

[0048] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0049] The embodiment of the present invention discloses a corn growth monitoring method combining remote sensing technology and image recognition, see Figure 1 , including the following steps:

[0050] Step 1: The drone is equipped with a visible light sensor and a multispectral sensor to obtain a color RGB image and a multi-band image of the area to be identified, respectively.

[0051] In this embodiment, a DJI multispectral drone (P4 Multispectral) is used to obtain data from the test area. The drone is equipped with one visible light sensor and five multispectral sensors, namely blue light, green light, red light, red edge and near infrared. The center of their bands are 450, 560, 650, 730 and 840 nm, respectively, and their resolutions are all 1.2 million pixels.

[0052] Step 2: Preprocess the color RGB image and the multi-band image:

[0053] Step 2.1, performing atmospheric correction on the color RGB image and the multi-band image to eliminate the influence of atmospheric absorption and scattering on the image quality;

[0054] Step 2.2, use filtering method to remove noise in the image and improve the signal-to-noise ratio and clarity of the image;

[0055] Step 2.3: Enhance the color and adjust the contrast of the image to make the features of the corn in the image more obvious, which is convenient for subsequent sub-region division and variety identification.

[0056] Step 3: Divide the sub-regions according to the preprocessed color RGB image. The specific method is as follows:

[0057] Step 3.1, using the color information in the preprocessed color RGB image, perform color segmentation by using the K-means clustering algorithm to distinguish the areas of different colors in the image;

[0058] Step 3.2: Based on color segmentation, further boundary detection is performed, using the Canny edge detector to identify the edges of the regions in the image.

[0059] Step 3.3: Based on the results of color segmentation and boundary detection, perform region segmentation. Specifically, start from a set of seed points according to the region growing algorithm and merge adjacent pixels or regions with similar attributes into a larger sub-region.

[0060] Step 3.4: Optimize and adjust the initially divided sub-regions, including removing isolated noise regions, merging overly fragmented sub-regions, and smoothing sub-region boundaries.

[0061] Step 4: Identify the corn variety type in each sub-region and select the corresponding growth cycle judgment model.

[0062] Among them, the method for identifying the type of corn variety in each sub-region is:

[0063] Step 4.1, extract key features from the preprocessed color RGB image of each sub-region, including at least color features (such as color histogram, color moment), texture features (such as gray level co-occurrence matrix, local binary pattern LBP), shape features (such as edge contour, regional shape parameters), etc., and combine the extracted features into a feature vector to represent the corn features of each sub-region;

[0064] Step 4.2, pre-establishing a database containing characteristics of multiple corn varieties, wherein the database contains typical characteristic data of each corn variety;

[0065] Step 4.3, performing data matching by calculating the similarity between the feature vector of the sub-region and the data in the corn variety feature database;

[0066] Step 4.4: Use random forest to classify the feature vector to identify the type of corn variety in each sub-region.

[0067] Step 5: Input the preprocessed multi-band image of each sub-region into the corresponding growth cycle judgment model to determine the growth period of corn in each sub-region.

[0068] In this embodiment, each corn variety type corresponds to a growth cycle judgment model, and the method for constructing the growth cycle judgment model is:

[0069] Data collection: Collect multi-band image data of corn fields of different corn varieties and different growth periods (divided into five periods: sowing period, jointing period, tasseling period, and milky period), and classify them according to corn variety types to obtain several image data sets;

[0070] Data annotation: Annotate the collected multi-band image data to identify the growth period of corn in each image;

[0071] Feature extraction: Extracting features related to the growth cycle from multi-band images, including at least the Normalized Difference Vegetation Index (NDVI), and in the specific implementation process, it may also include the Leaf Area Index (LAI), Fractional Vegetation Cover (FVC), etc.;

[0072] Model construction: Select the convolutional neural network CNN and build an initial model based on the number of input features and output labels;

[0073] Model training: Use several labeled image data sets to train the same number of initial models to obtain several growth cycle judgment models corresponding to different corn varieties.

[0074] Another embodiment of the present invention also provides a corn growth monitoring system combining remote sensing technology and image recognition, see Figure 2 ,include:

[0075] Image acquisition module, used for drones equipped with visible light sensors and multispectral sensors, to obtain color RGB images and multi-band images of the area to be identified respectively;

[0076] A preprocessing module, used for preprocessing the color RGB image and the multi-band image;

[0077] A sub-region division module is used to divide the sub-regions according to the pre-processed color RGB image;

[0078] Variety recognition module, used to identify the corn variety type in each sub-region and select the corresponding growth cycle judgment model;

[0079] The growth cycle judgment module is used to input the pre-processed multi-band image of each sub-region into the corresponding growth cycle judgment model to determine the growth period of corn in each sub-region.

[0080] As for the device disclosed in the embodiment, since it corresponds to the method disclosed in the embodiment, the description is relatively simple, and the relevant parts can be referred to the description of the method part.

[0081] The various embodiments in this specification are described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between the various embodiments can be referenced to each other.

[0082] The above description of the disclosed embodiments enables one skilled in the art to implement or use the present invention. Various modifications to these embodiments will be apparent to one skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to the embodiments shown herein, but rather to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A corn growth monitoring method combining remote sensing technology and image recognition, characterized in that: The following steps are involved: Step 1: The drone is equipped with a visible light sensor and a multispectral sensor to obtain a color RGB image and a multi-band image of the area to be identified respectively; Step 2: preprocessing the color RGB image and the multi-band image; Step 3: Divide the sub-regions according to the preprocessed color RGB image; Step 4: Identify the corn variety type in each sub-region and select the corresponding growth cycle judgment model; Step 5: Input the preprocessed multi-band image of each sub-region into the corresponding growth cycle judgment model to determine the growth period of corn in each sub-region.

2. A corn growth monitoring method combining remote sensing technology and image recognition according to claim 1, characterized in that: The growth period of corn is divided into five periods: sowing period, jointing period, trumpet period, tasseling period and milky period.

3. The corn growth monitoring method combining remote sensing technology and image recognition according to claim 1, characterized in that: In step 2, the method for image preprocessing is: Step 2.1, performing atmospheric correction on the color RGB image and the multi-band image to eliminate the influence of atmospheric absorption and scattering on the image quality; Step 2.2, use filtering method to remove noise in the image and improve the signal-to-noise ratio and clarity of the image; Step 2.3: Enhance the color and adjust the contrast of the image to highlight the features of corn in the image.

4. The corn growth monitoring method combining remote sensing technology and image recognition according to claim 1, characterized in that: In step 3, the method for dividing the sub-regions is: Step 3.1, using the color information in the preprocessed color RGB image, perform color segmentation by using the K-means clustering algorithm to distinguish the areas of different colors in the image; Step 3.2: Based on color segmentation, further boundary detection is performed, using the Canny edge detector to identify the edges of the regions in the image. Step 3.3: Based on the results of color segmentation and boundary detection, perform region segmentation. Specifically, start from a set of seed points according to the region growing algorithm and merge adjacent pixels or regions with similar attributes into a larger sub-region. Step 3.4: Optimize and adjust the initially divided sub-regions, including removing isolated noise regions, merging overly fragmented sub-regions, and smoothing sub-region boundaries.

5. The corn growth monitoring method combining remote sensing technology and image recognition according to claim 1, characterized in that: In step 4, the method for identifying the type of corn variety in each sub-region is: Step 4.1, extracting key features from the preprocessed color RGB image of each sub-region, including at least color features, texture features, and shape features, and combining the extracted features into a feature vector to represent the corn features of each sub-region; Step 4.2, pre-establishing a database containing characteristics of multiple corn varieties, wherein the database contains typical characteristic data of each corn variety; Step 4.3, matching the feature vector of each sub-region with the data in the corn variety feature database; Step 4.4: Use random forest to classify the feature vector to identify the type of corn variety in each sub-region.

6. A corn growth monitoring method combining remote sensing technology and image recognition according to claim 5, characterized in that: In step 4.3, data matching is performed by calculating the similarity between the feature vector of the sub-region and the data in the corn variety feature library.

7. The corn growth monitoring method combining remote sensing technology and image recognition according to claim 1, characterized in that: Each corn variety type corresponds to a growth cycle judgment model. The method for constructing the growth cycle judgment model is: Data collection: Collect multi-band image data of corn fields of different corn varieties and different growth periods, and classify them according to corn variety types to obtain several image data sets; Data annotation: Annotate the collected multi-band image data to identify the growth period of corn in each image; Feature extraction: Extracting features related to the growth cycle from multi-band images, including at least the Normalized Difference Vegetation Index (NDVI); Model construction: Select the convolutional neural network CNN and build an initial model based on the number of input features and output labels; Model training: Use several labeled image data sets to train the same number of initial models to obtain several growth cycle judgment models corresponding to different corn varieties.

8. A corn growth monitoring system combining remote sensing technology and image recognition, characterized in that: include: Image acquisition module, used for drones equipped with visible light sensors and multispectral sensors, to obtain color RGB images and multi-band images of the area to be identified respectively; A preprocessing module, used for preprocessing the color RGB image and the multi-band image; A sub-region division module is used to divide the sub-regions according to the pre-processed color RGB image; Variety recognition module, used to identify the corn variety type in each sub-region and select the corresponding growth cycle judgment model; The growth cycle judgment module is used to input the pre-processed multi-band image of each sub-region into the corresponding growth cycle judgment model to determine the growth period of corn in each sub-region.

Images