Crop growth information analysis device using multiplexed composite images and plant phenotype exploration method utilizing the same

Abstract

This invention provides a growth information analysis device that selects traits that can be measured without harvesting the plant (such as initial growth, internode length, and leaf area), and predicts growth by determining the relationship between these indicators and biomass. [Solution] The growth information analysis device 100 comprises a temperature analysis unit 110 that analyzes temperature from a thermal image, a spectral range analysis unit 120 that analyzes the spectral range from a hyperspectral image, a crop recognition unit 130 that separates and recognizes crops from the background in an RGB image, a video synthesis unit 140 that generates a multi-layered composite image using a merged image obtained by combining the thermal image, hyperspectral image, and RGB image, a data storage unit 150 that stores the multi-layered composite image and merged image synthesized by the video synthesis unit, and an artificial intelligence analysis unit 160 that uses the stored images to classify the merged images stored in the data storage unit according to the phenotype of the crop and construct a database.

Description

Technical Field

[0001] The invention of this application relates to a crop growth information analysis device using multiple synthetic images. More specifically, it relates to a technology for searching for the phenotypes of crops by combining image materials, which are crop growth phenotype materials, with crop growth information.

Background Art

[0002] As a prior art before the filing of the present invention, a method for measuring the phenotype of plants using images is disclosed. In this technology, a technology for measuring the phenotype of plants with respect to the vegetation index of crops using image processing technology is disclosed.

[0003] Also, as another prior art, a deep learning-based crop growth diagnosis method and system are disclosed. This technology secures a large number of stress phenotype RGB images of the growth state of bean crops taken due to nutrient stress, uses this as learning data to train a learning model, and discloses a configuration for constructing an object recognition model to diagnose the growth state of bean crops.

Prior Art Documents

Patent Documents

[0004]

Patent Document 1

Patent Document 2

Summary of the Invention

Problems to be Solved by the Invention

[0005] For successful crop cultivation in a standardized cultivation system like a smart farm, it is essential to select and cultivate crops with traits suited to the specific location and region. This requires accurately understanding the genetic properties of crops based on their phenotypes. Therefore, crop growth monitoring is essential, and it is necessary to classify and identify the genetic properties based on the crop's phenotype during the growth monitoring process. By utilizing this matching relationship between crop phenotypes and genetic properties, it will be possible to select and cultivate crops suitable for smart farms based on region, climate, and altitude.

[0006] To achieve the above objectives, accurate monitoring of crop growth is essential for identifying the physiological and ecological responses of crops to environmental stresses they may experience in the growing environment, such as drought, global warming, soil salinity, and heavy metals. Since non-destructive measurement is impossible for biomass, which is considered the most important measure of crop growth response, in order to observe changes in plant growth over various periods, numerous repetitions were set up from the beginning of the experiment for each treatment, and traits that could be measured without harvesting the plant (initial growth, internode length, leaf area, etc.) were selected using the relative growth method. After determining the relationship between these indicators and biomass, changes in crop growth due to the environment were observed by estimating changes in biomass.

[0007] To this end, we aim to provide a technology that predicts crop growth and quantity based on image data input from RGB, thermal, and hyperspectral imaging devices of crops. The multiplexed image device of this invention aims to provide a means to select the most suitable crop for the region where the smart farm is located by combining two or more image information from thermal imaging, hyperspectral imaging, and RGB image acquisition devices, and utilizing additional information through image synthesis in addition to analysis through existing image acquisition, comparing it with existing technologies to extract information related to effective crop growth, and combining this with phenotype.

[0008] This invention aims to provide a technology that allows for the selection of traits (such as initial growth, internode length, and leaf area) that can be measured without harvesting the plant using allometry, for use in studying the physiological and ecological responses of crops to extreme weather and environmental stress, and to predict growth by determining the relationship between these indicators and biomass. [Means for solving the problem]

[0009] To solve the above problems, the present invention provides the following means for solving the problems.

[0010] In a crop growth information analysis device using multiplexed composite images of crops and a plant phenotype exploration method utilizing the same, For capturing multiple composite images of the aforementioned crops, A thermal imaging camera and a temperature analysis unit that analyzes temperature from images acquired by the thermal imaging camera, A hyperspectral camera and a spectral analysis unit that analyzes the spectral range from the image acquired by the hyperspectral camera, An RGB camera and a crop recognition unit that separates and recognizes crops from the background based on the image acquired by the RGB camera, An image merging unit generates a single merged image from the image acquired by the thermal imaging camera, the image acquired by the hyperspectral camera, and the image acquired by the RGB camera. A video synthesis unit generates a multi-layered composite image using the images merged in the aforementioned image merging unit, A data storage unit that stores the image synthesized by the image synthesis unit and the merged image combined by the image merging unit, The present invention provides a crop growth information analysis device through multiplexed composite images and a plant phenotype search method using the same, characterized by comprising: an artificial intelligence analysis unit that uses the images stored in the data storage unit to classify the merged images stored in the data storage unit according to the phenotype of the crop and construct a database.

[0011] Furthermore, the image synthesis unit includes a heat stress analysis unit that analyzes heat stress using the thermal image and crop data recognized by the crop recognition unit,

[0012] A vegetation index analysis unit analyzes the vegetation index using the hyperspectral image and crop data recognized by the crop recognition unit, The present invention provides a crop growth information analysis device via multiplexed composite images and a plant phenotype exploration method utilizing the same, characterized by comprising an RGB analysis unit that performs RGB analysis using the aforementioned RGB images.

[0013] Furthermore, the RGB analysis unit is characterized by extracting trait information of the crop in order to analyze the characteristics of the crop based on its phenotypic properties. This provides a crop growth information analysis device through multiplexed composite images and a plant phenotype exploration method utilizing the same.

[0014] Furthermore, the present invention provides a crop growth information analysis device that uses multiplexed composite images, characterized in that the crop trait information includes initial growth, internode length, leaf width, and leaf length, and a plant phenotype exploration method utilizing the same.

[0015] Furthermore, the present invention provides a crop growth information analysis device via multiplexed composite images and a plant phenotype exploration method utilizing the same, characterized in that the information combined with the crop trait information includes fruit size, number of fruits, heat stress, and vegetation index.

[0016] Furthermore, the data storage unit is characterized by storing, along with the merged image, the following information about crop growth: crop type, water stress index, crop growth index, leaf length, leaf width, leaf height, crop surface temperature, growing medium temperature, video date, external temperature and humidity, smart farm internal temperature and humidity, crop number, and crop location. This provides a crop growth information analysis device through multiplexed composite video and a plant phenotype exploration method utilizing the same. [Effects of the Invention]

[0017] With the above-described configuration of the invention of the present application, by using trait information such as the initial growth, internode length, leaf width, leaf length, etc. of crops cultivated in a smart farm to classify the cultivated crops, whether to directly use the RGB image, thermal image, and hyperspectral image themselves, or to synthesize two or more of these images to monitor the suitability of the cultivation environment such as vegetation information and water stress, and by combining the relevance with the future yield related to the future income, etc., by obtaining and providing materials that combine the phenotype of the crop, the profitability of the crop, and the adaptability of the crop to the cultivation environment, even if the growers cultivate the same crop in different regions, climates, and topographies of the smart farm, they can select and cultivate the genetic characteristics of the crop according to the phenotype of the crop, which is a technology that can enhance the profitability of the smart farm.

Brief Description of Drawings

[0018] [Figure 1] This is the overall system configuration of the crop growth information analysis device through the multi-synthetic video of the present invention. [Figure 2] The sequence diagram shows the multi-synthetic video collection method using the crop growth information analysis device through the multi-synthetic video for collecting the crop trait data of the present invention. [Figure 3] The sequence diagram shows the data collection method for collecting the phenotype trait data of crops using the crop growth information analysis device through the multi-synthetic video of the present invention. [Figure 4] This shows the data structure for collecting the phenotype trait data of crops using the crop growth information analysis device through the multi-synthetic video of the present invention and storing it in a database. [Figure 5] This shows the user screen for collecting the phenotype trait data of crops using the crop growth information analysis device through the multi-synthetic video of the present invention, storing it in a database, and displaying it on the user screen. [Figure 6] The sequence diagram shows the process of collecting the phenotype trait data using the crop growth information analysis device through the multi-synthetic video of the present invention, storing it in a database, and analyzing the data stored in the database. [Figure 7]The present invention provides a step-by-step diagram illustrating a method for analyzing heat stress using a crop growth information analysis device that utilizes multiplexed composite images. [Figure 8] The present invention provides a step-by-step diagram illustrating a method for analyzing vegetation indices using a crop growth information analysis device that utilizes multiplexed composite images. [Figure 9] The present invention provides a sequence diagram illustrating an RGB image analysis method using a crop growth information analysis device that utilizes multiplexed composite images. [Figure 10] This invention describes a method for acquiring hyperspectral video images using a crop growth information analysis device that analyzes multiplexed composite images. [Figure 11] The conceptual diagram shows the complete system configuration of the crop growth information analysis device using multiplexed composite images according to the present invention. [Modes for carrying out the invention]

[0019] The effects and advantages of the present invention will be explained using the drawings as follows.

[0020] Accurate monitoring of crop growth is essential for studying the physiological and ecological responses of crops to environmental stresses such as drought, salinity, and heavy metals. Since non-destructive measurement is impossible for biomass, which is considered the most important measure of crop growth response, observing changes in plant growth over various periods requires numerous repetitions for each treatment from the initial stages of the experiment. Using relative growth methods, traits that can be measured without harvesting the plant (such as initial growth, internode length, and leaf area) are selected, and after determining the relationship between these indicators and biomass, changes in biomass can be estimated.

[0021] In the case of crops, leaf area is closely related to the biomass of the above-ground parts; therefore, if the leaf area can be accurately measured, changes in biomass can be estimated. In this invention, we attempted to predict crop growth and yield using images captured by RGB, thermal imaging, and hyperspectral imaging devices. To this end, we aimed to provide a technology that obtains information related to crop growth as accurately as possible by converting the images captured by the RGB, thermal imaging, and hyperspectral imaging devices into individual or multiple composite images.

[0022] Figure 1 shows the complete system configuration of the crop growth information analysis device using multiplexed composite images according to the present invention. The system includes a thermal imaging camera and a temperature analysis unit that analyzes the temperature from the image acquired by the thermal imaging camera for capturing multiplexed composite images of crops, a hyperspectral camera and a spectral analysis unit that analyzes the spectral range from the image acquired by the hyperspectral camera, and an RGB camera and a crop recognition unit that separates and recognizes crops from the background from the image acquired by the RGB camera.

[0023] The temperature analysis unit 110 measures the temperature of the crop surface and the temperature of the growing medium through the thermal imaging camera 111. The thermal imaging camera captures environmental changes such as water stress on the crop and provides essential information for evaluating the condition of the crop. Essential information that can be provided is that there are no temperature changes in the leaves due to leaf blight or pathogen infection.

[0024] The thermal imaging camera 111 uses infrared technology to sense the temperature distribution of crops and can confirm physiological changes in crops due to insufficient or excessive water.

[0025] The spectral range analysis unit 120 can measure the vegetation index (NDVI, etc.) and water stress of crops through hyperspectral images captured using the hyperspectral camera 121, and can detect pests and pest eggs by using different image acquisition wavelengths. The photosynthetic capacity and growth status of crops can be analyzed using hyperspectral images. For example, hyperspectral images of various wavelengths can be acquired, and the chlorophyll content and water status of crops can be analyzed based on the acquired wavelength bands. Through such analysis, crops that can withstand diseases, pests, and environmental stresses caused by heat and water supply can be identified, and by checking the phenotype of the crop, the genotype possessing environmental resistance based on the crop's phenotype can be identified.

[0026] The crop recognition unit 130 uses an RGB camera 131 to measure the size and morphology of the crop (e.g., fruit size, leaf length, leaf width, leaf height, etc.). Through this, the degree of crop growth and the presence or absence of visible pests and diseases can be determined. The RGB image uses light in a visually perceptible range to grasp the external shape of the crop and is used to monitor growth rate and size changes. The method for recognizing the crop can use existing image processing methods and the latest artificial intelligence technologies such as YOLO (You Only Look Once V7.0). In this invention, the artificial intelligence recognition method is configured to measure fruit size, leaf length, leaf width, leaf height, etc. in a single image by using additional learning data, and the image is configured to include the positional information of the crop. Furthermore, to enable further analysis in the future, the RGB camera captures R (red), G (green), and B (blue) images separately on CCD elements and saves them as BMP images without image loss.

[0027] The image synthesis unit 140 combines data obtained from the thermal imaging camera 111, the hyperspectral camera 121, and the RGB camera 131 to generate a stereo image. Since each image provides different information, combining them allows for a more comprehensive analysis of the crop's condition. Stereo image synthesis combines two or more images taken at the same location to provide three-dimensional information, which allows for the estimation of even more precise growth information.

[0028] Furthermore, in order to facilitate the management and use of the multiplexed composite video image file of the present invention, the system is configured to allow the RGB, thermal image, and hyperspectral image to be separated using a header, and to store and use them side by side in a single file. In the case of RGB video, the CCD element is used to acquire and store each R, G, and B as a lossless image in BMP format, the thermal image is also stored as a lossless BMP image, and the hyperspectral image is stored in a quantity equal to the number of wavelength bands captured by hyperspectral imaging.

[0029] The header is a header that can confirm that it is a multi-composite associative image, and is configured to start the file as MCI (Multi-Composition Image), then to indicate the RGB image merge position, it consists of the letter "R" + R image start position, the letter "R" + G image start position, the letter "R" + B image start position, to indicate the infrared merge position, it consists of the letter "I" + image start position, to indicate the hyperspectral image merge position, it consists of the letter "U" + number of images, the letter "U" + 1st image, ..., the letter "U" + last image position, image data from the RGB images, ..., the last hyperspectral image data, and the image termination tag "END", so that multi-composite image images composed of three types of image acquisition devices can be combined into a single file for easy use. Various information can be added after "END", and each data field can be divided into 8-character groups of letters and numbers and added to. In particular, the phenotype of a crop can be divided into 8 letters after "END", and an additional 8-character number can be added.

[0030] The system includes a data storage unit that stores the image synthesized by the image synthesis unit and the merged image combined by the image merging unit, and an artificial intelligence analysis unit that uses the images stored in the data storage unit to classify the merged images stored in the data storage unit according to the phenotype of the crop and constructs a database.

[0031] The image synthesis unit comprises a thermal stress analysis unit that analyzes thermal stress using the thermal image and crop data recognized by the crop recognition unit, a vegetation index analysis unit that analyzes vegetation index using the hyperspectral image and crop data recognized by the crop recognition unit, and an RGB analysis unit that performs RGB analysis using the RGB image.

[0032] The artificial intelligence analysis unit 160 analyzes the stored data through an artificial intelligence model.

[0033] The artificial intelligence analysis unit classifies RGB images, thermal images, and hyperspectral images stored in the data storage unit into phenotypes such as initial crop growth, internode length, leaf width, and leaf length. It further generates relationships between the classified phenotypes, classifying them into 5 to 20 categories according to the user's request. By combining attribute data such as fruit size, fruit count, water stress, and vegetation index, it can provide information that allows farmers to select and cultivate future crop phenotypes.

[0034] Furthermore, the artificial intelligence analysis unit can consist of a crop growth status determination unit, a pest and disease determination unit, and a water stress level determination unit, and is composed of a model trained using the multiplexed composite image. For this purpose, deep learning technology in the field of artificial intelligence may be used. In addition, the deep learning-based artificial intelligence algorithm can analyze the multiplexed composite image and provide the user with functions such as pest and disease diagnosis, growth trend analysis, and environmental stress response methods.

[0035] Figure 2 shows a step-by-step diagram of a method for collecting crop trait data using a crop growth information analysis device via multiplexed composite images according to the present invention. It shows a step-by-step diagram of trend analysis and prediction using the artificial intelligence data analysis unit of the present invention. RGB, thermal images, and hyperspectral images are acquired and combined into a stereo image. Unlike existing stereo images, this stereo image is generated from images of a single plant taken using two different image capture devices. This allows for comparison and analysis of two image images taken in different ways. The generated stereo image is analyzed by the artificial intelligence data analysis unit. Basically, the artificial intelligence data analysis unit includes a "crop growth state determination unit," a "disease and pest determination unit," and a "moisture stress level determination unit." To this end, it determines the crop's growth state by judging vertical growth from images measured for each individual crop, determines diseases and pests from image data of the crop's leaf surface, and primarily determines the moisture stress level using thermal images and hyperspectral images. The aforementioned artificial intelligence data analysis unit can be configured using deep learning technology and image processing technology, and can be realized through additional learning of a library distributed in package form or an artificial intelligence classification program. However, the unique crop artificial intelligence analysis technology of the present invention involves photographing crops in two or more different ways, creating stereo images from these images and using them for learning, using stereo images synthesized from RGB images and thermal images to identify crop speed and damage to growing points to understand the crop's growth state, using stereo images synthesized from RGB images and hyperspectral images to measure or predict crop diseases and pests, and generating a water stress index using a stereo composite image of thermal images and hyperspectral images.

[0036] Figure 3 shows a step-by-step diagram of a data collection method for collecting phenotypic data of crops using the crop growth information analysis device via multiplexed composite images of the present invention. It shows the process of acquiring RGB, thermal images, and hyperspectral images based on the crop's phenotype (initial growth, internode length, leaf area, etc.) and generating a composite image from them. At this time, if no similarity is formed between the images acquired by two different devices, the thermal image and hyperspectral image are acquired by adjusting the shooting conditions, and if the image similarity reaches 90% or more, the images are composited. At this time, the composite image was synthesized as a stereo image for convenience in future use.

[0037] Figure 4 shows the data structure for collecting and creating a database of crop phenotypic data using the crop growth information analysis device via multiplexed composite images of the present invention. Here, the abbreviations PK = primary key and FK = foreign key, indicating that they are used when referencing the PK key of another database, and 1:1 means that the data in the database have a one-to-one correspondence.

[0038] Figure 5 shows a user screen for displaying crop phenotypic data collected and stored in a database using the crop growth information analysis device via multiplexed composite images of the present invention. Users can connect to the database and select crop growth trends for individual crops or for the entire smart farm. At this time, the analysis results are displayed through RGB images, thermal images, and hyperspectral images selected by the user. In particular, the artificial intelligence analysis and results include simple image analysis as well as analysis results using stereo images synthesized from images taken using two or more methods.

[0039] Figure 6 shows a step-by-step diagram illustrating the process of collecting phenotypic data, creating a database, and analyzing the database data using the crop growth information analysis device via multiplexed composite images of the present invention. To classify, store, synthesize, analyze, and utilize RGB images, thermal images, and hyperspectral images captured and stored within the smart farm according to crop characteristics, it is necessary to verify that the RGB images, thermal images, and hyperspectral images were captured appropriately and can be used in the future. For example, the step-by-step diagram shows the process of checking for issues such as out-of-focus images, whether information obtained from images of the same crop at the same location is related, and whether the crop is not clearly visible in images where the light intensity or light source position is incorrect. These images are then saved and used for growth and yield prediction. If an image is found to be incorrect, it is either re-captured or the data is deleted to avoid incorrectly affecting the analysis.

[0040] Figure 7 shows a step-by-step diagram illustrating a method for analyzing heat stress using the crop growth information analysis device via multiplexed composite images according to the present invention. Using RGB images and thermal images, the position of the crop and the separation of the crop from the background are performed. While the RGB images are used, the image captured by the R (red) CCD may not be distinguished from the thermal image. Alternatively, for accurate analysis, it can be excluded to create a stereo image for analysis. The RGB images and thermal images are combined to measure the crop surface temperature and generate a thermal map reflecting the temperature characteristics of each crop. A critical temperature is set to distinguish between crops experiencing heat stress. Crops above the critical temperature are displayed in red, and crops below the critical temperature are displayed in blue to identify abnormal heat regions. The identified stereo image is input to the artificial intelligence data analysis unit to analyze growth information, visualize the thermal map, and provide it to the user. At this time, if necessary, comparative analysis is performed to determine whether the abnormal heat region is due to environmental factors such as the smart farm's installation location or differences in heat tolerance due to the crop's genetic traits. Experimental groups will be created using methods such as crop transplantation, and further analysis will be conducted in separate experiments.

[0041] Figure 8 shows a step-by-step diagram illustrating a method for analyzing the vegetation index using the crop growth information analysis device via multiplexed composite images according to the present invention. NDVI (Normalized Difference Vegetation Index) is a vegetation index used to quantify vegetation by measuring the difference between near-infrared light, which vegetation strongly reflects, and red light, which vegetation strongly absorbs. Here, vegetation refers to chlorophyll. In other words, healthy vegetation (chlorophyll) reflects more near-infrared (NIR) and green light and absorbs more red light compared to other wavelengths. This is why healthy plants appear green to our eyes.

[0042] Plant chlorophyll strongly absorbs visible light, while leaf cells strongly reflect near-infrared light. Therefore, by observing near-infrared (NIR) changes compared to red light, the state of chlorophyll can be accurately displayed, allowing for the analysis of plant health.

[0043] NDVI can be calculated using the formula: NDVI = (NIR - RED) / (NIR + RED).

[0044] An NDVI value of 0 indicates no green vegetation, while a value of 1 indicates the highest density of green foliage.

[0045] The RED value is obtained using a CCD that measures only RED in an RGB video image. After analyzing the RGB video image, the vegetation index of the crop is accurately calculated using only the RED information values ​​corresponding to the location of the crop. For the NIR value, an image taken in the near-infrared region, specifically wavelengths between 750nm and 2500nm, can be used, but it is preferable to use an image between 900nm and 1100nm.

[0046] Figure 9 shows a step-by-step diagram of the RGB image analysis method using the crop growth information analysis device via multiplexed composite images according to the present invention.

[0047] The RGB camera of this invention uses a camera equipped with separate CCD elements that capture red, green, and blue colors, respectively, and have high sensitivity in dark places. In this way, images can be captured and stored separately for each color, and combined and used when necessary. Specifically, the technology includes acquiring an image using a 3-channel visible light camera, separating the background to distinguish the crops from the acquired image, distinguishing the crops individually from the image with the background separated, and separating the phenotype by searching for parts related to the phenotypic traits of the distinguished crop individuals. Such information on individual crops, crop locations, and crop phenotype extraction is stored and combined with images acquired from other image acquisition technologies to create a stereo image, which can then be used in combination with the genetic characteristics linked to the phenotype. That is, since the same phenotype may have the same crop growth morphology, tolerance to water stress, yield, and the same or similar taste and aroma of the harvested crop or fruit, it is very important to search for the phenotype from the RGB image. Figure 9 shows this process in a step-by-step diagram. Naturally, the video images captured in RGB are used to monitor crop growth and are saved so that they can be used for analysis.

[0048] Figure 10 shows a method for acquiring hyperspectral images using the crop growth information analysis device via multiplexed composite images according to the present invention. To obtain hyperspectral images in the near-infrared region, a wavelength band for measuring hyperspectral data is set, and as many wavelengths as possible are selected and captured while increasing or decreasing the wavelengths of the image for the crop in that wavelength band.

[0049] Therefore, for a given location on a crop, information is measured in the range of 750nm to 2500nm, for example. The information for a given point is displayed as a two-dimensional graph, with the horizontal axis representing the selected light wavelength and the vertical axis representing the intensity of reflected light measured at each wavelength. Because a great deal of information is generated, it is possible to reduce the number of wavelengths measured through experiments and research. However, the more information obtained, the more we can learn about the crop using that information. By selecting and photographing near-infrared light wavelengths between 900 and 1100nm, chlorophyll, i.e., the vegetation index, can be measured. By selecting the near-infrared wavelength range of 1200 to 1300nm, insects, insect larvae, and eggs can be identified. By measuring hyperspectral images using the 3000nm near-infrared wavelength, increased crop leaf production can be confirmed. In other words, if the temperature of crop leaves becomes high when measured with a thermal imaging camera in the presence of sunlight, it is possible to confirm whether increased water production is occurring due to water evaporation from the crop leaves using hyperspectral images in the 3000nm band, and this can be used to generate a water stress index.

[0050] Figure 11 is a conceptual diagram showing the entire system configuration of the crop growth information analysis device using multiplexed composite images according to the present invention. One embodiment of the present invention involves capturing various images of a smart farm or open-field crop using an RGB camera, a thermal imaging camera, and a hyperspectral camera, verifying whether the captured images contain sufficient information about the crop using an artificial intelligence image verification algorithm, then synthesizing the image images captured by two or more devices to generate a stereo image, storing the generated stereo image in a database, and inputting this into an artificial intelligence model in real time or non-real time to predict crop growth status, growth rate, diseases and pests, and crop phenotype for yield and growth prediction. [Explanation of Symbols]

[0051] 100...Growth information analysis device, 110...Temperature analysis unit, 111...Thermal imaging camera, 120...Spectral range analysis unit, 121...Hyperspectral camera, 130...Crop recognition unit, 131...RGB camera, 140...Image synthesis unit, 141...Heat stress analysis, 142...Vegetation index analysis, 143...RGB analysis, 150...Data storage unit, 160...Artificial intelligence analysis

Claims

1. In a crop growth information analysis device that uses multiplexed composite images of crops, For capturing multiple composite images of the aforementioned crops, A thermal imaging camera and a temperature analysis unit that analyzes temperature from the thermal image acquired by the thermal imaging camera, A hyperspectral camera and a spectral analysis unit that analyzes the spectral range from the hyperspectral image acquired by the hyperspectral camera, An RGB camera and a crop recognition unit that separates and recognizes crops from the background using RGB images acquired by the RGB camera, An image merging unit generates a single merged image from the thermal image acquired by the thermal image camera, the hyperspectral image acquired by the hyperspectral camera, and the RGB image acquired by the RGB camera. A video synthesis unit generates a multi-layered composite image using the merged images that have been combined in the image merging unit, A data storage unit that stores the multiplexed composite video synthesized in the video synthesis unit and the merged image merged in the image merging unit, An artificial intelligence analysis unit uses the images stored in the data storage unit to classify the merged images stored in the data storage unit according to the phenotype of the crop and constructs a database. A crop growth information analysis device characterized by being equipped with the following.

2. The aforementioned image synthesis unit includes a thermal stress analysis unit that analyzes thermal stress using the thermal image and crop data recognized by the crop recognition unit, A vegetation index analysis unit analyzes the vegetation index using the hyperspectral image and crop data recognized by the crop recognition unit, It consists of an RGB analysis unit that performs RGB analysis using the aforementioned RGB image, A crop growth information analysis device according to claim 1, characterized in that...

3. The crop growth information analysis apparatus according to claim 2, characterized in that the RGB analysis unit extracts trait information of the crop in order to analyze the characteristics of the crop based on the phenotypic properties of the crop.

4. The crop growth information analyzer according to claim 3, characterized in that the phenotypic information of the crop is initial growth, internode length, leaf width, and leaf length.

5. The crop growth information analyzer according to claim 4, characterized in that the information combined with the trait information of the crop is fruit size, number of fruits, heat stress, and vegetation index.

6. The crop growth information analysis device according to claim 5, characterized in that the data storage unit stores, along with the merged image, the crop type, water stress index, crop growth index, leaf length, leaf width, leaf height, crop surface temperature, growing medium temperature, video date, external temperature and humidity, smart farm internal temperature and humidity, crop number, and crop location.

7. A method for analyzing crop growth information using a crop growth information analysis device through multiple composite images of crops, The aforementioned device In order to capture a superimposed image of the aforementioned crop, The temperature is analyzed from the thermal image camera and the thermal image image acquired by the thermal image camera. A hyperspectral camera and the range of the spectrum from the hyperspectral image acquired by the hyperspectral camera are used to analyze the spectrum. The system recognizes crops by separating them from the background using an RGB camera and the RGB images acquired by the RGB camera. A single merged image is generated from the thermal image acquired by the thermal imaging camera, the hyperspectral image acquired by the hyperspectral camera, and the RGB image acquired by the RGB camera. Using the aforementioned merged image, a multi-layered composite image is generated. The multiplexed composite video and the merged image are stored in the data storage unit. Using the images stored in the data storage unit, the combined images stored in the data storage unit are classified according to the phenotype of the crop, and a database is constructed. An analytical method that includes the following.

Images