A method and system for detecting water stress information of a greenhouse strawberry crop

Abstract

The present application relates to the technical field of crop water stress detection, and discloses a greenhouse strawberry crop water stress information detection method and system, wherein the greenhouse strawberry crop water stress information detection method comprises S1, which is composed of an inspection robot, a thermal imaging camera, a track and a client; the inspection robot is installed on the rear wall of the greenhouse, and the strawberry crops are planted in the greenhouse.The greenhouse strawberry crop water stress information detection method obtains water stress information based on a simplified water stress index formula of canopy temperature, compared with the traditional way of measuring canopy temperature which is limited to "point" or "local" area, resulting in the lack of sufficient representativeness of canopy temperature data, with the decrease of the cost of the inspection robot and the continuous improvement of the thermal infrared image temperature extraction technology, the thermal infrared sensor is carried by the inspection robot, which can quickly, real-time and large-scale capture the canopy thermal infrared image of the crops, and this technology provides strong support for accurately monitoring the water condition of the crops.

Description

Technical Field

[0001] This invention relates to the field of crop water stress detection technology, and in particular to a method and system for detecting water stress information in greenhouse strawberry crops. Background Technology

[0002] Greenhouse strawberries refer to strawberries cultivated in artificially controlled environments such as greenhouses or polytunnels. Strawberries are highly adaptable and easily cultivated, resulting in consistently high and stable economic benefits. Their development advantages are evident, and the cultivated area shows a steady upward trend. Besides being eaten fresh, strawberries can be processed into canned goods, dried strawberries, and other snack foods. They are also a top choice for the integration of agriculture and tourism, thus possessing broad market prospects. To maximize yield, strawberry production typically employs a one-year planting system. Fresh fruit cultivation involves planting in late August or early September, with harvesting starting in mid-November and continuing until May of the following year, coinciding with holidays such as Christmas, New Year's Day, and the Spring Festival, making it a seasonal winter fruit. Among various facility horticulture industries, strawberry cultivation offers significant economic advantages. Furthermore, high-quality seedlings are crucial for high strawberry yields, and seedling propagation represents a new economic growth point for the strawberry industry.

[0003] Water stress is a significant influencing factor in the growth of greenhouse strawberries. Water stress refers to the disruption or damage to the normal physiological functions of plants when water supply is insufficient or excessive. For greenhouse strawberries, water stress primarily occurs when soil moisture content is below or above the range required for normal growth.

[0004] Accurate and efficient diagnosis of crop water stress and maximizing the efficiency of agricultural water use are the foundation for achieving scientific irrigation in agriculture. Canopy temperature is determined by the heat transfer and conversion within the soil-plant-atmosphere continuum. Canopy temperature and its related indices can indicate the state of crop water stress, among which the crop water stress index is the most widely used.

[0005] Traditional methods of measuring canopy temperature are limited to "points" or "local" areas, making it difficult to fully reflect the overall distribution of canopy temperature in crop areas, resulting in canopy temperature data lacking sufficient representativeness. Summary of the Invention

[0006] Given that the existing traditional methods for measuring canopy temperature are limited to "points" or "local" areas, making it difficult to comprehensively reflect the overall distribution of canopy temperature in crop areas and resulting in a lack of sufficient representativeness in canopy temperature data, this invention is proposed.

[0007] Therefore, the purpose of this invention is to provide a method for detecting water stress information in greenhouse strawberry crops. The purpose is that, with the decrease in the cost of inspection robots and the continuous improvement of thermal infrared image temperature extraction technology, by using inspection robots equipped with thermal infrared sensors, it is possible to quickly, in real time and over a wide area capture thermal infrared images of crop canopies. This technology provides strong support for the accurate monitoring of crop water status.

[0008] To solve the above technical problems, the present invention provides the following technical solution: a method for detecting water stress information of greenhouse strawberry crops, including S1, which consists of an inspection robot, a thermal imaging camera, a track and a client. The inspection robot is installed on the back wall of the greenhouse, where strawberry crops are grown.

[0009] S2, the thermal imaging camera mounted on the rail-mounted inspection robot moves along the track to the target canopy position, and the camera takes pictures of the canopy at different positions by rotating the angle until all target canopies are collected;

[0010] S3. Using a self-made temperature calibration device, the conversion function formula between grayscale and temperature is obtained, and temperature correction is performed.

[0011] S4, threshold segmentation of the canopy thermal image is performed using the improved GA-Otsu algorithm;

[0012] S5, the original canopy thermal image and the image after threshold segmentation are added by image addition to generate a canopy mask carrying temperature information;

[0013] S6. Based on the selected strongly correlated environmental factor data, construct the dry and wet reference surface temperature prediction equation to achieve contactless acquisition of the dry and wet reference surface temperature.

[0014] S7. Based on the water stress index formula, the water stress index at each location is calculated using the obtained canopy temperature and dry / wet reference surface temperature.

[0015] S8. Based on the obtained water stress index at each location, a global water stress information heat map of the greenhouse is constructed using a bilinear interpolation algorithm to visualize the water status.

[0016] As a preferred embodiment of the present invention, it includes thermal infrared image acquisition. The accurate acquisition of canopy temperature is of core importance in the water stress information detection scheme. However, the accuracy of the infrared thermal imaging camera is affected by the shooting angle and distance.

[0017] As a preferred embodiment of the present invention, the original thermal infrared image obtained by the thermal infrared imager has a certain deviation in temperature information. In order to improve the accuracy of the image temperature information, it is necessary to calibrate the temperature information.

[0018] As a preferred embodiment of the present invention, the GA-Otsu threshold segmentation algorithm includes the following: the core idea is to calculate the inter-class variance after the image is segmented into two regions. However, in practical applications, when the gray level of the image to be segmented is high, the time required to find the optimal threshold will increase accordingly, thereby reducing the efficiency of image segmentation. To solve this problem, the genetic algorithm can be combined with the Otsu algorithm, which has the characteristics of global optimization and nonlinear fast convergence.

[0019] As a preferred embodiment of the present invention, the genetic algorithm is a search heuristic algorithm based on the natural selection and genetic mechanisms of Darwin's theory of evolution to simulate the natural evolution process in order to find the optimal solution. The essence of the genetic algorithm is to generate offspring with stronger adaptability by randomly exchanging information of excellent mothers. This is a cyclical selection process that is not limited by specific problems. The solution process is stable and has the advantages of strong robustness and suitability for parallel processing. When using the genetic algorithm to solve a problem, it is generally necessary to go through steps such as parameter encoding, population initialization, fitness function selection, genetic operation design, and control parameter setting.

[0020] In a preferred embodiment of the present invention, the canopy region mask extraction includes obtaining a binarized image of the canopy region by passing the image through a threshold segmentation algorithm. However, the binarized image only segments the canopy region and does not carry temperature information. Therefore, extracting the canopy mask that carries temperature information is a key step in obtaining the canopy temperature.

[0021] As a preferred embodiment of the present invention, the acquisition of the wet and dry reference surface temperature includes accurately determining the wet and dry reference surface temperature, which is the basis for determining the CWSI.

[0022] The beneficial effects of this invention are as follows: Water stress information is obtained by using a simplified water stress index formula based on canopy temperature. Compared to traditional methods that measure canopy temperature only at "points" or "local" areas, making it difficult to comprehensively reflect the overall distribution of canopy temperature in crop areas and resulting in insufficient representativeness of canopy temperature data, this method offers significant advantages. With the decreasing cost of inspection robots and the continuous improvement of thermal infrared image temperature extraction technology, inspection robots equipped with thermal infrared sensors can quickly, in real-time, and over a wide area capture canopy thermal infrared images of crops. This technology provides strong support for accurately monitoring crop water conditions. However, thermal infrared images contain a large amount of interfering background, greenhouse structures, and other ground feature information, resulting in a large number of mixed ground feature pixels in the extracted canopy temperature information, interfering with canopy temperature extraction. Therefore, removing interfering background from thermal infrared images is crucial for improving the accuracy of crop canopy temperature extraction. Attached Figure Description

[0023] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort. Wherein:

[0024] Figure 1 This is an experimental partition diagram of the present invention.

[0025] Figure 2 This is a schematic diagram of the temperature calibration device of the present invention.

[0026] Figure 3 This is the temperature grayscale conversion diagram for the first partition of the present invention.

[0027] Figure 4 This is the grayscale conversion diagram of the second zone temperature of the present invention.

[0028] Figure 5 This is the grayscale conversion diagram of the third zone temperature of the present invention.

[0029] Figure 6 This is the temperature correction diagram for the first zone of the present invention.

[0030] Figure 7 This is the temperature correction diagram for the second zone of the present invention.

[0031] Figure 8 This is the temperature correction diagram for the third zone of the present invention.

[0032] Figure 9 This is a flowchart of the image segmentation process of the GA-Otsu method of the present invention.

[0033] Figure 10 This is a diagram of the image algorithm of the present invention.

[0034] Figure 11 This is a dry and wet reference temperature difference diagram of the present invention.

[0035] Figure 12 This is the CWSI heat map of the present invention. Detailed Implementation

[0036] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0037] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the spirit of the invention. Therefore, the invention is not limited to the specific embodiments disclosed below.

[0038] Reference Figure 1-12 As an embodiment of the present invention, a method and system for detecting water stress information of greenhouse strawberry crops are provided. The method for detecting water stress information of greenhouse strawberry crops includes S1, which consists of an inspection robot, a thermal imaging camera, a track and a client. The inspection robot is installed on the back wall of the greenhouse, where strawberry crops are grown.

[0039] S2, the thermal imaging camera mounted on the rail-mounted inspection robot moves along the track to the target canopy position, and the camera takes pictures of the canopy at different positions by rotating the angle until all target canopies are collected;

[0040] S3. Using a self-made temperature calibration device, the conversion function formula between grayscale and temperature is obtained, and temperature correction is performed.

[0041] S4, threshold segmentation of the canopy thermal image is performed using the improved GA-Otsu algorithm;

[0042] S5, the original canopy thermal image and the image after threshold segmentation are added by image addition to generate a canopy mask carrying temperature information;

[0043] S6. Based on the selected strongly correlated environmental factor data, construct the dry and wet reference surface temperature prediction equation to achieve contactless acquisition of the dry and wet reference surface temperature.

[0044] S7. Based on the water stress index formula, the water stress index at each location is calculated using the obtained canopy temperature and dry / wet reference surface temperature.

[0045] S8. Based on the obtained water stress index at each location, a global water stress information heat map of the greenhouse is constructed using a bilinear interpolation algorithm to visualize the water status.

[0046] The system for detecting water stress information in greenhouse strawberry crops includes thermal infrared image acquisition. The accurate acquisition of canopy temperature is the core of the water stress information detection scheme. However, the accuracy of the infrared thermal imaging camera is affected by the shooting angle and distance.

[0047] To mitigate the impact on temperature data acquisition, this experiment will divide each row into the following sections when capturing images: Figure 1The three areas shown are: area one when the camera is at a 40° angle to the horizontal, area two when the angle is 30°, and area three when the angle is 25°. This is to allow for temperature correction by area in the following analysis, thereby improving the accuracy of canopy temperature extraction. The data acquisition time is from 09:00 to 15:00 (Beijing time), and the acquisition of the entire test area takes about 15 minutes.

[0048] The temperature information in the original thermal infrared image obtained by the thermal infrared imager has a certain deviation. In order to improve the accuracy of the temperature information in the image, it is necessary to calibrate the temperature information.

[0049] like Figure 2 The temperature calibration device shown uses a wet-bulb and dry-bulb temperature sensor placed in the water. A thermal imaging sensor captures thermal images of the water, and the temperature sensor measures the water temperature. The grayscale value of the water in the thermal infrared image is extracted. The water temperature for each zone (between 20 and 40°C) is measured 20 times. A conversion function formula between grayscale and temperature for each zone is constructed using the extracted measured water temperature and grayscale values. To reduce temperature errors, image temperature calibration is also required. This paper uses a portable handheld thermometer to measure the water temperature and establishes a calibration function between the measured temperature and the image temperature, based on the extracted water temperature. The temperature conversion for each group is shown below. Figure 3 , Figure 4 , Figure 5 Temperature corrections are described in detail below. Figure 6 , Figure 7 , Figure 8 .

[0050] The GA-Otsu threshold segmentation algorithm includes the following: The core idea is to calculate the inter-class variance after the image is segmented into two regions. However, in practical applications, when the gray level of the image to be segmented is high, the time required to find the optimal threshold will increase accordingly, which will reduce the efficiency of image segmentation. In order to solve this problem, the genetic algorithm can be combined with the Otsu algorithm with the characteristics of global optimization and non-linear fast convergence.

[0051] The genetic algorithm described is a search heuristic algorithm based on Darwin's theory of evolution, specifically natural selection and genetic mechanisms, used to simulate the natural evolutionary process to find optimal solutions. Essentially, the genetic algorithm generates more adaptable offspring by randomly exchanging information from superior mothers. This is a cyclical selection process, unrestricted by specific problems, with a stable solution process and advantages such as strong robustness and suitability for parallel processing. When using a genetic algorithm to solve a problem, it generally involves steps such as parameter encoding, population initialization, selecting a fitness function, designing genetic operations, and setting control parameters. Figure 9 The diagram illustrates the basic optimization process of GA in a practical problem;

[0052] Genetic algorithms can optimize Otsu's threshold selection process. The combined GA-Otsu algorithm can perform non-linear and fast iteration and adjustment of the optimal threshold parameters. This method can significantly improve the efficiency and performance of Otsu's algorithm when processing high-resolution images, thus enabling it to handle complex image segmentation tasks more efficiently, especially when the contrast between the target and the background is high, and thus achieve greater accuracy.

[0053] The canopy region mask extraction includes obtaining a binarized image of the canopy region by the image being processed by a threshold segmentation algorithm. However, the binarized image only segments out the canopy region and does not carry temperature information. Therefore, extracting the canopy mask that carries temperature information is the key step in obtaining the canopy temperature.

[0054] The masked portion of the canopy region is extracted by adding the original thermal image and the segmented image. Each pixel in the image has an integer pixel value. Adding the two images involves adding the pixel values ​​at the same positions. Finally, the calculation results are reconstructed into a new image based on the original positions. The principle is as follows: Figure 10 As shown;

[0055] In actual program development, the "+" operator is not used directly to add images. This is because if the result of the "+" operator exceeds 255, it takes the remainder of the sum divided by 255, which may actually make the pixel values ​​smaller after addition, affecting the accuracy of mask pixel extraction. However, OpenCV's `add()` method calculates the result equal to the pixel's saturation value of 255 (the maximum value) when the sum of the pixel values ​​of two pixels is greater than 255. The syntax of this method is as follows:

[0056] dst=cv2.add(src1,src2,mask,dtype)

[0057] Return value description:

[0058] dst: The image after addition. If the sum is greater than 255, then 255 is used.

[0059] Parameter description:

[0060] src1: The first image.

[0061] src2: The first image.

[0062] mask: Optional parameter, mask, usually the default value is used.

[0063] dtype: Optional parameter, image depth, usually using the default value.

[0064] The acquisition of the wet and dry reference surface temperature (WST) is fundamental to determining the CWSI. The wet WST is determined by spraying water onto the crop canopy leaves under fully irrigated conditions. The dry WST is equal to the temperature of non-transpirating leaves with fully closed stomata, measured by applying petroleum jelly to both sides of the crop canopy leaves under water stress conditions. Temperature differences are as follows: Figure 11 As shown;

[0065] The dry / wet reference surface temperature largely depends on the environmental conditions during the measurement process, such as air temperature and humidity, carbon dioxide concentration, and other environmental climatic factors. In this study, a handheld thermal imager was used to measure leaves with wet canopies and leaves coated with Vaseline. As shown in the figure, the treated leaves exhibit significant color differences in the thermal images. Simultaneously, 500 sets of data were collected, measuring atmospheric and greenhouse environmental factors. Based on the correlation between these environmental data and the dry / wet reference surface temperature, a multiple regression fitting formula was determined, enabling the prediction of the dry / wet reference surface temperature without the need for any reference surface.

[0066] Based on the selected strongly correlated environmental factor data, a prediction equation for dry and wet reference surface temperature is constructed as follows:

[0067] T wet =0.729T ro -0.270H ro -0.958T air +0.108H air +0.404T box +0.188H box +10.156

[0068] T dry =0.894T ro -0.326H ro -0.855T air +0.073H air +0.540T box +0.236H box +6.773

[0069] ___________________________________________________

[0070] In the formula: T wet -Wet reference surface temperature;

[0071] T dry - Dry reference surface temperature;

[0072] T ro -Temperature of sensors on the robot;

[0073] H ro - Sensors on the robot measure temperature and atmospheric humidity;

[0074] T air -Atmospheric temperature;

[0075] H air -Atmospheric humidity;

[0076] T box - Greenhouse wet-bulb sensor temperature;

[0077] H box - Humidity sensor for greenhouse wet-bulb system.

[0078] Based on practical production experience, strawberry crop stress of 0.6 or less is defined as mild stress, 0.6–0.8 as moderate stress, and 0.8 or greater as severe stress. To facilitate users' intuitive viewing of greenhouse water stress distribution, the client calculates the water stress index by reading the predicted dry and wet reference surface temperature from the sensor database and the canopy temperature obtained from the thermal image. However, single-point data cannot comprehensively reflect the overall water status of the greenhouse. Therefore, a bicubic interpolation algorithm is introduced to visualize the water stress data, thereby generating a global thermal map of greenhouse water stress, such as... Figure 12 The image shows a heat map of greenhouse water stress distribution at 09:00 AM on May 21, 2024. The color gradient visually displays the severity of water stress in different areas. Through this innovative method, greenhouse managers can quickly and intuitively grasp the distribution and changing trends of water stress, providing strong support for precision irrigation.

[0079] Accurate and efficient diagnosis of crop water stress, and maximizing the efficiency of agricultural water use, are fundamental to achieving scientific irrigation in agriculture. Canopy temperature is determined by the heat transfer and conversion within the soil-plant-atmosphere continuum. Canopy temperature and related indices can indicate crop water stress conditions, with the Crop Water Stress Index (CWSI) being the most widely used. This study uses a simplified water stress index formula based on canopy temperature to obtain water stress information. The formula is:

[0080]

[0081] In the formula: T C —Current average canopy temperature;

[0082] T wet —The canopy temperature of a crop when its stomata are fully open and its transpiration rate is at its maximum is also known as the wet reference surface temperature.

[0083] T dry —The canopy temperature of a crop when its stomata are completely closed and it is not transpiring is also known as the dry reference surface temperature.

[0084] Traditional methods of measuring canopy temperature are limited to "points" or "local" areas, failing to comprehensively reflect the overall distribution of canopy temperature in crop areas, resulting in canopy temperature data lacking sufficient representativeness. With the decreasing cost of inspection robots and the continuous improvement of thermal infrared image temperature extraction technology, using inspection robots equipped with thermal infrared sensors can quickly, in real-time, and over a wide area capture canopy thermal infrared images of crops. This technology provides strong support for accurately monitoring crop moisture status. However, thermal infrared images contain a large amount of interfering background information, greenhouse structures, and other ground features, resulting in a large number of mixed ground feature pixels in the extracted canopy temperature information, interfering with canopy temperature extraction. Therefore, removing interfering background information from thermal infrared images is crucial to improving the accuracy of crop canopy temperature extraction.

[0085] Accurate acquisition of canopy temperature and dry / wet reference surface temperature is fundamental to obtaining information on water stress in greenhouse strawberries. Accurate acquisition of canopy temperature mainly involves accurately removing background interference from thermal infrared images and using temperature correction formulas. Accurate acquisition of dry / wet reference surface temperature mainly involves screening for highly correlated environmental factors for prediction.

[0086] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A method for detecting water stress information in greenhouse strawberry crops, characterized in that, include: S1, the inspection device consists of an inspection robot, a thermal imaging camera, a track and a client. The inspection robot is installed on the back wall of the greenhouse, where strawberry crops are grown. S2, the thermal imaging camera mounted on the rail-mounted inspection robot moves along the track to the target canopy position, and the camera takes pictures of the canopy at different positions by rotating the angle until all target canopies are collected; S3. Using a self-made temperature calibration device, the conversion function formula between grayscale and temperature is obtained, and temperature correction is performed. S4, threshold segmentation of the canopy thermal image is performed using the improved GA-Otsu algorithm; S5, the original canopy thermal image and the image after threshold segmentation are added by image addition to generate a canopy mask carrying temperature information; S6. Based on the selected strongly correlated environmental factor data, construct the dry and wet reference surface temperature prediction equation to achieve contactless acquisition of the dry and wet reference surface temperature. S7. Based on the water stress index formula, the water stress index at each location is calculated using the obtained canopy temperature and dry / wet reference surface temperature. S8. Based on the obtained water stress index at each location, a global water stress information heat map of the greenhouse is constructed using a bilinear interpolation algorithm to visualize the water status.

2. A system for detecting water stress in greenhouse strawberry crops, characterized in that... This method is applied to the detection method for water stress information in greenhouse strawberry crops as described in claim 1.

Images