Rice seedling age automatic detection method and device and seedling stage intelligent management and control system

Abstract

The invention belongs to the technical field of rice seedling age detection, and particularly discloses an automatic rice seedling age detection method and device and a seedling stage intelligent management and control system.The automatic rice seedling age detection method comprises the steps that rice seedling image information is preprocessed through water, a preprocessed image is obtained, and a single seedling is extracted from the preprocessed image; the rice seedling age automatic detection device comprises an image acquisition module and a microprocessor, and the seedling stage intelligent management and control system comprises a rice seedling age automatic detection device, an environmental data monitoring module, a storage module, a micro-jet and roller shutter opening and closing state acquisition module, a micro-jet and roller shutter opening and closing control module and a communication module. According to the method, a digital image processing technology with higher resolution is combined with a suitable target detection and recognition algorithm, the seedling age is judged based on the digital image processing technology, automatic detection of the seedling age in the rice seedling raising process is realized, and the problems of high labor intensity, time consumption in detection and inaccurate detection result caused by a manual observation method are effectively avoided.

Description

Technical Field

[0001] The present invention belongs to the technical field of rice seedling age detection technology, and in particular relates to a rice seedling age automatic detection method, device and seedling stage intelligent management and control system. Background Art

[0002] The cultivation of rice seedlings in cold regions is mainly carried out in greenhouses using the dry seedling cultivation method. Controlling temperature and humidity during different growth stages of the seedlings is the key to cultivating strong seedlings.

[0003] The growth period of seedlings is based on the age of seedlings. From sowing to the first leaf exposure is the seed root development period, the optimum temperature is 25℃~28℃, and generally no watering is required during this period. If the humidity is too high or the area is too wet, the film should be removed to disperse the moisture; from emergence to the first complete leaf expansion is the first complete leaf elongation period, the greenhouse temperature should be controlled at 22℃~25℃, and the maximum should not exceed 28℃. Ventilate and harden the seedlings in time, and achieve early ventilation, early seedling hardening and small seedling hardening. For water management, except for replenishing water in the too dry places of the seedbed, generally less watering or no watering is required to keep the seedbed Dry-raising state; from the tip of the second leaf to the third leaf unfolding is the weaning period, the maximum temperature does not exceed 25℃, the optimum temperature for the 2-leaf stage is 22℃~24℃, the optimum temperature for the 3-leaf stage is 20℃~22℃, especially at the 2.5-leaf stage, the temperature does not exceed 25℃, at this time, the seedlings should be ventilated and hardened, and the seedlings should also be ventilated and hardened on rainy days when the humidity in the greenhouse is high; from 3.1 to 3.5 leaves is the preparation period before transplanting, the focus is on controlling water to squat the seedlings and strengthen the roots, so that the seedlings are in a hungry and thirsty state, so as to facilitate good rooting and quick greening after transplanting. From the above seedling raising operation procedures, it can be seen that the detection of rice seedling age is crucial for the control of greenhouse temperature and humidity.

[0004] For a long time, the detection of rice seedling age in greenhouses has mainly been based on manual observation, which often requires plucking a seedling on site and then manually calculating the seedling age based on a complex seedling age calculation method. This is not only time-consuming and labor-intensive, and requires a large amount of work for the staff, but also poses a risk of inaccurate calculation results. Summary of the invention

[0005] In view of the above-mentioned problems, the object of the present invention is to provide a method and device for automatically detecting the age of rice seedlings and an intelligent management and control system for the seedling stage, which can detect the age of rice seedlings simply, quickly and accurately.

[0006] The technical solution of the present invention is: a method for automatically detecting the age of rice seedlings, comprising the following steps:

[0007] S1, preprocessing the rice seedling image information to obtain a preprocessed image;

[0008] S2, extracting individual seedlings from the preprocessed image;

[0009] S3, calculating the seedling age of a single seedling: finding all characteristic points of the single seedling according to the extracted single seedling, wherein the characteristic points include a root vertex, a leaf vertex and a stem-leaf intersection point;

[0010] Multiple preset ranges are screened by retaining the center point, retaining the center feature point and removing other feature points to obtain the target point;

[0011] Take the outer contour of the image of a single seedling, calculate the closest distance from each target point to the outer contour, set a threshold k, and when the distance is less than the threshold k, identify the target point as a root vertex or a leaf vertex, and then the one below is determined as a root vertex, otherwise it is a leaf vertex; when the distance is greater than the threshold k, identify the target point as a stem-leaf intersection;

[0012] Determine the number of leaves according to the number of leaf vertices; calculate the distance from the leaf vertex to the corresponding intersection point, which is the leaf length, sort according to the length, and determine the top two leaves;

[0013] The age of the seedlings is calculated by the length of the top two leaves and the number of leaves.

[0014] Furthermore, the specific steps of preprocessing the rice seedling image information are as follows: removing the background of the rice seedling image to obtain a pure seedling image; graying the pure seedling image to obtain a single-channel image; sharpening the single-channel image to obtain a sharpened image; equalizing the sharpened image to obtain a balanced image; extracting the stems and leaves of the rice seedlings in the balanced image, and thinning the extracted image to make the stems and leaves become thin lines, and then enriching the thin lines through an expansion algorithm to make the thin lines continuous lines to obtain an expanded image; and then denoising the expanded image to obtain a preprocessed image.

[0015] Furthermore, the specific steps of removing the background of the rice seedling image are: using the difference between the colors of other objects in the greenhouse and the green seedlings, using the RGB image, retaining the green channel, removing the background, and obtaining a pure seedling image.

[0016] Furthermore, the specific steps of sharpening the single channel image are: sharpening the single channel image by using Laplace transform, wherein the setting of the Laplace and matrix in the Laplace transform is digitally set according to the upward growth characteristic of rice.

[0017] Furthermore, the specific steps of performing equalization processing on the sharpened image are: performing equalization processing on the sharpened image using an adaptive histogram.

[0018] Furthermore, the specific steps of thinning the extracted image are: thinning the extracted image using a cross convolution method.

[0019] Furthermore, the specific steps for extracting a single seedling are as follows: using the region growing method, a cyclic method is used to search for the root point of the seedling from bottom to top as the seed point, the region growing algorithm is applied to extract the seed point area, and segmentation is performed to determine whether the segmented image is a complete single seedling. If not, continue searching; if complete, the extraction of a single seedling is achieved.

[0020] Furthermore, the specific steps for finding all the feature points of a single seedling are as follows: through the edge detection algorithm, all the feature points are traversed, and the angle calculation method is used to draw a circle with the current feature point as the center point and r as the radius, and other points on the image outside the current feature point are obtained on the circle, and the current feature point is used as the vertex to calculate the angle between the current feature point and the other two points on the circle. When the angle is an acute angle, the current feature point is determined to be the root vertex or leaf vertex or stem-leaf intersection in the individual plant, and so on, to find all the feature points in the seedling plant image.

[0021] Furthermore, the specific steps for calculating the age of the rice seedlings are:

[0022] The number of leaves is a, and the lengths of the top two leaves are L1 and L2 respectively, among which the length of the top leaf is L1, and the length of the second top leaf is L2. The ratio of the length of the top leaf to the length of the second top leaf is i, and the seedling age correction coefficient is j. The value of b is calculated according to the formula b=i×j. If b≥1, it means that the top leaf is a complete leaf and the seedling age is the whole leaf age, that is, the leaf age is a; if b<1, it means that the top leaf is an incomplete leaf, and the seedling age is (a-1)+0.1b.

[0023] A rice seedling age automatic detection device, comprising:

[0024] Image acquisition module, used to obtain rice seedling image information in real time;

[0025] The microprocessor is connected to the image acquisition module. The microprocessor is equipped with a seedling age recognition model. The seedling age recognition model calculates the rice seedling age based on the rice seedling age automatic detection method.

[0026] A seedling stage intelligent management and control system, applied to a cold-region rice seedling cultivation shed, comprises the above-mentioned rice seedling age automatic detection device, and also comprises:

[0027] An environmental data monitoring module is connected to the microprocessor and is used to obtain greenhouse environmental data in real time, the greenhouse environmental data including soil moisture data, soil temperature data, air temperature data and air humidity data;

[0028] A storage module, connected to the microprocessor, for storing picture data, greenhouse soil moisture data, soil temperature data, air temperature data and air humidity data;

[0029] The micro-spray and roller shutter opening and closing status acquisition module includes a travel switch and a switch quantity acquisition module. The micro-spray opening and closing status is directly acquired by the switch quantity acquisition module, and the roller shutter opening and closing status needs to be acquired by the switch quantity acquisition module to assist the travel switch to realize the state acquisition of the roller shutter opening, fully opening and fully closing;

[0030] A micro-spray and rolling curtain opening and closing control module, comprising a curtain motor, a solenoid valve and a control circuit, wherein the control circuit is connected to the microprocessor, the curtain motor is connected to the control circuit for opening and closing the rolling curtain, and the solenoid valve is connected to the control circuit for opening and closing the micro-spray;

[0031] The communication module uploads the seedling age detection results, real-time greenhouse environment data and the opening and closing status of the roller shutter through the network, providing data support for data archiving and remote control terminals.

[0032] Compared with the prior art, the beneficial effects of the present invention are as follows: the present invention calculates the seedling age during the rice greenhouse seedling raising period. At this time, the seedlings are small and densely clustered. The seedling age is judged based on the digital image processing technology using higher resolution combined with a suitable target detection and recognition algorithm, thereby realizing automatic detection of the seedling age during the rice seedling raising process, and effectively avoiding the problems of high labor intensity, time-consuming detection, and inaccurate detection results caused by manual observation methods. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] Figure 1 This is a flow chart of the method for automatically detecting the age of rice seedlings of the present invention;

[0034] Figure 2 It is a module diagram of the intelligent management and control system for seedling stage of the present invention. DETAILED DESCRIPTION

[0035] The following is combined with Figure 1 To the attached Figure 2 , the specific embodiments of the present invention are described in detail. In the description of the present invention, it should be understood that the terms "center", "upper", "lower", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inside", "outside" and the like indicate positions or positional relationships based on the positions or positional relationships shown in the drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as limiting the present invention.

[0036] The terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. Therefore, the features defined as "first" and "second" may explicitly or implicitly include one or more of the features; in the description of the present invention, unless otherwise specified, "plurality" means two or more.

[0037] It should be noted that the circuit connections involved in the present invention all adopt conventional circuit connection methods and do not involve any innovation.

[0038] Example 1

[0039] like Figure 1 The method for automatically detecting the age of rice seedlings shown comprises the following steps:

[0040] S1. Preprocessing rice seedling image information to obtain a preprocessed image.

[0041] S2. Extract individual seedlings from the preprocessed image.

[0042] S3, calculate the age of a single seedling: find all the feature points of a single seedling according to the extracted single seedling, and the feature points include the root vertex, leaf vertex and stem-leaf intersection. Multiple preset ranges are screened by retaining the center point, retaining the feature points at the center and removing other feature points to obtain the target point. Take the outer contour of the image of a single seedling, calculate the closest distance from each target point to the outer contour, set a threshold k, and when the distance is less than the threshold k, identify the target point as a root vertex or a leaf vertex, and then the one below is determined as a root vertex, otherwise it is a leaf vertex. When the distance is greater than the threshold k, identify the target point as a stem-leaf intersection. Determine the number of leaves according to the number of leaf vertices. Calculate the distance from the leaf vertex to the corresponding intersection, which is the leaf length, sort according to the length, and determine the top two leaves. The seedling age is calculated by the length of the top two leaves combined with the number of leaves.

[0043] This embodiment applies the latest technological achievements in digital image processing and combines the agronomic requirements of rice seedling detection to achieve automatic detection of rice seedling age.

[0044] Preferably, the specific steps of preprocessing the rice seedling image information are: removing the background of the rice seedling image to obtain a pure seedling image; graying the pure seedling image to obtain a single-channel image; sharpening the single-channel image to obtain a sharpened image; equalizing the sharpened image to obtain a balanced image; extracting the stems and leaves of the rice seedlings in the balanced image, and thinning the extracted image to make the stems and leaves become thin lines, and then enriching the thin lines through an expansion algorithm to make the thin lines continuous lines to obtain an expanded image; and then denoising the expanded image to obtain a preprocessed image.

[0045] Removing the background can completely focus on the seedlings and avoid the interference of background information on subsequent analysis. For example, when studying the growth status of seedlings, there is no need to consider irrelevant factors such as soil and weeds in the background, which improves the pertinence of the analysis. Reduce the amount of unnecessary data in the image. The background usually contains a lot of complex and useless information for seedling analysis. After removing the background, the workload of data processing and the storage space requirements can be effectively reduced.

[0046] In subsequent processing, the computational complexity of single-channel images is lower than that of color images. For example, when performing sharpening processing, there is no need to consider the relationship between multiple color channels, and the design and execution of the algorithm are simpler, which can improve processing efficiency. For seedling images, grayscale can highlight features such as the shape and structure of the seedlings. In some cases, color information may obscure the morphological characteristics of the seedlings themselves. Grayscale helps to highlight these key features, facilitating subsequent analysis of structures such as stems and leaves.

[0047] Sharpening can make the edges of stems and leaves in the seedling image clearer. When extracting stem and leaf features, clear edges help to accurately define their contours and improve the accuracy of feature extraction. Making the structure of the seedlings more obvious allows both manual observation and subsequent automated analysis to see the details of the seedlings more clearly, helping to determine the health status and growth stage of the seedlings.

[0048] Equalization can enhance the overall contrast of the image, making the stems and leaves of the seedlings more prominent in the image. In cases such as uneven lighting, equalization can adjust the brightness distribution of the image to ensure that the features of the stems and leaves are not obscured by shadows or overly bright areas. For images of seedlings collected under different lighting conditions, equalization helps to achieve a more consistent visual effect for images, facilitating unified analysis and processing.

[0049] Converting stems and leaves into thin lines can simplify their representation and present the basic structure of the seedlings more clearly. This is very helpful for tasks such as analyzing the topological structure of seedlings and calculating the proportion of stems and leaves. The thinned image can more accurately extract features such as the length and width of stems and leaves. The thin line form can reduce errors and make feature measurements more accurate. The lines of stems and leaves are continuous, ensuring the integrity of the seedling structure. In subsequent analysis, such as calculating the connectivity of stems and leaves and judging the continuity of growth, this complete line structure is very important. It can better preserve the overall characteristics of stems and leaves, avoid feature loss due to thin line breakage, and help to comprehensively and accurately analyze the morphological characteristics of seedlings.

[0050] Noise removal can avoid the interference of noise on the extraction and analysis of seedling stem and leaf features. Noise may lead to misjudgment of features such as the edge and length of stems and leaves. Denoising can improve the accuracy of the analysis results. In the subsequent various analysis operations on the preprocessed images, the analysis results are more stable and reliable due to the absence of noise interference. Whether it is long-term monitoring of seedling growth trends or comparative analysis between different seedlings, more reliable results can be obtained.

[0051] Preferably, the specific steps of removing the background of the rice seedling image are: using the difference between the color of other objects in the greenhouse and the color of the green seedlings, using the RGB image, retaining the green channel, removing the background, and obtaining a pure rice seedling image. Natural distinguishability

[0052] In RGB images, green seedlings usually have obvious color differences from other objects in the greenhouse. Green is the typical color of seedlings. Using this color difference for background removal is a natural method based on the characteristics of the object itself. For example, the soil in the greenhouse may be brown, and the shed may be gray. These colors are quite different from green in the RGB color space and are easy to distinguish. This method does not require complex shape analysis or prior knowledge to determine the background and seedlings. The distinction can be made directly by color features, and the design and understanding of the algorithm are relatively simple. In practical applications, it is easy to implement and can quickly obtain preliminary seedling image segmentation results. The green channel retains the information of the green part of the seedlings in the RGB image most directly. The chlorophyll of green plants has a high reflectivity of green light. Retaining the green channel can highlight the characteristics of the seedlings such as color and texture to the greatest extent. This helps to more accurately observe the growth status of the seedlings in subsequent analysis, such as the health of the leaves and the density of the leaves. Compared with using the full RGB image, retaining only the green channel reduces the amount of data. In subsequent processing that does not require color information, such as analysis of geometric features such as seedling shape and area, reducing the amount of data can improve processing efficiency and also reduce the demand for storage space.

[0053] Preferably, the specific steps of sharpening the single channel image are: sharpening the single channel image by Laplace transform, and the setting of the Laplace and matrix in the Laplace transform is digitally set according to the characteristic of rice growing upward. Sharpening can make the edges and details in the image clearer by enhancing the high-frequency components of the image, thereby improving the visual effect of the image.

[0054] Preferably, the specific steps of equalizing the sharpened image are: equalizing the sharpened image using an adaptive histogram. Adaptive histogram equalization can better preserve the details of the image and improve the local contrast of the image by performing histogram equalization on the local area of ​​the image. This method is particularly suitable for the situation where there are different brightness areas in the image, and the contrast can be enhanced within these areas, making the texture and structure of the image clearer.

[0055] Preferably, the specific step of thinning the extracted image is: thinning the extracted image using a cross convolution method. The cross convolution kernel can extract features in the horizontal and vertical directions through its unique structure, which helps to capture the line and edge information in the image. This structure can enhance the texture details of the image and make the structural features of the image more distinct.

[0056] Preferably, the specific steps for extracting a single seedling are: using the region growing method, searching for the root point of the seedling as the seed point from bottom to top in a cyclic manner, applying the region growing algorithm to extract the seed point area, performing segmentation, and judging whether the segmented image is a complete single seedling. If not, continue searching; if it is complete, the extraction of the single seedling is achieved.

[0057] Preferably, the specific steps for finding all the feature points of a single seedling are: through the edge detection algorithm, all the feature points are traversed, and the angle calculation method is used to draw a circle with the current feature point as the center point and r as the radius, and other points on the image outside the current feature point are obtained on the circle, and the current feature point is used as the vertex to calculate the angle between the current feature point and the other two points on the circle. When the angle is an acute angle, the current feature point is determined to be the root vertex or leaf vertex or stem-leaf intersection in the individual plant, and so on, to find all the special target points in the seedling plant image.

[0058] Preferably, the specific steps of calculating the age of the rice seedlings are:

[0059] The number of leaves is a, and the lengths of the top two leaves are L1 and L2 respectively, among which the length of the top leaf is L1, and the length of the second top leaf is L2. The ratio of the length of the top leaf to the length of the second top leaf is i, and the seedling age correction coefficient is j. The value of b is calculated according to the formula b=i×j. If b≥1, it means that the top leaf is a complete leaf and the seedling age is the whole leaf age, that is, the leaf age is a; if b<1, it means that the top leaf is an incomplete leaf, and the seedling age is (a-1)+0.1b.

[0060] Example 2

[0061] A rice seedling age automatic detection device comprises an image acquisition module and a microprocessor.

[0062] The image acquisition module is used to obtain high-definition color images of rice seedlings in real time. The microprocessor is connected to the image acquisition module, and the microprocessor is deployed with a seedling age recognition model, which calculates the seedling age based on the rice seedling age automatic detection method.

[0063] The image acquisition module obtains high-definition color pictures of rice seedlings with a pixel count of 6 million in real time, and transmits the uncompressed images to the microprocessor in real time through the USB protocol; the seedling age recognition model has been deployed in the microprocessor, and the model is the automatic detection method for rice seedling age proposed in Example 1.

[0064] Example 3

[0065] The invention discloses an intelligent management and control system for seedling stage, which is applied to the cultivation greenhouse of rice seedlings in cold regions. The planting system is placed in the greenhouse, which meets the use requirements of industrial-grade electronic equipment and realizes intelligent and automatic management and control of the growth of rice seedlings in cold regions.

[0066] like Figure 2 As shown, the intelligent management and control system includes an automatic rice seedling age detection device, an environmental data monitoring module, a storage module, a micro-spraying and rolling curtain opening and closing status acquisition module, a micro-spraying and rolling curtain opening and closing control module, and a communication module.

[0067] The environmental data monitoring module is connected to the microprocessor and is used to obtain greenhouse environmental data in real time. The greenhouse environmental data includes soil moisture data, soil temperature data, air temperature data and air humidity data; the environmental data monitoring module specifically includes a soil moisture sensor, a soil temperature sensor and an air temperature and humidity sensor.

[0068] The storage module is connected to the microprocessor and is used for storing picture data, greenhouse soil moisture data, soil temperature data, air temperature data and air humidity data.

[0069] The micro-spray and rolling shutter opening and closing status acquisition module includes a travel switch and a switch quantity acquisition module. The micro-spray opening and closing status is directly acquired by the switch quantity acquisition module, and the rolling shutter opening and closing status needs to be assisted by the travel switch through the switch quantity acquisition module to realize the status acquisition of the rolling shutter open, fully open and fully closed.

[0070] The micro-spray and rolling curtain opening and closing control module includes a curtain motor, a solenoid valve and a control circuit. The control circuit is connected to the microprocessor, the curtain motor is connected to the control circuit for controlling the opening and closing of the rolling curtain, and the solenoid valve is connected to the control circuit for controlling the opening and closing of the micro-spray; based on the collected opening and closing status of the micro-spray and rolling curtain, combined with the seedling age and greenhouse environment data collection results, according to the agronomic requirements of cold-region rice seedlings in different growth stages, the micro-spray and rolling curtain are opened and closed; the solenoid valve is a micro-spray solenoid valve, which is used to control the circulation of the micro-spray.

[0071] The communication module uploads the seedling age detection results, real-time greenhouse environment data and the opening and closing status of the roller shutter through the network, providing data support for data archiving and remote control terminals.

[0072] It should be noted that: this embodiment also includes a power management module, which provides various power sources such as 24V, 12V, 5V, 3.3V, etc. for the entire planting.

[0073] The seedling intelligent management and control system proposed in this embodiment was used to collect images of rice seedlings in the greenhouse during the rice seedling raising process in the No. 1 experimental greenhouse of the Rice Science and Technology Demonstration Park of the 856 Farm in Heilongjiang Province from April to May 2024. 15,000 data were collected from April 17, 2024, stored according to the date, and the data set was screened and cropped. The seedling image data set was obtained and the seedling age recognition model was deployed. Relying on digital image processing algorithms and high-precision image acquisition technology, the rice seedling images in the greenhouse are obtained in real time, and the seedling age discrimination model is established through the digital image processing algorithm. The model is deployed in the embedded monitoring equipment of the greenhouse. According to the relationship between seedling age and growth period, combined with the agronomic requirements of different growth periods in the seedling raising stage, the image acquisition-seedling age recognition-roller curtain ventilation and micro-spray irrigation automatic opening and closing of the greenhouse intelligent detection and control are realized.

[0074] The specific models of the above electronic components are not specially specified, and ordinary products available on the market can be selected as long as they can meet the use requirements of the present invention.

[0075] The specific embodiments described above further illustrate the objectives, technical solutions and beneficial effects of the present invention in detail. It should be understood that the above description is only a specific embodiment of the present invention and does not limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention are included in the protection scope of the present invention.

Claims

1. A method for automatically detecting the age of rice seedlings, characterized in that: The following steps are involved: S1, preprocessing the rice seedling image information to obtain a preprocessed image; S2, extracting individual seedlings from the preprocessed image; S3, calculating the seedling age of a single seedling: finding all characteristic points of the single seedling according to the extracted single seedling, wherein the characteristic points include a root vertex, a leaf vertex and a stem-leaf intersection point; Multiple preset ranges are screened by retaining the center point, retaining the center feature point and removing other feature points to obtain the target point; Take the outer contour of the image of a single seedling, calculate the closest distance from each target point to the outer contour, set a threshold k, and when the distance is less than the threshold k, identify the target point as a root vertex or a leaf vertex, and then the one below is determined as a root vertex, otherwise it is a leaf vertex; when the distance is greater than the threshold k, identify the target point as a stem-leaf intersection; Determine the number of leaves according to the number of leaf vertices; calculate the distance from the leaf vertex to the corresponding intersection point, which is the leaf length, sort according to the length, and determine the top two leaves; The age of the seedlings is calculated by the length of the top two leaves and the number of leaves.

2. The method for automatically detecting the age of rice seedlings according to claim 1, characterized in that: The specific steps of preprocessing the rice seedling image information are as follows: removing the background of the rice seedling image to obtain a pure rice seedling image; graying the pure rice seedling image to obtain a single channel image; The single channel image is sharpened to obtain a sharpened image; the sharpened image is equalized to obtain a balanced image; the stems and leaves of the rice seedlings in the balanced image are extracted, and the extracted image is thinned to make the stems and leaves into thin lines, and then the thin lines are enriched by the expansion algorithm to make the thin lines into continuous lines to obtain an expanded image; then the expanded image is denoised to obtain a preprocessed image.

3. The method for automatically detecting the age of rice seedlings according to claim 2, characterized in that: The specific steps of removing the background of the rice seedling image are as follows: using the difference between the colors of other objects in the greenhouse and the green seedlings, using the RGB image, retaining the green channel, removing the background, and obtaining a pure seedling image.

4. The method for automatically detecting the age of rice seedlings according to claim 2, characterized in that: The specific steps of sharpening the single channel image are as follows: sharpening the single channel image by using Laplace transform, wherein the setting of the Laplace and matrix in the Laplace transform is digitally set according to the upward growth characteristic of rice.

5. The method for automatically detecting the age of rice seedlings according to claim 2, characterized in that: The specific steps of equalizing the sharpened image are: using an adaptive histogram to equalize the sharpened image.

6. The method for automatically detecting the age of rice seedlings according to claim 1, characterized in that: The specific steps for extracting a single seedling are as follows: using the region growing method, a circular method is used to search for the root point of the seedling from bottom to top as the seed point, the region growing algorithm is applied to extract the seed point area, and segmentation is performed. The segmented image is judged to see whether it is a complete single seedling. If not, the search continues; if it is complete, the extraction of a single seedling is achieved.

7. The method for automatically detecting the age of rice seedlings according to claim 1, characterized in that: The specific steps to find all the feature points of a single seedling are: through the edge detection algorithm, all feature points are traversed, and the angle calculation method is used to draw a circle with the current feature point as the center point and r as the radius. The other points on the image outside the current feature point are obtained on the circle, and the current feature point is used as the vertex to calculate the angle between the current feature point and the other two points on the circle. When the angle is an acute angle, the current feature point is determined to be the root vertex or leaf vertex or stem-leaf intersection in the individual plant, and so on, to find all the special target points in the seedling plant image.

8. The method for automatically detecting the age of rice seedlings according to claim 1, characterized in that: The specific steps for calculating the age of seedlings are: The number of leaves is a, and the lengths of the top two leaves are L1 and L2 respectively, among which the length of the top leaf is L1, and the length of the second top leaf is L2. The ratio of the length of the top leaf to the length of the second top leaf is i, and the seedling age correction coefficient is j. The value of b is calculated according to the formula b=i×j. If b≥1, it means that the top leaf is a complete leaf and the seedling age is the whole leaf age, that is, the leaf age is a; if b<1, it means that the top leaf is an incomplete leaf, and the seedling age is (a-1)+0.1b.

9. A rice seedling age automatic detection device, characterized in that: include: Image acquisition module, used to obtain rice seedling image information in real time; A microprocessor is connected to the image acquisition module. The microprocessor is equipped with a seedling age recognition model. The seedling age recognition model calculates the rice seedling age based on the rice seedling age automatic detection method according to any one of claims 1 to 8.

10. A seedling stage intelligent management and control system, applied to a cold-region rice seedling cultivation shed, characterized in that: The device for automatically detecting the age of rice seedlings according to claim 9 further comprises: An environmental data monitoring module is connected to the microprocessor and is used to obtain greenhouse environmental data in real time, the greenhouse environmental data including soil moisture data, soil temperature data, air temperature data and air humidity data; A storage module, connected to the microprocessor, for storing picture data, greenhouse soil moisture data, soil temperature data, air temperature data and air humidity data; The micro-spray and roller shutter opening and closing status acquisition module includes a travel switch and a switch quantity acquisition module. The micro-spray opening and closing status is directly acquired by the switch quantity acquisition module, and the roller shutter opening and closing status needs to be acquired by the switch quantity acquisition module to assist the travel switch to realize the state acquisition of the roller shutter opening, fully opening and fully closing; A micro-spray and rolling curtain opening and closing control module, comprising a curtain motor, a solenoid valve and a control circuit, wherein the control circuit is connected to the microprocessor, the curtain motor is connected to the control circuit for opening and closing the rolling curtain, and the solenoid valve is connected to the control circuit for opening and closing the micro-spray; The communication module uploads the seedling age detection results, real-time greenhouse environment data and the opening and closing status of the roller shutter through the network, providing data support for data archiving and remote control terminals.

Images