Multi-spectral intelligent regulation and control system and method for precise cultivation of plants

Abstract

The invention discloses a multispectral intelligent regulation and control system and method for precise cultivation of plants. The regulation and control system comprises a control system, a six-channel multispectral plant growth regulation and control lamp, a binocular camera and a monocular camera. According to the multi-spectral plant growth regulation and control lamp, six optical films with specific spectral characteristics are carried, and a light source position regulation mechanism driven by a steering engine is matched, so that the limitation of a fixed spectral ratio in the prior art is broken through. The color temperature-spectrum two-dimensional regulation and control module can switch corresponding spectrum combinations in real time according to specific light requirements of different growth stages such as a plant germination stage and a picking stage, and the problem that a traditional fixed spectrum cannot be regulated and controlled in a customized mode is solved. The decision engine adopts a'real-time monitoring-deviation correction 'linkage mode, can drive the switching of the six-equal-division circular film within 3 minutes and reduce the illumination intensity, and avoids the problem of spindling and out-of-control caused by lagging of a traditional system. Illumination waste is reduced, and the light utilization efficiency is improved.

Description

TECHNICAL FIELD

[0001] The application belongs to the field of intelligent equipment for facility agriculture, and relates to a multi-spectral intelligent regulation system and method for precise plant cultivation, which is especially suitable for the closed-loop regulation of the light environment of plant growth. BACKGROUND

[0002] With the rapid development of facility agriculture, greenhouse planting of crops has broken the geographical and seasonal limitations of traditional planting due to its strong adaptability to the natural environment, and has become a key mode for improving the economic benefits of agriculture. As a core environmental factor for plant growth and development, light has a direct impact on the photosynthetic efficiency, secondary metabolite accumulation, and morphological development of plants. Therefore, precise regulation of the light environment has become a core technical requirement for facility plant cultivation. The multi-spectral intelligent regulation system can significantly improve quality and yield by dynamically matching the light needs of plant growth, and has important application value in modern greenhouse construction.

[0003] Chinese Patent CN219761999U discloses an angle-adjustable full-spectrum plant growth lamp, which realizes mechanical adjustment of the lighting angle through a rotating lamp head structure and uses full-spectrum LED light sources to simulate solar radiation, thereby improving the adaptability of the light to a certain extent. However, this technology has significant limitations: the angle adjustment relies on pre-set gears and cannot continuously and dynamically adapt to the plant canopy form, and the full-spectrum is fixed in proportion and the specific spectral ratio is not specified. The system lacks a feedback module for the plant growth state, and cannot form a closed loop of "monitoring - decision - regulation".

[0004] In the article "Analysis of Full-spectrum Plant Lighting Lamp Development Technology" by Liu Xue of Shanghai Qiangling Electronics Co., Ltd. in Electronic Production, Vol. 33, No. 4, pp. 99-102, a plant lighting scheme is introduced, which realizes the improvement of spectral stability through the mixing of specific proportion of lamp beads and corrosion-resistant technology. However, the static spectral formula used cannot be dynamically adjusted with plant growth, and lacks real-time sensing and feedback mechanism for plant growth state, making it difficult to cope with the complex and variable light needs in crop cultivation.

[0005] Therefore, it is of great significance to develop a dedicated system that integrates dynamic spectral regulation, multi-dimensional growth feedback, and intelligent decision-making to promote the development of precise plant cultivation technology. SUMMARY

[0006] In view of the defects in the prior art such as fixed spectrum of plant light, single visual monitoring dimension, and lack of plant-specific photoperiod model in decision-making, the present application proposes a multi-spectral intelligent regulation system and method for precise plant cultivation, which realizes precise regulation of "spectrum - space - decision" for the whole growth cycle of plants.

[0007] To achieve the above objectives, the technical solution of the present invention is as follows: A multispectral intelligent control system for precision plant cultivation includes a control system, a six-channel multispectral plant growth control lamp, a binocular camera, and a monocular camera. M of the six-channel multispectral plant growth control lamps are fixed to the top of a greenhouse. The binocular camera is fixed to the top of the greenhouse via a platform with multi-degree-of-freedom rotation adjustment in both horizontal and vertical directions. The binocular camera is installed 1-2 meters above the plant canopy. The binocular cameras are evenly distributed within the greenhouse; specifically, N binocular cameras are deployed using a grid-based method based on the spatial topological characteristics and canopy density of the planting area to achieve systematic sampling of the entire planting area. L monocular cameras are deployed at eye level, facing the plants horizontally, and are installed inside the greenhouse via brackets. The monocular camera's shooting angle is aimed at the plant group to achieve complete capture of the overall plant morphology. The six-channel multispectral plant growth control lamp, binocular camera, and monocular camera are all connected to the control system via data cables. The supplemental lighting range of the six-channel multispectral plant growth control lamp covers all plants, the binocular camera acquires images of the plant canopy, and the monocular camera acquires images of the entire plant.

[0008] Furthermore, the six-channel multispectral plant growth control lamp includes a lampshade frame, a light-transmitting lampshade, a hanging point, a film changing mechanism, and a light source position adjustment mechanism; the six-channel multispectral plant growth control lamp has an overall cylindrical structure, and the light-transmitting lampshade surrounds the lampshade frame; The lampshade frame includes a base, columns, and three horizontal frames. There are four columns, which are evenly distributed around the perimeter of the base and are fixedly connected to the base, the lower horizontal frame, the middle horizontal frame, and the upper horizontal frame from bottom to top. A suspension point is provided at the top of each column. There are four light source position adjustment mechanisms in total. These four mechanisms have identical structures and are evenly distributed on the upper horizontal frame on the same side in the same direction of rotation. Each light source position adjustment mechanism has a cover plate above it. The light source position adjustment mechanism at the front left corner includes a servo pad, servo A, servo B, servo C, a steering plate, a moving frame, and a light source. The steering plate is hinged to the upper surface of the upper horizontal frame, with the hinge position located on the front left side of the upper horizontal frame. The moving frame is fixed to the upper surface of the steering plate, and the light source is installed in the center of the lower surface of the moving frame for heat dissipation. The plate is installed in the center of the upper surface of the moving frame; servo A is fixed to the lower surface of the upper horizontal frame through servo pad A, and is fixedly connected to the front side of the moving frame through the moving swing arm A, controlling the movement of the moving frame along the X-axis; servo B is fixed to the lower surface of the upper horizontal frame through servo pad B, and is fixedly connected to the left side of the moving frame through the moving swing arm B, controlling the movement of the moving frame along the Y-axis; servo bracket C is fixed to the upper surface of the middle horizontal frame and fixedly connected to servo C, and servo C is fixedly connected to the steering plate through the steering swing arm, realizing the steering plate's steering movement; The film-changing mechanism includes a servo motor D, a servo motor bracket D, a light-blocking cloth, a drive gear, a driven gear, and a six-part circular film. The servo motor bracket D is fixed to the upper surface of the lower horizontal frame and is fixedly connected to the servo motor D. The servo motor D is connected to the drive gear via a key. The driven gear is installed on the upper surface of the lower horizontal frame and meshes with the drive gear. The inner ring of the driven gear has a slot to hold the six-part circular film. Each 1 / 6 sector area of ​​the six-part circular film is provided with a film of a different spectrum. The lower surface of the lower horizontal frame is provided with a light-blocking cloth of the same size as the six-part circular film. The light-blocking cloth has two opposing 1 / 6 light-transmitting sector openings.

[0009] The working principle of the six-channel multispectral plant growth control lamp is as follows: When the control system issues a light source movement command, servo motor A starts and drives the moving arm A to rotate, which in turn drives the moving frame connected to it to move along the X-axis; servo motor B starts and drives the moving arm B to rotate, which in turn drives the moving frame connected to it to move along the Y-axis; when the control system issues a direction adjustment command, servo motor C starts and drives the direction adjustment arm to rotate, and the direction adjustment plate connected to it rotates around the hinge axis connecting itself to the lamp cover frame. The two commands together realize the adjustment of the illumination angle and range. When the control system issues a film replacement command based on the plant's growth status and specifies the target optical film type, such as red-yellow type A or red-yellow type B, the servo motor D in the film replacement mechanism is activated. The active gear drives the driven gear to rotate 60°, and the six-part circular film rotates synchronously with the driven gear, so that the two target films are located at the light-transmitting openings of the light-blocking cloth. The two spectra of light can pass through the light-transmitting openings to complete the spectral switching.

[0010] The color temperature range of the six-part circular films is 1800K±100K; the peak wavelength combinations are as follows: Type A (Red-Yellow): 620nm-630nm red light and 570nm-590nm yellow light; Type B (Red-Yellow): 660nm-680nm deep red light and 590nm-610nm orange light; Red-Yellow Type C: 690nm-750nm far-red light; A multispectral intelligent regulation method for precision plant cultivation, utilizing a multispectral intelligent regulation system for precision plant cultivation, includes the following steps: A. Data acquisition by binocular cameras N binocular cameras were evenly arranged in the greenhouse to systematically sample the plant canopy. The binocular cameras were fixed vertically downwards at the top of the greenhouse, 1-2 meters above the plant canopy, to ensure that the field of view completely covered the target area. A light shield was used to reduce stray light interference. The white balance of 5500K, exposure time and low ISO were manually optimized and fixed. The infrared emitter was turned off to avoid interfering with RGB imaging. Environmental parameters were recorded during synchronous acquisition and compared with a standard color chart for subsequent calibration. A binocular camera measures the distance between a plant and the ground by taking two photos with two lenses. That is, it takes two images of the same plant area with two lenses, calculates the phase difference between corresponding pixels in the two images, and then combines the camera's own parameters to calculate the actual physical distance from each point on the plant canopy surface to the ground through measurement. B. Monitoring plant height using a monocular camera Inside the greenhouse, based on the spatial distribution characteristics of the plant rows, L monocular cameras are configured horizontally. The monocular cameras perform distributed sampling of the plant population through multi-point visual coverage at a level view angle. The plant height is quantitatively detected using a machine vision-based proportional analysis method: the vertical edge of the calibration rod is identified by Hough transform and its pixel height is calculated. The outline of the main stem of the plant is extracted by color threshold segmentation. Finally, the absolute biological height of the plant is derived based on the ratio between the actual height of the calibration rod and the pixel height. The measurement accuracy is controlled within ±1cm. Data is collected hourly to calibrate the growth rate curve.

[0011] C. Processing and chlorophyll analysis of images acquired by the monocular camera Image preprocessing and white balance correction are performed. The corrected pixel values ​​(red, green, and blue channels) are then used for vegetation segmentation using the greenness index and Otsu thresholding method to obtain the greenness index image. The formula is as follows: (1) in, ExG(x, y) For the green index image, For pixel values, the red channel. For pixel values, the green channel, Pixel value blue channel; A binary mask is generated by applying the Otsu thresholding method to the green index image. The optimal threshold, automatically determined by the Otsu thresholding algorithm, is used in vegetated areas. Calculate the following two representative indices: Normalized Green-Red Difference Index : (2) Green leaf index : (3) in, Used to avoid division by zero errors; The two-dimensional features of the overgreen index image for each vegetation type are transformed into a single quantity, and the arithmetic mean of the vegetation area is extracted: (4) in, The total number of pixels in the vegetation area. , , The extracted arithmetic mean of the vegetation area As variables, the following multiple linear regression model is established: (5) in, The intercept is... For regression coefficients, This represents the normalized green-red difference index relative to the chlorophyll content. The weight and direction of influence; Indicates the green index to The influence weight and direction of the value; Indicates the green leaf index to The influence weights and directions of the values ​​are determined; they are estimated using the least squares method, and the full graph is calculated. The statistics for the predicted values ​​are as follows: (6) Define low chlorophyll regions: (7) D. Construct a dual-dimensional decision engine based on spectrum and color temperature. The spectrum-color temperature dual-dimensional decision engine is the core unit connecting plant growth status monitoring and light environment regulation.

[0012] During the monitoring phase: First, the monitoring data is standardized, and the arithmetic mean of the relative chlorophyll content of the plants is obtained by least squares inversion. Based on this, and the optimal chlorophyll reference value for the entire growth cycle of this variety. Normalization calculations were performed to obtain the chlorophyll state index. Quantify whether the current chlorophyll level is suitable for photosynthetic requirements; combine the actual plant height measured by a monocular camera based on Hough transform and proportional analysis. H Plant height benchmark value in the standard growth model of the same variety during the same period Comparison, calculation of plant height growth index, To determine whether the rate of plant morphogenesis conforms to a normal rhythm; extract leaves. RGB In the image G The actual average brightness of the component Compared with the baseline value of G component in healthy leaves of the same variety After normalization, the leaf vigor index was obtained. This reflects the immediate metabolic activity of the plant, as observed in the leaves collected above. RGB Extracting from images G The process of calculating the actual average brightness of the components and normalizing it transforms discrete monitoring data into unified quantitative decision indicators.

[0013] During the control execution phase, the control system generates control commands containing the target optical film type and light intensity parameters. These commands are transmitted to the six-channel multispectral plant growth control lamp, driving the servo motor D to rotate and engage the drive and driven gears, causing the target film to rotate to the working position. The entire switching process has a response time of less than 3 seconds. Simultaneously, the moving frame slides along the X and Y axes to adjust the lateral coverage of the beam, ensuring a perfect match between the spectral projection area and the plant canopy morphology. After control is complete, the binocular camera re-acquires images of the plant canopy, plant height, and leaves within 3 minutes. G For component data, repeat the aforementioned data processing flow to calculate new... , , The value is compared with the deviation of the actual indicator and the target indicator. If the deviation is less than 5%, the current strategy is maintained. If the deviation reaches or exceeds 5%, data standardization is restarted. Based on the three unified and quantified plant growth status indicators of chlorophyll status index, plant height growth index and leaf vitality index obtained by recalculation, the appropriate optical film type and light intensity control parameters are matched and the instruction execution process is carried out until the deviation falls within the allowable range, so as to continuously ensure the dynamic adaptation of light environment and plant growth needs.

[0014] Compared with the prior art, the present invention has the following beneficial effects: 1. This invention utilizes a multispectral plant growth control lamp equipped with six optical films possessing specific spectral characteristics, coupled with a servo-driven light source position adjustment mechanism, overcoming the limitations of fixed spectral ratios in existing technologies. Its dual-dimensional "color temperature-spectrum" control module can switch the corresponding spectral combination in real time according to the specific light requirements of different growth stages such as plant budding and harvesting, solving the problem of traditional fixed-spectrum uncustomizable control.

[0015] 2. The decision engine of this invention adopts a "real-time monitoring-deviation correction" linkage mode: the multi-view vision network collects plant height data every 30 minutes and updates the relative chlorophyll content hourly (through...). ExG , NGRDI (Algorithm inversion), after data is transmitted to the engine, if the control deviation is ≥5%, a secondary decision is immediately initiated, reducing the control lag time by 60% compared to traditional feedback-free systems. For example, when excessive plant height is detected... When the light intensity is greater than 1.2, the engine can drive the switching of the red and yellow C-type film and reduce the light intensity within 3 minutes, avoiding the problem of uncontrolled growth caused by the lag in traditional systems.

[0016] 3. This invention uses a high-strength aluminum alloy lampshade frame as the core of its six-channel mechanical structure. Components such as a moving frame and a moving swing arm allow for adjustment of the beam angle and lateral position. Precision slots ensure stable positioning during film switching, guaranteeing precise light coverage of the plant canopy. Compared to traditional devices that rely on preset angle adjustments, this invention can continuously and dynamically adapt to the plant canopy morphology, reducing light waste and improving light utilization efficiency. Attached Figure Description

[0017] Figure 1 is a schematic diagram of the structure of a six-channel multispectral plant growth control lamp.

[0018] Figure 2 for Figure 1 AA top view.

[0019] Figure 3 for Figure 1 A top view of the BB.

[0020] Figure 4 for Figure 1 CC top view.

[0021] Figure 5 This is a schematic diagram of the appearance of a six-channel multispectral plant growth regulator lamp.

[0022] Figure 6 This is a schematic diagram of a monocular camera used to detect plant height.

[0023] Figure 7 This is a flowchart of the method of the present invention.

[0024] Figure 8 This is a schematic diagram of a six-part circular thin film structure.

[0025] Figure 9 This is a schematic diagram of the structure of the blackout fabric.

[0026] Figure 10 for Figure 1 One of the isometric views (without cover plate).

[0027] Figure 11 for Figure 1 Axonometric drawing 2 (without cover plate).

[0028] In the diagram: 1. Cover plate, 2. Light-transmitting lampshade, 3. Suspension point, 4. Servo pad, 5. Steering plate, 6. Servo bracket B, 7. Heat sink, 8. Moving frame, 9. Moving swing arm, 10. Servo C, 11. Servo D, 12. Drive gear, 13. Servo B, 14. Driven gear, 15. Servo A, 16. Steering swing arm, 17. Lampshade frame. Detailed Implementation

[0029] The invention will now be further described with reference to the accompanying drawings.

[0030] like Figures 1-11 As shown, a multispectral intelligent control system for precise plant cultivation includes a control system, a six-channel multispectral plant growth control lamp, a binocular camera, and a monocular camera. M of the six-channel multispectral plant growth control lamps are fixed to the top of a greenhouse. The binocular camera is fixed to the top of the greenhouse via a platform with multi-degree-of-freedom rotation adjustment in both horizontal and vertical directions. The binocular camera is installed 1-2 meters above the plant canopy. The binocular cameras are evenly distributed within the greenhouse; specifically, N binocular cameras are deployed using a grid-based method based on the spatial topological characteristics and canopy distribution density of the planting area, enabling systematic sampling of the entire planting area. L monocular cameras are deployed at eye level, facing the plants horizontally, and are installed inside the greenhouse via brackets. The monocular camera's shooting angle is aimed at the plant group, achieving complete capture of the overall plant morphology. The six-channel multispectral plant growth control lamp, binocular camera, and monocular camera are all connected to the control system via data cables. The supplemental lighting range of the six-channel multispectral plant growth control lamp covers all plants, the binocular camera acquires images of the plant canopy, and the monocular camera acquires images of the entire plant.

[0031] Furthermore, the six-channel multispectral plant growth control lamp includes a lampshade frame 17, a light-transmitting lampshade 2, a hanging point 3, a film changing mechanism, and a light source position adjustment mechanism; the six-channel multispectral plant growth control lamp has an overall cylindrical structure, and the light-transmitting lampshade 2 surrounds the lampshade frame 17. The lampshade frame 17 includes a base, columns, and three horizontal frames. There are four columns, which are evenly distributed around the perimeter of the base and are fixedly connected to the base, the lower horizontal frame, the middle horizontal frame, and the upper horizontal frame from bottom to top. A hanging point 3 is provided at the top of the column. There are four light source position adjustment mechanisms in total. These four mechanisms have identical structures and are evenly distributed on the upper horizontal frame on the same side in the same direction of rotation. Each light source position adjustment mechanism has a cover plate 1 above it. The light source position adjustment mechanism at the front left corner includes a servo pad 4, servo A15, servo B13, servo C10, a steering plate 5, a moving frame 8, and a light source. The steering plate 5 is hinged to the upper surface of the upper horizontal frame, with the hinge position located on the front left side of the upper horizontal frame. The moving frame 8 is fixed to the upper surface of the steering plate 5, and the light source is installed in the center of the lower surface of the moving frame 8. A heat sink 7 is installed... Mounted at the center of the upper surface of the movable frame 8; Servo A15 is fixed to the lower surface of the upper horizontal frame via servo pad A, and is fixedly connected to the front side of the movable frame 8 via movable swing arm A, controlling the movable frame 8 to move along the X-axis; Servo B13 is fixed to the lower surface of the upper horizontal frame via servo pad B4, and is fixedly connected to the left side of the movable frame 8 via movable swing arm B9, controlling the movable frame 8 to move along the Y-axis; Servo bracket C is fixed to the upper surface of the middle horizontal frame and is fixedly connected to servo C10, and servo C10 is fixedly connected to the steering plate 5 via steering swing arm 16, realizing the steering movement of steering plate 5; The film-changing mechanism includes a servo motor D11, a servo motor bracket D, a light-shielding cloth, a drive gear 12, a driven gear 14, and a six-part circular film. The servo motor bracket D is fixed to the upper surface of the lower horizontal frame and is fixedly connected to the servo motor D11. The servo motor D11 is connected to the drive gear 12 via a key. The driven gear 14 is installed on the upper surface of the lower horizontal frame and meshes with the drive gear 12. The inner ring of the driven gear 14 has a slot to hold the six-part circular film. Each 1 / 6 sector area of ​​the six-part circular film is provided with a film of a different spectrum. The lower surface of the lower horizontal frame is provided with a light-shielding cloth of the same size as the six-part circular film. The light-shielding cloth has two opposing 1 / 6 light-transmitting sector openings. The working principle of the six-channel multispectral plant growth control lamp is as follows: When the control system issues a light source movement command, servo motor A15 starts and drives the moving arm A to rotate, which in turn drives the moving frame 8 connected to it to move along the X-axis; servo motor B13 starts and drives the moving arm B9 to rotate, which in turn drives the moving frame 8 connected to it to move along the Y-axis; when the control system issues a direction adjustment command, servo motor C10 starts and drives the direction adjustment arm 16 to rotate, which in turn drives the direction adjustment plate 5 connected to it to rotate around the hinge axis that connects itself to the lamp cover frame 17. The two commands together realize the adjustment of the illumination angle and range. When the control system issues a film replacement command based on the plant's growth status and specifies the target optical film type, such as red-yellow type A or red-yellow type B, the servo motor D11 in the film replacement mechanism is activated, and the drive gear 12 drives the driven gear 14 to rotate 60°. The six-part circular film rotates synchronously with the driven gear 14, so that the two target films are located at the light-transmitting openings of the light-blocking cloth. The two spectra of light can pass through the light-transmitting openings to complete the spectral switching. The color temperature range of the six-part circular films is 1800K±100K; the peak wavelength combinations are as follows: Type A (Red-Yellow): 620nm-630nm red light and 570nm-590nm yellow light; Type B (Red-Yellow): 660nm-680nm deep red light and 590nm-610nm orange light; Red-Yellow Type C: 690nm-750nm far-red light.

[0032] A multispectral intelligent regulation method for precision plant cultivation, utilizing a multispectral intelligent regulation system for precision plant cultivation, includes the following steps: A. Data acquisition by binocular cameras N binocular cameras were evenly arranged in the greenhouse to systematically sample the plant canopy. The binocular cameras were fixed vertically downwards at the top of the greenhouse, 1-2 meters above the plant canopy, to ensure that the field of view completely covered the target area. A light shield was used to reduce stray light interference. The white balance of 5500K, exposure time and low ISO were manually optimized and fixed. The infrared emitter was turned off to avoid interfering with RGB imaging. Environmental parameters were recorded during synchronous acquisition and compared with a standard color chart for subsequent calibration. A binocular camera measures the distance between a plant and the ground by taking two photos with two lenses. That is, it takes two images of the same plant area with two lenses, calculates the phase difference between corresponding pixels in the two images, and then combines the camera's own parameters to calculate the actual physical distance from each point on the plant canopy surface to the ground through measurement. B. Monitoring plant height using a monocular camera Inside the greenhouse, based on the spatial distribution characteristics of the plant rows, L monocular cameras are configured horizontally. The monocular cameras perform distributed sampling of the plant population through multi-point visual coverage at a level view angle. The plant height is quantitatively detected using a machine vision-based proportional analysis method: the vertical edge of the calibration rod is identified by Hough transform and its pixel height is calculated. The outline of the main stem of the plant is extracted by color threshold segmentation. Finally, the absolute biological height of the plant is derived based on the ratio between the actual height of the calibration rod and the pixel height. The measurement accuracy is controlled within ±1cm. Data is collected hourly to calibrate the growth rate curve. C. Processing and chlorophyll analysis of images acquired by the monocular camera Image preprocessing and white balance correction are performed. The corrected pixel values ​​(red, green, and blue channels) are then used for vegetation segmentation using the greenness index and Otsu thresholding method to obtain the greenness index image. The formula is as follows: (1) in, ExG(x, y) For the green index image, For pixel values, the red channel. For pixel values, the green channel, Pixel value blue channel; A binary mask is generated by applying the Otsu thresholding method to the green index image. The optimal threshold, automatically determined by the Otsu thresholding algorithm, is used in vegetated areas. Calculate the following two representative indices: Normalized Green-Red Difference Index : (2) Green leaf index : (3) in, Used to avoid division by zero errors; The two-dimensional features of the overgreen index image for each vegetation type are transformed into a single quantity, and the arithmetic mean of the vegetation area is extracted: (4) in, The total number of pixels in the vegetation area. , , The extracted arithmetic mean of the vegetation area As variables, the following multiple linear regression model is established: (5) in, The intercept is... For regression coefficients, This represents the normalized green-red difference index relative to the chlorophyll content. The weight and direction of influence; Indicates the green index to The influence weight and direction of the value; Indicates the green leaf index to The influence weights and directions of the values ​​are determined; they are estimated using the least squares method, and the full graph is calculated. The statistics for the predicted values ​​are as follows: (6) Define low chlorophyll regions: (7) D. Construct a dual-dimensional decision engine based on spectrum and color temperature. The spectrum-color temperature dual-dimensional decision engine is the core unit connecting plant growth status monitoring and light environment regulation; During the monitoring phase: First, the monitoring data is standardized, and the arithmetic mean of the relative chlorophyll content of the plants is obtained by least squares inversion. Based on this, and the optimal chlorophyll reference value for the entire growth cycle of this variety. Normalization calculations were performed to obtain the chlorophyll state index. Quantify whether the current chlorophyll level is suitable for photosynthetic requirements; combine the actual plant height measured by a monocular camera based on Hough transform and proportional analysis. H Plant height benchmark value in the standard growth model of the same variety during the same period Comparison, calculation of plant height growth index, To determine whether the rate of plant morphogenesis conforms to a normal rhythm; extract leaves. RGB In the image G The actual average brightness of the component Compared with the baseline value of G component in healthy leaves of the same variety After normalization, the leaf vigor index was obtained. This reflects the immediate metabolic activity of the plant, as observed in the leaves collected above. RGB Extracting from images G The process of calculating the actual average brightness of the components and normalizing it transforms discrete monitoring data into unified quantitative decision indicators. During the control execution phase, the control system generates control commands containing the target optical film type and light intensity parameters. After being transmitted to the six-channel multispectral plant growth control lamp, the servo motor D11 is driven to rotate, which in turn drives the active gear 12 and the driven gear 14 to mesh and rotate the target film to the working position. The entire switching process has a response time of less than 3 seconds. At the same time, the moving frame 8 slides along the X and Y axes to adjust the lateral coverage of the beam, ensuring that the spectral projection area is completely matched with the plant canopy morphology. After the control is completed, the binocular camera re-acquires images of the plant canopy, plant height, and leaves within 3 minutes. G For component data, repeat the aforementioned data processing flow to calculate new... , , The value is compared with the deviation of the actual indicator and the target indicator. If the deviation is less than 5%, the current strategy is maintained. If the deviation reaches or exceeds 5%, data standardization is restarted. Based on the three unified and quantified plant growth status indicators of chlorophyll status index, plant height growth index and leaf vitality index obtained by recalculation, the appropriate optical film type and light intensity control parameters are matched and the instruction execution process is carried out until the deviation falls within the allowable range, so as to continuously ensure the dynamic adaptation of light environment and plant growth needs.

[0033] In the light source control stage of this invention, four light source position adjustment mechanisms serve as core execution components. When the control system issues commands to adjust the illumination angle and range, servo motors A 15 and B 13, fixed to the lower surface of the upper horizontal frame, drive the rotating arm 9 to rotate, causing the movable frame 8, which is hinged to it, to translate along the X and Y axes. Simultaneously, servo motor C10 on the middle horizontal frame drives the directional arm to rotate, causing the directional plate, which is hinged to the upper horizontal frame, to rotate around the hinge axis. The two work together to achieve dynamic adjustment of the illumination angle and coverage range of the light source. In the film replacement process, after receiving commands from the control system, servo motor D 11 of the film replacement mechanism drives the six-division circular film fixed in the inner groove of the driven gear 14 to rotate synchronously by 60° or 120° through the drive gear 12. This allows the target spectral film to be precisely aligned with the light-transmitting opening of the light-blocking cloth on the lower surface of the lower horizontal frame, completing the spectral switching and achieving dynamic spectral control adapted to the plant's growth state.

[0034] The binocular camera is vertically fixed 1-2 meters above the canopy at the top of the greenhouse, equipped with a shade cover, white balance set to 5500K, and infrared emitter turned off; the monocular camera is deployed along both sides of the greenhouse, parallel to the planting rows, for monitoring plant height. The spectrum-color temperature dual-dimensional decision engine is deployed on the greenhouse control center server and is connected to the control unit of the mechanical structure and the image acquisition terminal of the visual monitoring network via wired communication. Through the above standardization process, the originally scattered monitoring data is transformed into normalized indicators, providing a unified quantitative basis for subsequent decision-making. The workflow of this invention is as follows: First, growth status data acquisition and preprocessing: The top binocular camera collects data on the plant canopy every hour. RGB Image, through ExG The index and Otsu threshold method were used to segment vegetation areas and calculate... ExG , NGRDI , GLI The index, substituted into the multiple linear regression model, yields... SPAD Predicted values, and then the chlorophyll state index is calculated. The monocular camera acquires images every three hours, identifies a 100cm calibration pole using Hough transform, and calculates the plant height growth index. Simultaneously, the average brightness of the G component in the RGB image of the leaf is extracted to obtain the leaf vitality index. All data is transmitted to the decision engine in real time.

[0035] Secondly, the decision matching is based on a two-dimensional spectrum and color temperature: the decision engine calls upon the spectral requirements library for different plant growth stages and performs threshold comparisons on standardized indicators, as shown in the table below: Matching table of multi-indicator regulation strategies for plant light environment

[0036] This table focuses on multi-indicator input and control strategy output, presenting complex plant growth states: chlorophyll status index. Plant height growth index Leaf vigor index The light environment control parameters are standardized and mapped to optical thin films and light intensity; the engine collects data in real time. , , The data can be directly matched with the "indicator range" in the table, and the corresponding optical thin film type and illumination parameters can be quickly retrieved to realize the automated control of real-time monitoring → table lookup and matching → strategy output. Next, dynamic control is executed: the decision engine generates control commands, which drive the servo B 13 to rotate through the servo controller, causing the drive gear 12 and driven gear 14 to rotate, so that the target film rotates 60° to the light-transmitting opening of the light-shielding cloth. The moving frame 8 adjusts the lateral coverage range of the beam along the X and Y axes, and the moving swing arm 9 adjusts the irradiation angle to ensure the stability of the film's optical performance.

[0037] Finally, closed-loop feedback optimization: Three minutes after the regulation is completed, the visual monitoring network re-collects data and calculates the deviation between the actual and target indicators. If the deviation is less than 5%, the current regulation strategy is maintained; if the deviation is ≥5%, the decision engine restarts the above decision-making process, adjusting the film type or light parameters until the deviation is within the allowable range, thus achieving precise regulation of the light environment throughout the plant's entire growth cycle.

[0038] As described above, those skilled in the art can make various other corresponding changes and modifications based on the technical solutions and concepts of this invention, and all such changes and modifications should fall within the protection scope of the claims of this invention.

Claims

1. A multispectral intelligent control system for precision plant cultivation, characterized in that: The system includes a control system, six-channel multispectral plant growth control lights, binocular cameras, and monocular cameras. There are M six-channel multispectral plant growth control lights, fixed to the top of the greenhouse. The binocular cameras are fixed to the top of the greenhouse via a platform with multi-degree-of-freedom rotation adjustment in both horizontal and vertical directions. The binocular cameras are installed 1-2 meters above the plant canopy. The binocular cameras are evenly distributed within the greenhouse, with N binocular cameras deployed using a grid-based method based on the spatial topological characteristics and canopy density of the planting area, enabling systematic sampling of the entire planting area. There are L monocular cameras, all deployed at eye level, horizontally facing the plants, and mounted inside the greenhouse via brackets. The monocular cameras' shooting angle is aimed at the plant group, achieving complete capture of the overall plant morphology. The six-channel multispectral plant growth control lights, binocular cameras, and monocular cameras are all connected to the control system via data cables. The supplemental lighting range of the six-channel multispectral plant growth control lamp covers all plants, the binocular camera acquires images of the plant canopy, and the monocular camera acquires images of the entire plant.

2. The multispectral intelligent control system for precision plant cultivation according to claim 1, characterized in that: The six-channel multispectral plant growth control lamp includes a lampshade frame (17), a light-transmitting lampshade (2), a hanging point (3), a film changing mechanism, and a light source position adjustment mechanism; the six-channel multispectral plant growth control lamp is a cylindrical structure, and the light-transmitting lampshade (2) surrounds the lampshade frame (17). The lampshade frame (17) includes a base, columns and three horizontal frames. There are four columns, which are evenly distributed around the base and are fixedly connected to the base, the lower horizontal frame, the middle horizontal frame and the upper horizontal frame from bottom to top. The top of the column is provided with a hanging point (3). There are four light source position adjustment mechanisms in total. The four light source position adjustment mechanisms have the same structure and are evenly distributed on the upper horizontal frame on the same side in the same rotation direction. Each light source position adjustment mechanism is equipped with a cover plate (1). The light source position adjustment mechanism at the left front corner includes a servo pad (4), servo A (15), servo B (13), servo C (10), a steering plate (5), a moving frame (8), and a light source. The steering plate (5) is hinged to the upper surface of the upper horizontal frame, and the hinge position is located on the left front side of the upper horizontal frame. The moving frame (8) is fixed to the upper surface of the steering plate (5), and the light source is installed in the center of the lower surface of the moving frame (8). The heat sink (7) is installed on the upper surface of the moving frame (8). The center of the upper surface of the movable frame (8); servo A (15) is fixed to the lower surface of the upper horizontal frame through servo pad A, and is fixedly connected to the front side of the movable frame (8) through the movable swing arm A, controlling the movable frame (8) to move along the X-axis; servo B (13) is fixed to the lower surface of the upper horizontal frame through servo pad B (4), and is fixedly connected to the left side of the movable frame (8) through the movable swing arm B (9), controlling the movable frame (8) to move along the Y-axis; servo bracket C is fixed to the upper surface of the middle horizontal frame and is fixedly connected to servo C (10), servo C (10) is fixedly connected to the steering plate (5) through the steering swing arm (16), realizing the steering movement of the steering plate (5); The film-changing mechanism includes a servo motor D (11), a servo motor bracket D, a light-shielding cloth, a drive gear (12), a driven gear (14), and a six-part circular film. The servo motor bracket D is fixed to the upper surface of the lower horizontal frame and is fixedly connected to the servo motor D (11). The servo motor D (11) is connected to the drive gear (12) by a key. The driven gear (14) is installed on the upper surface of the lower horizontal frame and meshes with the drive gear (12). The inner ring of the driven gear (14) is provided with a slot to hold the six-part circular film. Each 1 / 6 sector area of ​​the six-part circular film is provided with a film of a different spectrum. The lower surface of the lower horizontal frame is provided with a light-shielding cloth of the same size as the six-part circular film. The light-shielding cloth is provided with two opposite 1 / 6 light-transmitting sector openings. The working principle of the six-channel multispectral plant growth control lamp is as follows: When the control system issues a light source movement command, servo motor A (15) starts and drives the moving arm A to rotate, which in turn drives the moving frame (8) connected to it to move along the X-axis; servo motor B (13) starts and drives the moving arm B (9) to rotate, which in turn drives the moving frame (8) connected to it to move along the Y-axis; when the control system issues a direction adjustment command, servo motor C (10) starts and drives the direction adjustment arm (16) to rotate, and the direction adjustment plate (5) connected to it rotates around the hinge axis connected to itself and the lamp cover frame (17). The two commands together realize the adjustment of the illumination angle and range. When the control system issues a film replacement command based on the plant's growth status and specifies the target optical film type, such as red-yellow type A or red-yellow type B, the servo motor D (11) in the film replacement mechanism is activated, and the active gear (12) drives the driven gear (14) to rotate 60°. The six-part circular film rotates synchronously with the driven gear (14), so that the two target films are located at the light-transmitting openings of the light-blocking cloth. The light of the two spectra can be transmitted through the light-transmitting openings to complete the spectrum switching. The color temperature range of the six-part circular films is 1800K±100K; the peak wavelength combinations are as follows: Type A (Red-Yellow): 620nm-630nm red light and 570nm-590nm yellow light; Type B (Red-Yellow): 660nm-680nm deep red light and 590nm-610nm orange light; Red-Yellow Type C: 690nm-750nm far-red light.

3. A multispectral intelligent regulation method for precise plant cultivation, utilizing the multispectral intelligent regulation system for precise plant cultivation as described in claim 1 or 2, comprising the following steps: A. Data acquisition by binocular cameras N binocular cameras were evenly arranged in the greenhouse to systematically sample the plant canopy. The binocular cameras were fixed vertically downwards at the top of the greenhouse, 1-2 meters above the plant canopy, to ensure that the field of view completely covered the target area. A light shield was used to reduce stray light interference. The white balance of 5500K, exposure time and low ISO were manually optimized and fixed. The infrared emitter was turned off to avoid interfering with RGB imaging. Environmental parameters were recorded during synchronous acquisition and compared with a standard colorimeter for subsequent calibration. A binocular camera measures the distance between a plant and the ground by taking two photos with two lenses. That is, it takes two images of the same plant area with two lenses, calculates the phase difference between corresponding pixels in the two images, and then combines the camera's own parameters to calculate the actual physical distance from each point on the plant canopy surface to the ground through measurement. B. Monitoring plant height using a monocular camera Inside the greenhouse, based on the spatial distribution characteristics of the plant rows, L monocular cameras are configured horizontally. The monocular cameras perform distributed sampling of the plant population through multi-point visual coverage at a level view angle. The plant height is quantitatively detected using a machine vision-based proportional analysis method: the vertical edge of the calibration rod is identified by Hough transform and its pixel height is calculated. The outline of the main stem of the plant is extracted by color threshold segmentation. Finally, the absolute biological height of the plant is derived based on the ratio between the actual height of the calibration rod and the pixel height. The measurement accuracy is controlled within ±1cm. Data is collected hourly to calibrate the growth rate curve. C. Processing and chlorophyll analysis of images acquired by the monocular camera. Image preprocessing and white balance correction are performed. The corrected pixel values ​​(red, green, and blue channels) are then used for vegetation segmentation using the greenness index and Otsu thresholding method to obtain the greenness index image. The formula is as follows: (1) in, ExG(x,y) For the green index image, For pixel values, the red channel. For pixel values, the green channel, Pixel value blue channel; A binary mask is generated by applying the Otsu thresholding method to the green index image. The optimal threshold, automatically determined by the Otsu thresholding algorithm, is used in vegetated areas. Calculate the following two representative indices: Normalized Green-Red Difference Index : (2) Green leaf index : (3) in, Used to avoid division by zero errors; The two-dimensional features of the overgreen index image for each vegetation type are transformed into a single quantity, and the arithmetic mean of the vegetation area is extracted: (4) in, The total number of pixels in the vegetation area. , , The extracted arithmetic mean of the vegetation area As variables, the following multiple linear regression model is established: (5) in, The intercept is... For regression coefficients, This represents the normalized green-red difference index relative to the chlorophyll content. The weight and direction of influence; Indicates the green index to The influence weight and direction of the value; Indicates the green leaf index to The influence weights and directions of the values ​​are determined; they are estimated using the least squares method, and the full graph is calculated. The statistics for the predicted values ​​are as follows: (6) Define low chlorophyll regions: (7) D. Construct a dual-dimensional decision engine based on spectrum and color temperature. The spectrum-color temperature dual-dimensional decision engine is the core unit connecting plant growth status monitoring and light environment regulation; During the monitoring phase: First, the monitoring data is standardized, and the arithmetic mean of the relative chlorophyll content of the plants is obtained by least squares inversion. Based on this, and the optimal chlorophyll reference value for the entire growth cycle of this variety. Normalization calculations were performed to obtain the chlorophyll state index. Quantify whether the current chlorophyll level is suitable for photosynthetic requirements; combine the actual plant height measured by a monocular camera based on Hough transform and proportional analysis. H Plant height benchmark value in the standard growth model of the same variety during the same period Comparison, calculation of plant height growth index, To determine whether the rate of plant morphogenesis conforms to a normal rhythm; extract leaves. RGB In the image G The actual average brightness of the component Compared with the baseline value of G component in healthy leaves of the same variety After normalization, the leaf vigor index was obtained. This reflects the immediate metabolic activity of the plant, as observed in the leaves collected above. RGB Extracting from images G The process of calculating the actual average brightness of the components and normalizing it transforms discrete monitoring data into unified quantitative decision indicators. During the control execution phase, the control system generates control commands containing the target optical film type and light intensity parameters. After transmitting the commands to the six-channel multispectral plant growth control lamp, the servo motor D (11) is driven to operate and drive the active gear (12) and driven gear (14) to mesh and rotate the target film to the working position. The entire switching process has a response time of less than 3 seconds. At the same time, the moving frame (8) slides along the X and Y axes to adjust the lateral coverage of the beam, ensuring that the spectral projection area is completely matched with the plant canopy morphology. After the control is completed, the binocular camera re-acquires plant canopy images, plant height, and leaves within 3 minutes. G For component data, repeat the aforementioned data processing flow to calculate new... , , The value is compared with the deviation of the actual indicator and the target indicator. If the deviation is less than 5%, the current strategy is maintained. If the deviation reaches or exceeds 5%, data standardization is restarted. Based on the three unified and quantified plant growth status indicators of chlorophyll status index, plant height growth index and leaf vitality index obtained by recalculation, the appropriate optical film type and light intensity control parameters are matched and the instruction execution process is carried out until the deviation falls within the allowable range, so as to continuously ensure the dynamic adaptation of light environment and plant growth needs.

Images