Groove type photo-thermal sun position double-closed-loop real-time tracking control system and method based on machine vision

Abstract

The invention provides a groove type photo-thermal sun position double-closed-loop real-time tracking control system and method based on machine vision, the system is composed of a sun sensor, a camera, a development board, a plastic rod, a receiving panel, a heat collecting groove and a driving motor, and double-closed-loop real-time automatic tracking control is carried out in a mode of combining the sun sensor and the camera. According to the method, a sun sensor is used as an outer ring for coarse adjustment, pre-orientation of a tracking system is carried out, fine adjustment is carried out based on a machine vision technology, and a target rotation angle of a heat collection groove is calculated through geometric calibration and perspective transformation, brightness correction and gray level conversion, filtering noise reduction, binarization processing and edge detection. And a fuzzy control algorithm is adopted to control the rotating angle of the heat collecting groove through the driving motor for accurate adjustment, so that the sun is always focused on the heat collecting pipe. The sun position can be accurately identified, the tracking precision is improved, and the system is simple in composition, easy to implement and convenient to popularize and apply.

Description

Technical Field

[0001] The present invention relates to the technical field of solar energy utilization, and in particular to a trough-type solar thermal solar position double closed-loop real-time tracking control system and method based on machine vision. Background Art

[0002] As the global demand for clean energy continues to grow, solar energy, as an inexhaustible renewable energy source, has received widespread attention and application. Trough solar thermal power generation technology is a technology that uses parabolic trough reflectors to focus sunlight onto collector tubes for heating or power generation. Accurately tracking the position of the sun is the key to improving system efficiency.

[0003] Traditionally, there are two ways to track the position of the sun in trough-type solar thermal systems: dual-axis tracking and single-axis tracking. The dual-axis tracking mode is less used in actual engineering applications due to its complex structure, high cost, and the need for a more complex control system to coordinate the movement of the two axes. In the single-axis tracking mode, the open-loop method does not require sensors to provide control feedback signals, has a simple structure and low cost, but has low tracking accuracy and is prone to affecting thermal efficiency.

[0004] The closed-loop method generally uses photoelectric sensors to achieve tracking, such as four-quadrant photoelectric sensors. The advantages of this method are high tracking accuracy, fast response speed, and the ability to adjust in real time according to light conditions. Sensors are easily affected by weather conditions. In rainy, cloudy or hazy weather, the intensity and uniformity of sunlight will change, causing the light signal received by the sensor to become unstable or blurred. Because the sensor outputs a light intensity signal, the signal obtained by the photoelectric sensor needs to be processed complexly before it can be converted into an instruction to control the angle adjustment of the trough-type photothermal device. This not only has certain requirements on the computing power of the control system, but also requires more time and effort in the algorithm design and debugging process. The sensing range of the photoelectric sensor is limited. When the sunlight deviates from the effective sensing range of the sensor, the sensor will not be able to accurately detect the light intensity distribution, thereby affecting the continuity of tracking. However, during long-term use, due to the aging of electronic components, the photodiode in the sensor may experience performance degradation over time, which reduces the accuracy of light intensity detection and affects the long-term stability and tracking accuracy of the system.

[0005] Compared with the prior art, the differences are as follows:

[0006] Technical comparison with patent CN115437410A "A solar panel control system and method"

[0007] The control system of patent CN115437410A includes communication modules, image acquisition modules, etc., emphasizing the interaction with external information and environmental feedback; while this patent uses a dual closed-loop system combining a sun sensor and a camera for tracking control, focusing on the fine adjustment of the inner loop. The two are essentially different in system architecture and functional focus.

[0008] CN115437410A uses weather status information and image acquisition modules for environmental feedback, emphasizing the sun tracking function under different weather conditions; this patent uses geometric calibration and fuzzy control algorithms for high-precision tracking, emphasizing real-time and accuracy. There are obvious differences in the implementation ideas of the control methods of the two.

[0009] CN115437410A mainly focuses on the energy receiving efficiency of solar panels, aiming to optimize deployment; while this patent focuses on the efficient heat collection of trough-type solar thermal systems, aiming to improve the heat collection effect. The two are essentially different in terms of application objects and technical goals.

[0010] Technical comparison with patent CN211293758U "Automatic solar tracking device based on machine vision recognition"

[0011] CN211293758U controls the rotation of solar panels through the ratio of light and shadow lengths, based on light and shadow recognition by visual cameras; while this patent combines machine vision technology with fuzzy control algorithms to perform complex image processing and precise adjustment. The two are essentially different in terms of the complexity of recognition technology and implementation methods.

[0012] CN211293758U focuses on improving solar energy utilization efficiency and controlling costs, and emphasizes anti-interference ability; while this patent ensures high-precision solar position tracking through a double closed-loop control system to improve heat collection efficiency. There are obvious differences between the two in terms of control accuracy and focus.

[0013] The device of CN211293758U has a simple structure and mainly relies on light and shadow as input signals; this patent integrates multiple sensors and control algorithms, focusing on real-time precise control. There is an essential difference between the two in terms of the complexity of system composition and functional realization. Summary of the invention

[0014] In order to solve the above technical problems, the present invention proposes a dual closed-loop real-time tracking control system and method for the trough type solar thermal solar position based on machine vision, which adopts dual feedback combining a solar sensor and a camera to perform closed-loop automatic tracking control to improve the accuracy, real-time performance, stability and reliability of the trough type solar thermal solar position tracking system.

[0015] To achieve the above object, the technical solution adopted by the present invention is:

[0016] The dual closed-loop real-time tracking control system for the solar position of trough-type solar thermal system based on machine vision is composed of a solar sensor based on photodiode, a camera, a development board, an IO module, a plastic rod and a receiving panel, a solar trough and a driving motor;

[0017] The solar sensor, camera, development board, plastic rod and receiving panel are fixed on the top of the heat collecting tank through a bracket. The plastic rod must be arranged parallel to the heat collecting tube of the heat collecting tank and will not be blocked by any nearby objects;

[0018] The receiving panel must be arranged parallel to the opening plane of the heat collecting tank, and the midline of the receiving panel is located in the plane determined by the normal vector of the heat collecting tank and the heat collecting tube. The receiving panel has a rectangular shape and a light-colored outer surface;

[0019] The drive motor adopts a DC motor and includes at least two speed adjustment gears of high speed and low speed. When the inner loop is controlled, the drive motor rotates at high speed, and when the outer loop is controlled, the drive motor rotates at low speed to improve the control accuracy;

[0020] The dual feedback of the sun sensor and the camera is used for closed-loop automatic tracking control. The photodiode sensor is used for rough control to preliminarily determine the position of the sun. The camera captures the projection image of the plastic pole on the receiving panel in real time, and uses machine vision image processing technology to identify the projection contour. The rotation angle of the heat collecting trough is controlled for precise adjustment to ensure that the sun is always focused on the collector tube.

[0021] As a further improvement of the system of the present invention, the development board adopts a Raspberry Pi development board, and the camera is inserted into the CSI interface on the Raspberry Pi mainboard through a flexible cable.

[0022] As a further improvement of the system of the present invention, the development board is externally connected to an AD chip or circuit to access the two voltages output by the solar sensor, converting the analog signal into a digital signal, or the development board communicates with the solar sensor by transmitting data via a bus, and uses an SPI or GPIO analog sensor transmission interface protocol to collect data, thereby realizing serial communication.

[0023] The present invention provides a control method for a trough-type solar thermal solar position double closed-loop real-time tracking control system based on machine vision, and the specific steps are as follows:

[0024] S101: When the sun sensor detects cloudy weather, the driving mechanism remains stationary, otherwise, the process proceeds to step S102;

[0025] S102: When the voltage signal difference output by the two photodiodes of the solar sensor is greater than the set voltage difference, the outer loop coarse adjustment is entered, and the development board controls the driving mechanism to rotate the solar collector to reduce the difference to below the set voltage difference;

[0026] S103: When the voltage signal difference output by the two photodiodes of the solar sensor is less than the set voltage difference, the plastic rod will leave a projected shadow on the light-colored receiving panel. The inner ring is fine-tuned and the camera is used to capture the projected shadow image of the plastic rod on the receiving panel in real time. The development version uses machine vision image processing technology for real-time processing, identifies the projection contour, and calculates the angle between the normal vector of the receiving plate and the normal vector of the solar collector. The fuzzy control algorithm is used to drive the motor to control the rotation angle of the solar collector for precise adjustment, so that the sun is always focused on the solar collector tube.

[0027] As a further improvement of the control method of the present invention, in step S103, the projected shadow image is processed in real time using machine vision image processing technology, and the specific steps are as follows:

[0028] S201: Perform geometric calibration and perspective transformation on the camera;

[0029] S202: Perform brightness correction and grayscale conversion on the input image;

[0030] S203: Filter and reduce noise on the input image;

[0031] S204: Binarization processing is performed on the input image;

[0032] S205: Detect the edge of the shadow band and calculate the distance between the center of mass of the shadow band and the center of the receiving panel.

[0033] S206: Calculate the rotation angle of the heat collecting tank.

[0034] As a further improvement of the control method of the present invention, the step S202 performs brightness correction and grayscale conversion on the input image as follows:

[0035] When using histogram equalization for brightness correction, first count the frequency of each gray level in the image and construct a gray histogram; then calculate the mapping value of each gray level based on the cumulative distribution function CDF; suppose the gray level range of the original image is [0, L-1], the total number of pixels is n, and the frequency of gray level k is n k , then the cumulative distribution function By taking the CDF (k) Multiply by the maximum gray level [L-1] and round to the integer to get a new gray level mapping value, thereby correcting the image brightness so that both dark and bright areas in the image can show the details of the sun's projection more clearly.

[0036] As a further improvement of the control method of the present invention, the step S204: binarization of the input image is specifically as follows;

[0037] Binarization is performed to enhance the contrast between the shadow band and the white background of the receiving panel. The shadow band is separated from the white background using the threshold segmentation method. First, the grayscale histogram of the image is calculated. Suppose the grayscale level of the image is [0, L-1], the total number of pixels is n, and the frequency of grayscale level i is p. i ; Then traverse all possible thresholds t and calculate the inter-class variance σ b 2 (t), the formula is σ b 2 (t)=w0(t)(μ0(t)-μ T ) 2 +w1(t)(μ1(t)-μ T ) 2 Where w0(t) and w1(t) are the probabilities of the two types of pixels after threshold segmentation, μ0(t) and μ1(t) are the average grayscale values ​​of the two types of pixels, and μ T is the total average gray value of the image; find b 2 (t) The maximum threshold t is the Otsu threshold. This threshold is used to segment the image and obtain the projection image segmentation effect.

[0038] As a further improvement of the control method of the present invention, the step S205 performs edge detection of the shadow band and calculates the distance between the centroid of the shadow band and the center of the receiving panel as follows:

[0039] The center of mass is calculated by taking a weighted average of the coordinates of the edge pixel points of the plastic rod projection. Assuming that the coordinates of the edge pixel points are (x i ,y i ), whose weight is determined by the brightness or other characteristics of the pixel; the centroid coordinate (x c ,y c ) is calculated by the following formula, and the obtained centroid position is the projection center position of the plastic plate;

[0040]

[0041] Where n is the number of edge pixels, w i is the weight.

[0042] As a further improvement of the control method of the present invention, the specific steps of calculating the rotation angle of the heat collecting tank in step S206 are as follows:

[0043] It is known that the distance from the plastic plate to the receiving panel is h, and the distance from the center of mass of the plastic rod to the center of the receiving panel is d. Then the angle formed by the normal vector of the receiving panel and the normal vector of the heat collecting tank is:

[0044] As a further improvement of the control method of the present invention, the fuzzy control algorithm in step S103 includes the following steps:

[0045] S301: taking the deviation between the sun position and the target position and the rate of change of the deviation as input variables, and converting them into fuzzy variables using a fuzzification method;

[0046] S302: Building a fuzzy rule base based on experience and experimental data to determine the fuzzy relationship between input variables and output variables;

[0047] S303: Calculating the fuzzy value of the output variable using a fuzzy reasoning method according to the fuzzy value of the input variable and the fuzzy rule base;

[0048] S304: Convert the fuzzy value of the output variable into a clear value as a rotation angle control signal of the reflector.

[0049] Beneficial effects: The present invention has the following advantages:

[0050] This method can adapt to different geographical latitudes and seasonal changes. As long as the sun has a certain light intensity in the sky, it can effectively track the changing characteristics of the local sun position. The present invention adopts image processing technology and advanced control algorithms, which can accurately identify the sun's position and control the rotation angle of the reflector so that the sun is always focused on the collector tube, thereby improving the tracking accuracy. It can adapt to different environmental conditions and system parameter changes, and improve the stability of the system. The solar sensor used in the present invention is coarsely adjusted, and the requirements for the accuracy of the photodiode are not high. The accuracy and aging problems of the photodiode components have little effect on the system. The system composition of the present invention is simple, and the image processing algorithm and control algorithm are easy to implement, which is convenient for promotion and application. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] Figure 1 It is a schematic diagram of the sun position tracking control process of the method of the present invention;

[0052] Figure 2 It is a schematic diagram of calculating the rotation angle of the heat collecting tank under the inner loop control in the present invention. DETAILED DESCRIPTION

[0053] The present invention is further described in detail below in conjunction with the accompanying drawings and specific embodiments:

[0054] A dual-closed-loop real-time tracking control system and method for trough-type solar thermal solar position based on machine vision of the present invention comprises the following steps:

[0055] The system consists of a solar sensor, a camera, a development board, a plastic pole and a receiving panel, a solar trough, and a drive motor. It uses dual feedback of a solar sensor and a camera for closed-loop automatic tracking control. This method uses a photodiode sensor as the outer loop for coarse adjustment to preliminarily determine the position of the sun, that is, to pre-orient the tracking system. Through coarse adjustment, the plastic pole will leave a projected shadow on the light-colored receiving panel. After that, fine adjustment is performed based on machine vision technology to calculate the target rotation angle of the solar trough and track the position of the sun in real time. A digital switch device is used to control the switching between the inner and outer loops. Figure 1 shown.

[0056] Here are the steps:

[0057] S101: When the sun sensor detects cloudy weather, the driving mechanism remains stationary, otherwise, the process proceeds to step S102;

[0058] S102: When the voltage signal difference output by the two photodiodes of the solar sensor is greater than the set voltage difference, the outer loop coarse adjustment is entered, and the development board controls the driving mechanism to rotate the solar collector to reduce the difference to below the set voltage difference;

[0059] S103: When the voltage signal difference output by the two photodiodes of the solar sensor is less than the set voltage difference, the plastic rod will leave a projected shadow on the light-colored receiving panel. The inner ring is fine-tuned and the camera is used to capture the projected shadow image of the plastic rod on the receiving panel in real time. The development version uses machine vision image processing technology for real-time processing, identifies the projection contour, and calculates the angle between the normal vector of the receiving plate and the normal vector of the solar collector. The fuzzy control algorithm is used to drive the motor to control the rotation angle of the solar collector for precise adjustment, so that the sun is always focused on the solar collector tube.

[0060] The solar sensor, camera, development board, plastic rod and receiving panel are fixed on the top of the heat collecting tank through a bracket. The plastic rod must be arranged parallel to the heat collecting tube of the heat collecting tank and will not be blocked by any nearby objects.

[0061] The receiving panel must be arranged parallel to the opening plane of the heat collecting tank, and the center line of the receiving panel is located in the plane determined by the normal vector of the heat collecting tank and the heat collecting tube. The receiving panel has a rectangular shape and a light-colored outer surface.

[0062] Preferably, the development board is a Raspberry Pi development board, and the camera is inserted into a CSI interface on a Raspberry Pi mainboard via a flexible flat cable.

[0063] The sun sensor comprises 2 photodiodes.

[0064] Preferably, the development board is externally connected to an AD chip or circuit to access the two voltages output by the solar sensor to convert the analog signal into a digital signal.

[0065] Optionally, the development board communicates with the solar sensor by using bus data transmission, and uses SPI or GPIO to simulate the sensor transmission interface protocol for data acquisition to achieve serial communication.

[0066] The driving motor is a DC motor and includes at least two speed adjustment gears: high speed and low speed.

[0067] The camera is a camera with high resolution and appropriate sensitivity. High resolution can more accurately capture the details of the projected image, such as the edge and shape changes of the projection. The sensitivity should be adjusted according to the lighting conditions of the use environment to ensure that clear images can be obtained in different weather (sunny, cloudy, etc.) and time (early morning, evening, etc. when the light is dark). The camera needs to face the projection plane, and its field of view should be able to completely cover the area where the projection may appear.

[0068] Furthermore, during inner loop control, the drive motor rotates at high speed, and during outer loop control, the drive motor rotates at low speed to improve control accuracy.

[0069] In step S103, the steps of processing the captured image based on machine vision are as follows:

[0070] S201: Perform geometric calibration and perspective transformation on the camera;

[0071] S202: Perform brightness correction and grayscale conversion on the input image

[0072] S203: Filter and reduce noise on the input image;

[0073] S204: Binarization processing is performed on the input image;

[0074] S205: Detect the edge of the shadow band and calculate the distance between the center of mass of the shadow band and the center of the receiving panel.

[0075] S206: Calculate the rotation angle of the heat collecting tank.

[0076] In step S201, the camera geometry calibration and perspective transformation are used to provide a conversion relationship between the camera coordinate system and the world coordinate system. In the perspective transformation, the pixel coordinates of the image are matched with the world coordinates, and the distortion is removed and the error caused by the camera direction is corrected to ensure accurate measurement.

[0077] The step S202 uses histogram equalization to perform brightness correction, which is a method for enhancing image contrast by adjusting the grayscale histogram of the image. Due to changes in lighting conditions (such as different time and weather, etc.), the image may have uneven brightness. Based on the grayscale probability distribution of the image, the grayscale histogram of the original image is transformed into a uniformly distributed form, thereby expanding the brightness range of the image.

[0078] First, count the frequency of each gray level in the image and construct a gray level histogram. Then calculate the mapping value of each gray level based on the cumulative distribution function (CDF). For example, suppose the gray level range of the original image is [0, L-1], the total number of pixels is n, and the frequency of gray level k is n. k , then the cumulative distribution function By taking the CDF (k) Multiply by the maximum gray level [L-1] and round to the integer to get a new gray level mapping value, thereby correcting the image brightness so that both the darker and brighter areas in the image can show the details of the sun projection more clearly.

[0079] The step S202, grayscale processing: the color image is converted into a grayscale image by using a weighted average method for grayscale conversion, which reduces the amount of calculation and improves the processing speed. By converting the color image into a grayscale image in this way, the main shape and brightness characteristics of the plastic rod projection are retained while reducing the amount of data. The formula used is as follows: Y = 0.299R + 0.587G + 0.114B, where Y is the grayscale value, and R, G, and B are the pixel values ​​of the red, green, and blue channels in the color image, respectively.

[0080] In step S203, the noise in the image may come from a variety of factors, such as the electronic noise of the camera itself, particle scattering in the atmosphere, etc. The median filter method is used for filtering and noise reduction, which is to sort the pixel values ​​in the neighborhood by size, and then take the middle value to replace the value of the central pixel. This method is very effective in removing impulse noise, because impulse noise is usually isolated and has a large difference from the surrounding pixel values. Median filtering can effectively remove this noise without destroying details such as image edges.

[0081] In step S204, a binarization process is performed to enhance the contrast between the shadow band and the white background of the receiving panel, and the shadow band is separated from the white background by using a threshold segmentation method. The grayscale histogram of the image is divided into two parts using an optimal threshold value so that the variance between the two classes is maximized, i.e., the separability is maximized. Thus, the binary segmentation of the image is achieved. The Otsu threshold method is a method for automatically determining the optimal threshold based on the grayscale histogram of the image. Its goal is to maximize the inter-class variance of the foreground and background pixels after segmentation. First, the grayscale histogram of the image is calculated, assuming that the grayscale level of the image is [0, L-1], the total number of pixels is n, and the frequency of occurrence of grayscale level i is p. i Then traverse all possible thresholds t and calculate the inter-class variance σ b 2 (t), the formula is σ b 2 (t)=w0(t)(μ0(t)-μ T ) 2 +w1(t)(μ1(t)-μ T ) 2

[0082] Where w0(t) and w1(t) are the probabilities of the two types of pixels after threshold segmentation, μ0(t) and μ1(t) are the average grayscale values ​​of the two types of pixels, and μ T is the total average gray value of the image. Find b 2 (t) The maximum threshold t is the Otsu threshold. Using this threshold to segment the image can obtain a better projection image segmentation effect.

[0083] In step S205, Canny edge detection is used to perform edge detection, and the gradient amplitude and direction of the image are calculated. The gradient amplitude represents the intensity of pixel changes in the image, and the gradient direction represents the direction in which the pixel changes fastest. Through non-maximum suppression, the pixels that are not the strongest on the edge (i.e., the pixels whose gradient amplitude is not the local maximum) are suppressed, thereby refining the edge. Finally, through dual threshold detection, the true edge pixels are determined according to the high and low thresholds. The high threshold is used to detect strong edges, and the low threshold is used to connect weak edges, so that a relatively complete edge of the sun can be obtained.

[0084] In step S205, the coordinates of the edge pixel points of the plastic rod projection are weighted averaged to calculate the centroid. Assume that the coordinates of the edge pixel points are (x i ,y i ), whose weights can be determined based on the brightness or other features of the pixel. c ,y c ) can be calculated by the following formula, and the obtained centroid position is the projection center position of the plastic plate.

[0085]

[0086] Where n is the number of edge pixels, w i is the weight

[0087] The step S206, Figure 2 As shown, it is known that the distance from the plastic plate to the receiving panel is h, and the distance from the center of mass of the plastic rod to the center of the receiving panel is d, then the angle formed by the normal vector of the receiving panel and the normal vector of the heat collecting tank is:

[0088]

[0089] The field of view of the shadow tracking device based on machine vision, which is composed of the camera, plastic rod and receiving panel, cannot be designed to be too large, because the larger the field of view, the lower the accuracy of image recognition and tracking. Due to the limitation of the structural size, the field of view cannot be designed to be too large. The field of view is generally 4-7°. When the shadow of the plastic rod exceeds this range, the shadow tracking device based on machine vision cannot detect the image of the plastic rod.

[0090] At the same time, in order to achieve high-precision tracking, the shadow tracking device based on machine vision is used as the inner loop for control, and the sun sensor is used as the outer loop for control. Both the inner and outer loops are closed-loop control, thus forming a double closed-loop control. The outer loop is coarsely adjusted, and the inner loop is finely adjusted.

[0091] The solar sensor has a wide field of view, generally 40-50°. The solar sensor uses two photodiodes. When the light of the photodiodes is strong, the output voltage is low, and when the light of the photodiodes is weak, the output voltage is high. According to the difference in output voltage, the angle formed between the normal vector of the receiving plate and the normal vector of the heat collecting tank can be calculated, thereby controlling the rotation of the drive motor.

[0092] Since solar sensors are usually susceptible to diffuse solar radiation and temperature changes, have low linearity and repeatability, and have time-varying responses, the accuracy of this loop is low. Therefore, it is not possible to accurately calculate the deviation angle of the sun's rays using only solar sensors;

[0093] However, this system only uses the sun sensor for coarse adjustment, that is, to pre-align the tracking system. After the pre-alignment is completed, the digital switch device is used to switch the system to the inner control loop, that is, fine adjustment. After setting the dead zone,

[0094] According to the deviation between the sun's position and the target position, the fuzzy control algorithm is used to calculate the rotation angle of the reflector so that the sun is always focused on the collector tube. The fuzzy control algorithm has strong robustness and adaptability, and can adapt to different environmental conditions and system parameter changes.

[0095] The fuzzy control algorithm comprises the following steps:

[0096] S301: taking the deviation between the sun position and the target position and the rate of change of the deviation as input variables, and converting them into fuzzy variables using a fuzzification method.

[0097] S302: Based on experience and experimental data, a fuzzy rule base is established to determine the fuzzy relationship between input variables and output variables.

[0098] S303: Calculate the fuzzy value of the output variable using a fuzzy reasoning method according to the fuzzy value of the input variable and the fuzzy rule base.

[0099] S304: Convert the fuzzy value of the output variable into a clear value as a rotation angle control signal of the reflector.

[0100] The above description is only a preferred embodiment of the present invention and does not constitute any other form of limitation to the present invention. Any modification or equivalent change made based on the technical essence of the present invention still falls within the scope of protection required by the present invention.

Claims

1. A dual closed-loop real-time tracking control system for the position of the sun in a trough-type solar thermal system based on machine vision, characterized by: It consists of a photodiode-based solar sensor, a camera, a development board, an IO module, a plastic rod and a receiving panel, a heat collecting tank and a driving motor; The solar sensor, camera, development board, plastic rod and receiving panel are fixed on the top of the heat collecting tank through a bracket. The plastic rod must be arranged parallel to the heat collecting tube of the heat collecting tank and will not be blocked by any nearby objects; The receiving panel must be arranged parallel to the opening plane of the heat collecting tank, and the midline of the receiving panel is located in the plane determined by the normal vector of the heat collecting tank and the heat collecting tube. The receiving panel has a rectangular shape and a light-colored outer surface; The drive motor adopts a DC motor and includes at least two speed adjustment gears of high speed and low speed. When the inner loop is controlled, the drive motor rotates at high speed, and when the outer loop is controlled, the drive motor rotates at low speed to improve the control accuracy; The dual feedback of the sun sensor and the camera is used for closed-loop automatic tracking control. The photodiode sensor is used for rough control to preliminarily determine the position of the sun. The camera captures the projection image of the plastic pole on the receiving panel in real time, and uses machine vision image processing technology to identify the projection contour. The rotation angle of the heat collecting trough is controlled for precise adjustment to ensure that the sun is always focused on the collector tube.

2. The dual closed-loop real-time tracking control system for trough-type solar thermal solar position based on machine vision according to claim 1 is characterized in that: The development board adopts a Raspberry Pi development board, and the camera is inserted into the CSI interface on the Raspberry Pi mainboard through a flexible cable.

3. The dual closed-loop real-time tracking control system for trough-type solar thermal solar position based on machine vision according to claim 2 is characterized in that: The development board is connected to an external AD chip or circuit to access the two voltages output by the solar sensor, converting the analog signal into a digital signal or the development board communicates with the solar sensor by using a bus to transmit data, and uses the SPI or GPIO analog sensor transmission interface protocol to collect data and realize serial communication.

4. The control method of the trough type solar thermal solar position double closed-loop real-time tracking control system based on machine vision according to any one of claims 1 to 3, characterized in that: The specific steps are as follows: S101: When the sun sensor detects cloudy weather, the driving mechanism remains stationary, otherwise, the process proceeds to step S102; S102: When the voltage signal difference output by the two photodiodes of the solar sensor is greater than the set voltage difference, the outer loop coarse adjustment is entered, and the development board controls the driving mechanism to rotate the solar collector to reduce the difference to below the set voltage difference; S103: When the voltage signal difference output by the two photodiodes of the solar sensor is less than the set voltage difference, the plastic rod will leave a projected shadow on the light-colored receiving panel. The inner ring is fine-tuned and the camera is used to capture the projected shadow image of the plastic rod on the receiving panel in real time. The development version uses machine vision image processing technology for real-time processing, identifies the projection contour, and calculates the angle between the normal vector of the receiving plate and the normal vector of the solar collector. The fuzzy control algorithm is used to drive the motor to control the rotation angle of the solar collector for precise adjustment, so that the sun is always focused on the solar collector tube.

5. The control method of the trough type solar thermal solar position double closed-loop real-time tracking control system based on machine vision according to claim 4 is characterized by: In step S103, the projected shadow image is processed in real time using machine vision image processing technology, and the specific steps are as follows: S201: Perform geometric calibration and perspective transformation on the camera; S202: Perform brightness correction and grayscale conversion on the input image; S203: Filter and reduce noise on the input image; S204: Binarization processing is performed on the input image; S205: Detect the edge of the shadow band and calculate the distance between the center of mass of the shadow band and the center of the receiving panel. S206: Calculate the rotation angle of the heat collecting tank.

6. The control method of the trough type solar thermal solar position double closed-loop real-time tracking control system based on machine vision according to claim 5 is characterized by: The step S202 performs brightness correction and grayscale conversion on the input image as follows: When using histogram equalization for brightness correction, first count the frequency of each gray level in the image and construct a gray histogram; then calculate the mapping value of each gray level based on the cumulative distribution function CDF; suppose the gray level range of the original image is [0, L-1], the total number of pixels is n, and the frequency of gray level k is n k , then the cumulative distribution function By taking the CDF (k) Multiply by the maximum gray level [L-1] and round to the integer to get a new gray level mapping value, thereby correcting the image brightness so that both dark and bright areas in the image can show the details of the sun's projection more clearly.

7. The control method of the trough type solar thermal solar position double closed-loop real-time tracking control system based on machine vision according to claim 5 is characterized by: The step S204: binarization of the input image is specifically as follows; Binarization is performed to enhance the contrast between the shadow band and the white background of the receiving panel. The shadow band is separated from the white background using the threshold segmentation method. First, the grayscale histogram of the image is calculated. Suppose the grayscale level of the image is [0, L-1], the total number of pixels is n, and the frequency of grayscale level i is p. i ; Then traverse all possible thresholds t and calculate the inter-class variance σ b 2 (t), the formula is σ b 2 (t)=w0(t)(μ0(t)-μ T ) 2 +w1(t)(μ1(t)-μ T ) 2 , where w0(t) and w1(t) are the probabilities of the two types of pixels after threshold segmentation, μ0(t) and μ1(t) are the average grayscale values ​​of the two types of pixels, and μ T is the total average gray value of the image; find b 2 (t) The maximum threshold t is the Otsu threshold. This threshold is used to segment the image and obtain the projection image segmentation effect.

8. The control method of the trough type solar thermal solar position double closed-loop real-time tracking control system based on machine vision according to claim 5 is characterized by: The steps of performing edge detection of the shadow band and calculating the distance between the centroid of the shadow band and the center of the receiving panel in step S205 are as follows: The center of mass is calculated by taking a weighted average of the coordinates of the edge pixel points of the plastic rod projection. Assuming that the coordinates of the edge pixel points are (x i ,y i ), whose weights are determined based on the brightness or other characteristics of the pixels; Center of mass coordinates (x c ,y c ) is calculated by the following formula, and the obtained centroid position is the projection center position of the plastic plate; Where n is the number of edge pixels, w i is the weight.

9. The control method of the trough type solar thermal solar position double closed-loop real-time tracking control system based on machine vision according to claim 5 is characterized by: The specific steps of calculating the rotation angle of the heat collecting tank in step S206 are as follows: It is known that the distance from the plastic plate to the receiving panel is h, and the distance from the center of mass of the plastic rod to the center of the receiving panel is d. Then the angle formed by the normal vector of the receiving panel and the normal vector of the heat collecting tank is:

10. The control method of the trough type solar thermal solar position double closed-loop real-time tracking control system based on machine vision according to claim 4, characterized in that: The fuzzy control algorithm in step S103 includes the following steps: S301: taking the deviation between the sun position and the target position and the rate of change of the deviation as input variables, and converting them into fuzzy variables using a fuzzification method; S302: Building a fuzzy rule base based on experience and experimental data to determine the fuzzy relationship between input variables and output variables; S303: Calculating the fuzzy value of the output variable using a fuzzy reasoning method according to the fuzzy value of the input variable and the fuzzy rule base; S304: Convert the fuzzy value of the output variable into a clear value as a rotation angle control signal of the reflector.

Images