Automatic blanking control method and system for automatic concrete production

Abstract

The present invention relates to the technical field of concrete production. More specifically, the present invention relates to an automatic blanking control method and system for automated concrete production, including: collecting an image of a feed hopper, grayscale processing the image to obtain a grayscale image; constructing an edge recognition model and a key point detection model based on a neural network, and training the constructed edge recognition model and key point detection model using sample data; inputting the preprocessed grayscale image into the trained edge recognition model to output a recognition result. According to the method of the neural network, the present invention monitors the blanking situation of the feed hopper, greatly avoiding the problem that the concrete in the feed hopper is likely to overflow when there is too much. And in another embodiment, the data measured by a sensor is also combined to monitor the blanking situation of the feed hopper, further improving the accuracy of detection and avoiding the problem of inaccurate detection relying only on image processing.

Description

Technical Field

[0001] The present invention relates to the technical field of concrete production. More specifically, the present invention relates to an automatic blanking control method and system for automated concrete production. Background Art

[0002] Concrete refers to a general term for engineering composite materials in which aggregate is cemented into a whole by a gelling material. Usually, the term "concrete" refers to cement concrete, also known as ordinary concrete, which is made of cement as the gelling material, sand and stone as aggregate, and mixed with water in a certain proportion and stirred. It is widely used in civil engineering. As a basic material in modern construction industry, concrete has a huge market demand, so concrete production lines have emerged as the times require. A concrete production line is an industrial production device integrating automation and mechanization, mainly used for producing concrete precast parts and cast-in-place concrete components, etc.

[0003] The concrete production line includes conveying equipment, which is used to convey raw materials and mixed concrete to each production link. Common conveying equipment includes conveyor belts, concrete pump trucks, mixer trucks, etc. These devices can ensure the timely supply of raw materials and concrete and improve production efficiency.

[0004] Currently, when the mixed concrete is conveyed to a mixer truck through a conveyor belt in a concrete production line, the feeding hopper on the mixer truck is needed to receive the material. However, due to the relatively fast feeding speed of the small material in the cement production device, if the staff makes a supervision mistake during work, it is easy to cause the cement in the feeding hopper to overflow, and then cause the concrete in the feeding hopper to overflow, resulting in waste of materials. Summary of the Invention

[0005] The present invention provides an automatic blanking control method for automated concrete production, aiming to solve the problem in the related art that due to the relatively fast feeding speed of the small material in the cement production device, if the staff makes a supervision mistake during work, it is easy to cause the cement in the feeding hopper to overflow, and then cause the concrete in the feeding hopper to overflow, resulting in waste of materials.

[0006] In a first aspect, the present invention provides an automatic blanking control method for automated concrete production, including: collecting an image of a feed hopper, grayscaling the image to obtain a grayscale image; constructing an edge recognition model and a key point detection model based on a neural network, and training the constructed edge recognition model and key point detection model using sample data; inputting the preprocessed grayscale image into the trained edge recognition model, and outputting a recognition result, where the recognition result includes the demarcation edge between the concrete area and the feed hopper area, and the upper edge of the feed hopper in the feed hopper area; for the grayscale image after the recognition result, inputting it into the trained key point detection model, and outputting a first key point on the demarcation edge and a second key point on the upper edge of the feed hopper; calculating the material capacity of the feed hopper, where the material capacity is positively correlated with the distance between the first key point and the second key point and the area of the concrete area; controlling the opening and closing of the closing plate on the feed hopper according to the size of the material capacity of the feed hopper.

[0007] Further, the preprocessed grayscale image includes: preprocessing the grayscale image using a semantic segmentation method to obtain a concrete area and a feed hopper area.

[0008] Further, constructing an edge recognition model based on a neural network includes: obtaining multiple preprocessed grayscale image samples; annotating the demarcation edge between the concrete area and the feed hopper area and the upper edge of the feed hopper in the feed hopper area in the grayscale image samples to obtain annotated grayscale image samples; using the annotated grayscale image samples as training samples to iteratively train the neural network; determining the difference between the edge prediction information output by the neural network in each iteration process and the information of the annotated grayscale image samples; adjusting the network parameters of the neural network based on the difference until the training stop condition is met and the training ends to obtain the edge detection model.

[0009] Further, calculating the material capacity of the feed hopper, the calculation formula is: ; in the formula, represents the material capacity of the feed hopper, represents the distance between the first key point and the second key point, represents the area of the concrete area.

[0010] Further, controlling the opening and closing of the closing plate on the feed hopper according to the size of the material capacity of the feed hopper includes: in response to the material capacity of the feed hopper being greater than a first preset threshold, closing the closing plate on the feed hopper; in response to the material capacity of the feed hopper being less than or equal to the first preset threshold, not closing the closing plate on the feed hopper.

[0011] Further, according to the material capacity level of the feed hopper, the opening and closing of the closing plate on the feed hopper is controlled, and it further includes: in response to the material capacity level of the feed hopper being greater than a second preset threshold, controlling the movement of the closing plate on the feed hopper to reduce the material discharge gap between the feed hopper and the closing plate.

[0012] In response to the material capacity level of the feed hopper being greater than a third preset threshold, closing the closing plate on the feed hopper.

[0013] Further, it further includes: collecting the current pressure of the sensor; calculating the storage coefficient of the feed hopper according to the difference between the current pressure and the maximum bearing capacity, where the maximum bearing capacity is the pressure value measured after the feed hopper bears concrete to its maximum capacity, and the storage coefficient is inversely correlated with the difference; controlling the opening and closing of the closing plate on the feed hopper according to the sum of the storage coefficient of the feed hopper and the material capacity level of the feed hopper.

[0014] Further, controlling the opening and closing of the closing plate on the feed hopper according to the sum of the storage coefficient of the feed hopper and the material capacity level of the feed hopper includes: calculating the average value between the storage coefficient of the feed hopper and the material capacity level of the feed hopper; in response to the average value being greater than a fourth preset threshold, closing the closing plate on the feed hopper;

[0015] Further, the empirical value of the first preset threshold is 0.7.

[0016] In the second aspect of the present invention, an automatic material discharge control system for automated concrete production is further provided, which is characterized by including a processor and a memory, where the memory stores a computer program, and the processor executes the computer program to implement the automatic material discharge control method for automated concrete production described in any one of the above.

[0017] Beneficial effects: According to the method of neural network, the material discharge situation of the feed hopper is monitored, which greatly avoids the problem that the concrete in the feed hopper may overflow when there is too much. And in another embodiment, the data measured by the sensor is also combined to monitor the material discharge situation of the feed hopper, further improving the detection accuracy and avoiding the problem of inaccurate detection relying only on image processing. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] By referring to the accompanying drawings and reading the following detailed description, several embodiments of the present invention are shown in an exemplary rather than restrictive manner, and the same or corresponding reference numerals represent the same or corresponding parts, where:

[0019] Figure 1 It is a flowchart schematically showing the calculation of the material capacity situation of the feed hopper according to an embodiment of the present invention.

[0020] Figure 2It is a flowchart schematically showing the storage coefficient calculation of a feed hopper according to an embodiment of the present invention.

[0021] Figure 3 It is a schematic structural diagram of a system according to an embodiment of the present invention.

[0022] Figure 4 It is a schematic structural diagram of a feed hopper according to an embodiment of the present invention.

[0023] Reference numerals

[0024] 1. Processor; 2. Communication bus; 3. Memory; 4. Communication interface. Detailed implementation manners

[0025] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, rather than all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative efforts belong to the protection scope of the present invention.

[0026] Next, the detailed implementation manners of the present invention will be described in detail in conjunction with the accompanying drawings.

[0027] The concrete production line includes conveying equipment, which is used to convey raw materials and mixed concrete to each production link. Common conveying equipment includes conveyor belts, concrete pump trucks, transport vehicles, etc. These devices can ensure the timely supply of raw materials and concrete and improve production efficiency. Currently, when the mixed concrete in the concrete production line is conveyed from the production device to the transport vehicle through the production device, the production device is used to produce cement, and a baffle is installed at the discharge port of the production device. However, due to the unstable discharge speed of the production device, when the manual monitoring of the discharging process is neglected, the concrete in the feed hopper overflows, resulting in waste of materials. Therefore, the following method is proposed to solve the above problems.

[0028] In one embodiment, it includes a transport vehicle with a tank on it. The tank is used to store concrete, and there is also a feed hopper on the tank. At the same time, a baffle is installed at the discharge port of the production device, and the movement of the baffle is controlled by a hydraulic mechanism, thereby realizing the opening and closing of the upper and lower discharge ports of the production device, avoiding the overflow of concrete in the feed hopper and resulting in waste of materials.

[0029] S101: Collect images of the feed hopper.

[0030] In one embodiment, an image acquisition device is installed on a production device so that the camera can capture a frontal image of the feed hopper on the transport vehicle. Specifically, in the frontal image, the feed hopper is located in the exact middle of the image. In this embodiment, the image acquisition device can be a camera or an industrial camera, etc. With the above structure, an image of the feed hopper can be obtained. Then, the image of the feed hopper is grayscale processed to obtain a grayscale image. Additionally, noise in the grayscale image can be removed. For example, filtering techniques (such as Gaussian filtering) are applied to remove image noise, facilitating subsequent processing of the grayscale image. It should be noted that Gaussian filtering is a widely used linear smoothing filtering technique in image processing, mainly used to eliminate Gaussian noise in the image, improve the signal-to-noise ratio (SNR) of the image, and thus improve the image quality.

[0031] In one embodiment, after obtaining the grayscale image, preprocessing of the grayscale image is also required. Here, preprocessing refers to segmenting the grayscale image, which can obtain the concrete area and the feed hopper area in the grayscale image. Specifically, the grayscale image can be segmented by a semantic segmentation method to obtain the concrete area and the feed hopper area, facilitating subsequent determination of whether the feed hopper is filled with concrete.

[0032] In one embodiment, the feed hopper area and the concrete area in the grayscale image are segmented by semantic segmentation. Semantic segmentation is a task in the field of computer vision, aiming to assign a specific semantic label to each pixel in the image, that is, classify each pixel so that each object area in the image can be accurately divided and recognized. Specifically, semantic segmentation includes the following steps. Build a dataset. Specifically, images of all models of feed hoppers and images of concrete involved are collected to form a dataset. Data annotation. Specifically, the labels of pixel points in the image are divided into two categories. Among them, the pixel points in the concrete area of the image are labeled as 1, and the pixel points in the feed hopper area are labeled as 0. Network selection. Specifically, in this embodiment, the Unet network structure is used, and implementers can select network modules within the Unet network structure according to actual situations.

[0033] S102: Construct an edge recognition model and a key point detection model based on a neural network.

[0034] In one embodiment, an edge recognition model is constructed based on a neural network, including: first, obtaining multiple preprocessed grayscale image samples; annotating the boundary edges between the concrete area and the hopper area and the upper edge of the hopper in the hopper area in the grayscale image samples to obtain the annotated grayscale image samples; using the annotated grayscale image samples as training samples to iteratively train the neural network; determining the difference between the edge prediction information output by the neural network and the information of the annotated grayscale image samples during each iteration; adjusting the network parameters of the neural network based on the difference until the training stops when the training stop condition is met, and obtaining the edge detection model.

[0035] In another embodiment, to recognize the edges of a grayscale image, in addition to using the neural network algorithm, an edge detection algorithm can also be directly used for detection. The specific steps are as follows: perform edge detection on the grayscale image to obtain the boundary edges between the concrete area and the hopper area and the upper edge of the hopper in the hopper area. In this embodiment, the Canny edge detection method can be used to perform segmentation processing on the grayscale image. The Canny detection algorithm, that is, the Canny edge detection algorithm, is a widely used image processing algorithm. The purpose of this algorithm is to find the optimal edges in the image, ensuring good detection, positioning, and minimum response. The characteristics of using the Canny detection algorithm are: high accuracy: The Canny algorithm can accurately detect the real edges in the image and precisely locate the edge pixels. Noise suppression: Through steps such as Gaussian filtering and non-maximum suppression, the Canny algorithm can effectively suppress noise interference and improve the accuracy of edge detection. Edge connection: By using double-threshold detection and edge connection algorithms, the Canny algorithm can connect weak edges with strong edges to obtain continuous edge lines. Parameter adjustability: The parameters of the Canny algorithm (such as the size of the Gaussian filter, the setting of double thresholds, etc.) can be adjusted according to specific applications to obtain edge detection results that meet the requirements.

[0036] In one embodiment, a key point detection model is constructed based on a neural network, including: first, obtaining multiple edge images output from the edge recognition model; and annotating the edge images. The annotation includes: marking the longitudinal center line of the edge image, and determining the point where the longitudinal center line intersects the boundary edge as the first key point and the point where it intersects the upper edge of the hopper as the second key point, thereby obtaining the annotated edge image; using the annotated edge image as a training sample to iteratively train the neural network; determining the difference between the edge prediction information output by the neural network and the information of the annotated edge image during each iteration; adjusting the network parameters of the neural network based on the difference until the training stops when the training stop condition is met, and obtaining the key point detection model.

[0037] S103: Calculate the material capacity level of the feed hopper.

[0038] Input the preprocessed grayscale image into the trained edge recognition model to output the recognition result, where the recognition result includes the demarcation edge between the concrete area and the feed hopper area, and the upper edge of the feed hopper in the feed hopper area. For the grayscale image after the recognition result, input it into the trained key point detection model to output the first key point on the demarcation edge and the second key point on the upper edge of the feed hopper.

[0039] In one embodiment, calculate the material capacity level of the feed hopper, and the material capacity level is positively correlated with the distance between the first key point and the second key point and the area of the concrete area.

[0040] In one embodiment, calculating the material capacity level of the feed hopper includes: calculating the material capacity level of the feed hopper according to the distance between the first key point and the second key point. The smaller the distance between the first key point and the second key point, the greater the material capacity level of the feed hopper, indicating that the concrete is closer to the top edge of the feed hopper, indicating that there is more concrete in the feed hopper.

[0041] In one embodiment, calculating the material capacity level of the feed hopper further includes: The calculation formula is: ; In the formula, represents the material capacity level of the feed hopper, represents the distance between the first key point and the second key point, represents the area of the concrete area; in this embodiment, when calculating the material capacity level of the feed hopper, the area of the concrete area is also used to correct the material capacity level of the feed hopper. The reason is that: since the shape of the feed hopper is basically funnel-shaped, that is, wider at the top and narrower at the bottom, when the area of the concrete area is larger, it indicates that the concrete is closer to the upper part of the feed hopper, then the concrete inside the feed hopper is larger at this time, and the material capacity level of the feed hopper is larger.

[0042] S104: Control the opening and closing of the closing plate on the feed hopper according to the size of the material capacity level of the feed hopper.

[0043] In one embodiment, if the material capacity level of the feed hopper is greater than the first preset threshold, it indicates that the cement in the feed hopper is about to overflow, then close the closing plate on the feed hopper; if the material capacity level of the feed hopper is less than or equal to the first preset threshold, it indicates that the cement in the feed hopper is not full yet, then there is no need to close the closing plate on the feed hopper. In this embodiment, the empirical value of the first preset threshold is 0.7. In other embodiments, the empirical value of the first preset threshold can be 0.8 or 0.85, etc., and can be adjusted according to the actual situation.

[0044] In another embodiment, if the capacity of the feed hopper is greater than a second preset threshold, the closing plate on the feed hopper is controlled to move to reduce the gap between the feed hopper and the closing plate for material discharge. In response to the capacity of the feed hopper being greater than a third preset threshold, the closing plate on the feed hopper is closed. That is to say, in this embodiment, the closing plate can be gradually closed as the capacity of the feed hopper increases. When the cement inside the feed hopper is almost full, the amount of cement discharged from the production device is reduced to avoid the problem of cement overflowing from the feed hopper. In this embodiment, the empirical value of the second preset threshold is 0.65, and the empirical value of the third preset threshold is 0.95. In other embodiments, the empirical value of the third preset threshold can be 0.9 or 0.92, etc.

[0045] In one embodiment, since the camera is often at the concrete filling site, some dust or concrete may cause the camera to take blurred pictures, resulting in deviations in the final calculated results and inaccurate calculation results. Therefore, another method can be used to improve the accuracy of the calculation results. The specific method is as follows.

[0046] S201: Collect the current pressure of the sensor.

[0047] In one embodiment, the sensor is a pressure sensor, and there are two sensors, which are symmetrically mounted on the inner walls of both sides of the feed hopper (such as Figure 4 As shown in the figure (the shaded area represents the sensor), the higher the pressure measured by the pressure sensor, the more concrete in the feed hopper and the more likely it is to overflow. It is important to note that when there are multiple sensors in the feed hopper, the average of the pressures measured by all sensors is used as the current pressure when calculating pressure. By using multiple sensors to measure pressure, this structure can reduce errors and improve the accuracy of measurement data.

[0048] S202: Calculate the storage coefficient of the feed hopper.

[0049] In one embodiment, the storage coefficient of the feed hopper is calculated based on the difference between the current pressure and the maximum bearing capacity. The storage coefficient is inversely correlated with the difference. Specifically, the difference between the current pressure and the maximum bearing capacity is normalized, and the normalized value is used as the storage coefficient of the feed hopper. The maximum bearing capacity is the pressure measured after the feed hopper has been filled with concrete to its maximum capacity. In other embodiments, the maximum bearing capacity can be obtained by performing the above steps multiple times, with the final average value being used as the maximum bearing capacity.

[0050] S203: Control the opening and closing of the closing plate on the feed hopper.

[0051] In one embodiment, the opening and closing of the closing plate on the feed hopper is controlled according to the sum of the storage coefficient of the feed hopper and the material capacity level of the feed hopper. Specifically, the average value between the storage coefficient of the feed hopper and the material capacity level of the feed hopper is calculated; if the average value is greater than a fourth preset threshold, the opening and closing of the closing plate on the feed hopper is reduced. In this embodiment, the empirical value of the fourth preset threshold is 0.75. In other embodiments, the empirical value of the fourth preset threshold can be 0.8 or 0.72, etc.

[0052] Exemplarily, it is calculated that the storage coefficient of the feed hopper is , and the material capacity level of the feed hopper is , then the average value is calculated, and the average value is equal to . If the average value is greater than 0.75, at this time, the opening and closing of the closing plate on the feed hopper is reduced.

[0053] In another embodiment, the opening and closing of the closing plate on the feed hopper can also be controlled solely by the storage coefficient of the feed hopper. Specifically, if the storage coefficient of the feed hopper alone is greater than a preset coefficient value, the closing plate on the feed hopper is closed. In this embodiment, the preset coefficient value is 0.9. In other embodiments, the preset coefficient value can be 0.91 or 0.92, etc.

[0054] Through the above steps, the feeding situation of the feed hopper can be monitored by means of an image processing method, greatly avoiding the problem that the concrete in the feed hopper is likely to overflow. And in another embodiment, the data measured by the sensor is also combined to monitor the feeding situation of the feed hopper, improving the detection accuracy.

[0055] The present invention also provides an automatic feeding control method for automated concrete production. As Figure 2 shown, the system includes a processor 1 and a memory 3. The memory 3 stores computer program instructions, and when the computer program instructions are executed by the processor 1, an automatic feeding control method for automated concrete production according to the first aspect of the present invention is implemented.

[0056] The system further includes a communication bus 2 and a communication interface 4 and other components well known to those skilled in the art. Their settings and functions are known in the art, so they will not be elaborated here.

[0057] In the present invention, the aforementioned memory 3 may be any tangible medium that contains or stores a program, which can be used by or in conjunction with an instruction execution system, apparatus, or device. For example, a computer-readable storage medium may be any suitable magnetic storage medium or magneto-optical storage medium, such as, resistive random access memory (RRAM), dynamic random access memory (DRAM), static random-access memory (SRAM), enhanced dynamic random access memory (EDRAM), high-bandwidth memory (HBM), hybrid memory cube (HMC), etc., or any other medium that can be used to store the required information and can be accessed by an application program, module, or both. Any such computer storage medium may be part of the device or accessible or connectable to the device. Any application or module described in the present invention may be implemented using computer-readable / executable instructions that can be stored or otherwise held by such a computer-readable medium.

[0058] In the description of this specification, the meanings of "a plurality of" and "several" are at least two, such as two, three, or more, etc., unless otherwise specifically defined.

[0059] The technical features of the above-described embodiments can be combined arbitrarily. For the sake of brevity of description, not all possible combinations of the technical features in the above-described embodiments are described. However, as long as there is no contradiction in the combination of these technical features, it should be considered as the scope described in this specification.

[0060] The above-described embodiments merely represent several implementation manners of the present invention. The description thereof is relatively specific and detailed, but it should not be construed as a limitation on the scope of the patent application. It should be noted that for those of ordinary skill in the art, without departing from the concept of the present invention, several modifications and improvements can still be made, and these all belong to the protection scope of the present invention.

Claims

1. An automatic blanking control method for automated concrete production, characterized in that, Including: Collect an image of the feeding hopper, grayscale the image to obtain a grayscale image; Preprocess the grayscale image using a semantic segmentation method to obtain a concrete area and a feeding hopper area; Construct an edge recognition model and a key point detection model based on a neural network, and train the constructed edge recognition model and key point detection model using sample data; Input the preprocessed grayscale image into the trained edge recognition model, and output a recognition result, where the recognition result includes the demarcation edge between the concrete area and the feeding hopper area, and the upper edge of the feeding hopper in the feeding hopper area; For the grayscale image after the recognition result, input it into the trained key point detection model, and output a first key point on the demarcation edge and a second key point on the upper edge of the feeding hopper; Calculate the material capacity level of the feed hopper. The material capacity level is positively correlated with the distance between the first key point and the second key point and the area of the concrete area. The calculation formula is: ; In the formula, represents the material capacity level of the feed hopper, represents the distance between the first key point and the second key point, represents the area of the concrete area; Control the opening and closing of the closing plate on the feeding hopper according to the material capacity degree of the feeding hopper; Also including: collecting the current pressure of the sensor; Calculate the storage coefficient of the feeding hopper according to the difference between the current pressure and the maximum bearing capacity, where the maximum bearing capacity is the pressure value measured after the feeding hopper bears concrete to the maximum capacity, and the storage coefficient is inversely correlated with the difference; Control the opening and closing of the closing plate on the feeding hopper according to the sum of the storage coefficient of the feeding hopper and the material capacity degree of the feeding hopper.

2. The automatic blanking control method for automated concrete production according to claim 1, characterized in that, Constructing an edge recognition model based on a neural network, including: Obtain multiple preprocessed grayscale image samples; Label the demarcation edge between the concrete area and the feeding hopper area and the upper edge of the feeding hopper in the feeding hopper area in the grayscale image sample to obtain a labeled grayscale image sample; Use the labeled grayscale image sample as a training sample to iteratively train the neural network; Determine the difference between the edge prediction information output by the neural network and the information of the labeled grayscale image sample during each iteration process; Adjust the network parameters of the neural network based on the difference until the training stop condition is met and the training ends to obtain the edge recognition model.

3. The automatic blanking control method for automated concrete production according to claim 1, wherein, Controlling the opening and closing of the closing plate on the feeding hopper according to the material capacity degree of the feeding hopper, including: In response to the material capacity degree of the feeding hopper being greater than a first preset threshold, close the closing plate on the feeding hopper; In response to the material capacity degree of the feeding hopper being less than or equal to the first preset threshold, do not close the closing plate on the feeding hopper.

4. The automatic blanking control method for automated concrete production according to claim 1, characterized in that Controlling the opening and closing of the closing plate on the feeding hopper according to the material capacity degree of the feeding hopper, further including: In response to the material capacity degree of the feeding hopper being greater than a second preset threshold, control the closing plate on the feeding hopper to move to reduce the material discharging gap between the feeding hopper and the closing plate; In response to the material capacity degree of the feeding hopper being greater than a third preset threshold, close the closing plate on the feeding hopper.

5. The automatic blanking control method for automated concrete production according to claim 1, characterized in that, Controlling the opening and closing of the closing plate on the feeding hopper according to the sum of the storage coefficient of the feeding hopper and the material capacity degree of the feeding hopper, including: Calculate the average value between the storage coefficient of the feeding hopper and the material capacity degree of the feeding hopper; In response to the average value being greater than a fourth preset threshold, close the closing plate on the feeding hopper.

6. An automatic blanking control system for automated concrete production, characterized in that, It includes a processor and a memory. The memory stores a computer program, and the processor executes the computer program to implement the automatic blanking control method for automated concrete production as described in any one of claims 1-5.

Images