Visual detection device and method for smelting end point of anode furnace
Through the combination of high-definition industrial cameras and deep learning models, high-precision, real-time detection of the anode furnace smelting end point is achieved, solving the problems of low detection accuracy and large environmental interference in existing technologies, and improving the purity of copper liquid and production efficiency.
Patent Information
- Application Number
- CN202510810044.9
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-06-17
- Publication Date
- 2025-10-10
AI Technical Summary
The existing technology for anode furnace smelting endpoint detection has the disadvantages of low accuracy, high labor intensity and susceptibility to environmental interference, and cannot meet the needs of high-precision real-time detection under high temperature, high dust and strong electromagnetic interference.
A high-definition industrial camera combined with adaptive histogram equalization and hybrid filtering algorithm is used for image preprocessing. A detection model is constructed by combining deep learning convolutional neural network and long short-term memory network. The accurate determination of the smelting endpoint is achieved through image feature extraction and model training.
The accuracy and real-time performance of anode furnace smelting endpoint detection are improved, the subjectivity of manual judgment and the influence of environmental interference are reduced, and the purity of copper liquid and production efficiency are guaranteed.
Smart Images
Figure CN120760024A_ABST
Abstract
Description
TECHNICAL FIELD
[0001] The present application relates to the technical field of non-ferrous smelting detection, more particularly, the present application relates to anode furnace smelting endpoint visual detection device and method. BACKGROUND
[0002] In the copper smelting industry, anode furnace refining is a key link for further purifying crude copper, and accurately determining the smelting endpoint is crucial to guarantee the quality and production efficiency of refined copper. If the endpoint is determined too early, impurities in the copper liquid cannot be fully removed, resulting in substandard product purity, and if the endpoint is determined too late, not only will it cause energy waste, but it may also affect the service life of the furnace body and increase production costs.
[0003] Currently, anode furnace smelting endpoint detection mainly relies on traditional methods. The manual observation method relies on the experience of the operator to determine the endpoint by observing the flame color, shape, and copper liquid surface state in the furnace, but this method is greatly influenced by individual experience, environmental light, high-temperature smoke, and other factors, is highly subjective, has low accuracy, and is highly labor-intensive. Detection methods based on physical and chemical parameters, such as monitoring current, voltage, temperature, and gas composition, achieve objective detection to some extent, but the correlation between these parameters and the smelting endpoint is not absolute, and other factors can easily cause misjudgment, making it difficult to fully and accurately reflect the smelting process.
[0004] With the development of computer vision technology, its application in anode furnace smelting endpoint detection has become a research hotspot. However, existing visual detection technologies have problems such as poor image quality affected by harsh environments, inaccurate feature extraction, and poor adaptability of detection models, which cannot meet the demand for high-precision, real-time detection of the smelting endpoint under complex working conditions such as high temperature, high dust, and strong electromagnetic interference in the anode furnace smelting process. Therefore, we designed an anode furnace smelting endpoint visual detection device and method. SUMMARY
[0005] In order to overcome the shortcomings of the prior art, the present application provides an anode furnace smelting endpoint visual detection device and method to solve the problems raised in the background art.
[0006] To achieve the above purpose, the present application provides the following technical scheme: an anode furnace smelting endpoint visual detection device, comprising a bottom plate, an annular plate is arranged below the bottom plate, a lifting mechanism is connected between the annular plate and the bottom plate on both sides above the annular plate, a high-definition industrial camera is arranged on both sides below the annular plate, and a sliding mechanism is arranged between the high-definition industrial camera and the annular plate above the high-definition industrial camera.
[0007] As a preferred technical solution of the present application, the lifting mechanism comprises an electric telescopic rod and a side plate, the side plate is fixed on both sides above the annular plate, one end of the electric telescopic rod is fixed to the bottom of the bottom plate, and the other end of the electric telescopic rod is fixed above the side plate.
[0008] As a preferred technical solution of the present invention, mounting holes are provided on both sides of the two side walls of the bottom plate.
[0009] As a preferred technical solution of the present invention, the sliding mechanism includes a convex slide groove and a convex slide plate. The convex slide groove is opened on both sides directly below the annular plate. One end of the convex slide plate is fixedly installed above the high-definition industrial camera, and the upper part of the convex slide plate is slidably arranged in the convex slide groove.
[0010] As a preferred technical solution of the present invention, grooves are provided on both sides of the upper side of the convex slide plate, and balls are movably connected in the grooves.
[0011] As a preferred technical solution of the present invention, both side walls of the convex slide grooves are provided with sliding holes, a horizontal plate is fixed to the side wall of one of the convex skateboards, one end of the horizontal plate passes through the two sliding holes and is fixed to the side wall of the other convex skateboard, a U-shaped plate is fixed above the annular plate, a motor is fixedly installed on the side wall above the U-shaped plate, a rotating shaft is connected to the bottom of the motor, and one end of the rotating shaft is fixed to the horizontal plate.
[0012] The present invention also provides a method for visually detecting the end point of anode furnace smelting, comprising the following steps: S: The bottom plate is fixedly installed at the anode furnace smelting site that needs to be inspected for use. Images of the anode furnace smelting can be collected through two high-definition industrial cameras. The annular plate can be driven up and down by controlling the electric telescopic rod, so that the height of the two high-definition industrial cameras can be adjusted. By starting the motor, the motor can drive the horizontal plate to rotate. The rotation of the horizontal plate can drive the two convex slides to slide in the convex slide groove. The sliding of the two convex slides can adjust the horizontal position of the two high-definition industrial cameras.
[0013] S2: The collected original image is transmitted to the image processing server, and the image brightness is adjusted using an adaptive histogram equalization algorithm to enhance the image contrast and make the surface details of the copper liquid clearer. A hybrid filtering algorithm based on a combination of median filtering and Gaussian filtering is used to reduce the noise of the image to remove noise interference caused by environmental factors such as high temperature and smoke, retaining the effective information of the image, and extracting the surface area of the copper liquid through an image segmentation algorithm, such as a segmentation method based on a combination of threshold and edge detection, to remove background interference.
[0014] S3: A detection model was constructed using a convolutional neural network (CNN) in deep learning combined with a long short-term memory (LSTM) network. Extracted color, texture, and dynamic features were used as model inputs, with the actual smelting endpoint as the label. A large amount of anode furnace smelting image data under different operating conditions, along with corresponding features and labels, was collected to form a training set. The model was trained using this training set. The model parameters were adjusted and the model structure optimized using a backpropagation algorithm, enabling the model to accurately learn the mapping relationship between different features and the smelting endpoint. During training, data augmentation techniques such as rotation, scaling, and flipping were used to expand the training data and improve the model's generalization capabilities.
[0015] S4: The real-time image features after preprocessing and feature extraction are input into the trained detection model. The model outputs the probability value that the current smelting state is close to the end point. A reasonable threshold is set. When the output probability value exceeds the threshold, it is determined that the anode furnace smelting has reached the end point, and a prompt signal is sent to the operator. At the same time, the detection results are transmitted to the production control system for subsequent operations. During the detection process, new image data and actual smelting end point data are continuously collected, and the model is regularly updated and optimized to ensure the detection accuracy of the model under different production conditions.
[0016] Compared with the prior art, the present invention has the following beneficial effects: 1. The present invention can capture images of anode furnace smelting through two high-definition industrial cameras. By controlling the electric telescopic rod, the annular plate can be driven to move up and down, thereby adjusting the height of the two high-definition industrial cameras.
[0017] 2. The present invention starts the motor, which can drive the horizontal plate to rotate. The rotation of the horizontal plate can drive the two convex slides to slide in the convex slide groove. The sliding of the two convex slides can adjust the horizontal positions of the two high-definition industrial cameras. BRIEF DESCRIPTION OF THE DRAWINGS
[0018] Figure 1 This is a schematic diagram of the main structure of the anode furnace smelting endpoint visual detection device and method of the present invention; Figure 2 Schematic diagram of the top view of the anode furnace smelting endpoint visual detection device and method of the present invention; Figure 3 It is an enlarged schematic diagram of the structure of part A of the anode furnace smelting endpoint visual detection device and method of the present invention.
[0019] In the figure: 1. Base plate; 2. Ring plate; 3. Side plate; 4. Electric telescopic rod; 5. Mounting hole; 6. Convex slide groove; 7. Convex slide plate; 8. High-definition industrial camera; 9. Horizontal plate; 10. U-shaped plate; 11. Motor; 12. Groove; 13. Ball bearing. DETAILED DESCRIPTION
[0020] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.
[0021] like Figures 1 to 3 As shown, the present invention provides a visual detection device for the end point of anode furnace smelting, including a base plate 1, mounting holes 5 are opened on both sides of the two side walls of the base plate 1, an annular plate 2 is arranged directly below the base plate 1, and a lifting mechanism is connected between the two sides above the annular plate 2 and the base plate 1, high-definition industrial cameras 8 are arranged on both sides directly below the annular plate 2, and a sliding mechanism is provided between the top of the high-definition industrial camera 8 and the annular plate 2, and the two high-definition industrial cameras 8 can be used to capture images of the anode furnace smelting, and the lifting mechanism can drive the annular plate 2 to move up and down, so that the height of the two high-definition industrial cameras 8 can be adjusted.
[0022] Among them, the lifting mechanism includes an electric telescopic rod 4 and a side panel 3. The side panel 3 is fixed on both sides above the annular plate 2. One end of the electric telescopic rod 4 is fixed at the bottom of the base plate 1, and the other end of the electric telescopic rod 4 is fixed above the side panel 3. By controlling the electric telescopic rod 4, the annular plate 2 can be driven to move up and down, so that the height of the two high-definition industrial cameras 8 can be adjusted.
[0023] Among them, the sliding mechanism includes a convex slide 6 and a convex slide 7. The convex slide 6 is opened on both sides directly below the annular plate 2. One end of the convex slide 7 is fixedly installed above the high-definition industrial camera 8, and the upper part of the convex slide 7 is slidably arranged in the convex slide 6. Grooves 12 are opened on both sides above the convex slide 7, and balls 13 are movably connected in the grooves 12. The balls 13 can make the convex slide 7 slide more smoothly in the convex slide 6.
[0024] Among them, the side walls of the two convex chutes 6 are provided with sliding holes, and a horizontal plate 9 is fixed to the side wall of one of the convex slides 7. One end of the horizontal plate 9 passes through the two sliding holes and is fixed to the side wall of the other convex slide 7. A U-shaped plate 10 is fixed above the annular plate 2, and a motor 11 is fixedly installed on the side wall above the U-shaped plate 10. A rotating shaft is connected below the motor 11, and one end of the rotating shaft is fixed on the horizontal plate 9. By starting the motor 11, the motor 11 can drive the horizontal plate 9 to rotate. The rotation of the horizontal plate 9 can drive the two convex slides 7 to slide in the convex chute 6. The sliding of the two convex slides 7 can adjust the horizontal position of the two high-definition industrial cameras 8.
[0025] The working principle and usage process of the present invention are as follows: two high-definition industrial cameras 8 can be used to capture images of the anode furnace smelting, and the annular plate 2 can be driven up and down by controlling the electric telescopic rod 4, so that the height of the two high-definition industrial cameras 8 can be adjusted. By starting the motor 11, the motor 11 can drive the horizontal plate 9 to rotate, and the rotation of the horizontal plate 9 can drive the two convex slides 7 to slide in the convex slide groove 6, and the sliding of the two convex slides 7 can adjust the horizontal position of the two high-definition industrial cameras 8.
[0026] The present invention also provides a method for visually detecting the endpoint of anode furnace smelting, comprising the following steps: S1: The base plate 1 is fixedly installed at the anode furnace smelting site that needs to be inspected for use. The image of the anode furnace smelting can be collected by two high-definition industrial cameras 8. The annular plate 2 can be driven up and down by controlling the electric telescopic rod 4, so that the height of the two high-definition industrial cameras 8 can be adjusted. By starting the motor 11, the motor 11 can drive the horizontal plate 9 to rotate. The rotation of the horizontal plate 9 can drive the two convex slides 7 to slide in the convex slide groove 6. The sliding of the two convex slides 7 can adjust the horizontal position of the two high-definition industrial cameras 8.
[0027] S2: The collected original image is transmitted to the image processing server, and the image brightness is adjusted using an adaptive histogram equalization algorithm to enhance the image contrast and make the surface details of the copper liquid clearer. A hybrid filtering algorithm based on a combination of median filtering and Gaussian filtering is used to reduce the noise of the image to remove noise interference caused by environmental factors such as high temperature and smoke, retaining the effective information of the image, and extracting the surface area of the copper liquid through an image segmentation algorithm, such as a segmentation method based on a combination of threshold and edge detection, to remove background interference.
[0028] S3: A detection model was constructed using a convolutional neural network (CNN) in deep learning combined with a long short-term memory (LSTM) network. Extracted color, texture, and dynamic features were used as model inputs, with the actual smelting endpoint as the label. A large amount of anode furnace smelting image data under different operating conditions, along with corresponding features and labels, was collected to form a training set. The model was trained using this training set. The model parameters were adjusted and the model structure optimized using a backpropagation algorithm, enabling the model to accurately learn the mapping relationship between different features and the smelting endpoint. During training, data augmentation techniques such as rotation, scaling, and flipping were used to expand the training data and improve the model's generalization capabilities.
[0029] S4: The real-time image features after preprocessing and feature extraction are input into the trained detection model. The model outputs the probability value that the current smelting state is close to the end point. A reasonable threshold is set. When the output probability value exceeds the threshold, it is determined that the anode furnace smelting has reached the end point, and a prompt signal is sent to the operator. At the same time, the detection results are transmitted to the production control system for subsequent operations. During the detection process, new image data and actual smelting end point data are continuously collected, and the model is regularly updated and optimized to ensure the detection accuracy of the model under different production conditions.
[0030] It should be noted that, in this document, relational terms such as first and second, etc., are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the terms "comprises," "comprising," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that includes a list of elements includes not only those elements but also other elements not explicitly listed, or elements inherent to such process, method, article, or apparatus.
[0031] While embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions, and variations may be made to these embodiments without departing from the principles and spirit of the invention, and that the scope of the invention is defined by the appended claims and their equivalents.
Claims
1. A visual detection device for the end point of anode furnace smelting, comprising a bottom plate (1), characterized in that: An annular plate (2) is provided directly below the bottom plate (1), a lifting mechanism is connected between the two sides above the annular plate (2) and the bottom plate (1), high-definition industrial cameras (8) are provided on both sides directly below the annular plate (2), and a sliding mechanism is provided between the top of the high-definition industrial camera (8) and the annular plate (2).
2. The anode furnace smelting endpoint visual detection device according to claim 1, characterized in that: The lifting mechanism comprises an electric telescopic rod (4) and side panels (3), wherein the side panels (3) are fixed on both sides above the annular plate (2), one end of the electric telescopic rod (4) is fixed to the bottom of the base plate (1), and the other end of the electric telescopic rod (4) is fixed above the side panels (3).
3. The anode furnace smelting endpoint visual detection device according to claim 1, characterized in that: Mounting holes (5) are provided on both sides of the two side walls of the bottom plate (1).
4. The anode furnace smelting endpoint visual detection device according to claim 2, characterized in that: The sliding mechanism comprises a convex chute (6) and a convex slide plate (7), wherein the convex chute (6) is provided on both sides directly below the annular plate (2), one end of the convex slide plate (7) is fixedly mounted above the high-definition industrial camera (8), and the upper portion of the convex slide plate (7) is slidably arranged in the convex chute (6).
5. The anode furnace smelting endpoint visual detection device according to claim 4, characterized in that: Grooves (12) are provided on both sides of the upper side of the convex slide plate (7), and balls (13) are movably connected in the grooves (12).
6. The anode furnace smelting endpoint visual detection device according to claim 4, characterized in that: The side walls of the two convex chute (6) are both provided with sliding holes, a transverse plate (9) is fixed to the side wall of one of the convex slide plates (7), one end of the transverse plate (9) passes through the two sliding holes and is fixed to the side wall of the other convex slide plate (7), a U-shaped plate (10) is fixed above the annular plate (2), a motor (11) is fixedly installed on the side wall above the U-shaped plate (10), a rotating shaft is connected to the bottom of the motor (11), and one end of the rotating shaft is fixed to the transverse plate (9).
7. A method for visually detecting the endpoint of anode furnace smelting, using the device according to claim 6, characterized in that: The method includes the following steps: S1: The bottom plate (1) is fixedly mounted on the anode furnace smelting site to be inspected for use, and an image of the anode furnace smelting is collected by two high-definition industrial cameras (8). The annular plate (2) is driven to move up and down by controlling the electric telescopic rod (4), thereby adjusting the height of the two high-definition industrial cameras (8). The motor (11) is started, and the motor (11) drives the horizontal plate (9) to rotate. The horizontal plate (9) rotates and drives the two convex slides (7) to slide in the convex chute (6). The two convex slides (7) slide to adjust the horizontal position of the two high-definition industrial cameras (8); S2: The original image collected is transferred to an image processing server, and an adaptive histogram equalization algorithm is utilized to adjust the brightness of the image to enhance the image contrast. A hybrid filtering algorithm based on median filtering and Gaussian filtering is adopted to reduce noise on the image to remove the noise interference caused by environmental factors. The image effective information is retained. By an image segmentation algorithm, the copper liquid surface area is extracted to remove background interference. S3: A detection model is constructed using a convolutional neural network in deep learning combined with a long short-term memory network. The extracted color features, texture features, and dynamic features are used as model inputs. The actual smelting endpoint is used as a label. A large amount of anode furnace smelting image data under different working conditions and the corresponding features and labels are collected to form a training set. The model is trained using the training set. The model parameters are adjusted through the back-propagation algorithm to optimize the model structure so that the model can accurately learn the mapping relationship between different features and the smelting endpoint. During the training process, data enhancement techniques such as rotation, scaling, and flipping are used to expand the training data and improve the generalization ability of the model. S4: The real-time image features after preprocessing and feature extraction are input into the trained detection model. The model outputs the probability value that the current smelting state is close to the end point. A reasonable threshold is set. When the output probability value exceeds the threshold, it is determined that the anode furnace smelting has reached the end point, and a prompt signal is sent to the operator. At the same time, the detection results are transmitted to the production control system for subsequent operations. During the detection process, new image data and actual smelting end point data are continuously collected, and the model is regularly updated and optimized to ensure the detection accuracy of the model under different production conditions.