Caked ore intelligent processing device and method based on enhanced visual identification trigger algorithm

Abstract

The invention discloses an intelligent caked ore processing device and method based on an enhanced visual identification trigger algorithm, and belongs to the technical field of metallurgical industry ore processing. The device comprises an ore disassembling prying device and a vibrating screen, the ore disassembling prying device comprises a power driving shaft combination device and a driven shaft combination device, an industrial camera and a controller are installed on the outer surface of the power driving shaft combination device, a ring hook and a cutting blade are installed on the outer surface of the driven shaft combination device, and the industrial camera right faces a vibrating screen on the vibrating screen. One end of the controller is connected with the industrial camera, the other end is connected with the signal processor, the signal processor is connected with the variable-speed motor, and the variable-speed motor is connected with the power driving shaft combination device. The operation speed of the device is controlled and adjusted through a series of methods, and large caked ore is pried or disassembled through the ring hook and the cutting blade. According to the multi-stage anti-deformation composite structure vibrating screen, the blocking problem existing when the multi-stage anti-deformation composite structure vibrating screen is used for treating large-size caked ore is effectively solved.

Description

Technical Field

[0001] The present invention relates to an intelligent processing device and method for agglomerated ore based on an enhanced visual recognition trigger algorithm, belonging to the technical field of ore processing in the metallurgical industry. Background Art

[0002] The processing equipment required for production exists in various industries, such as equipment manufacturing, metallurgical production, and parts manufacturing. This demanding nature requires pretreatment under varying working conditions. Currently, much of this pretreatment is still performed manually. This processing is often accompanied by complex conditions, including cramped production sites, high noise levels, and high levels of dust. These complex working conditions create long-term workloads for on-site workers, resulting in not only high workload intensity but also adverse effects on their comfort and health.

[0003] Due to the limited space in the ore crushing section, general-purpose machinery that meets these requirements is unavailable on the market, necessitating custom production based on specific needs. Furthermore, the ore raw materials, depending on their primary element grade, moisture content, and the content of various other metal elements, impose unique requirements on the crushing structure, making traditional mechanical equipment inadequate.

[0004] In addition, long-term operation of ore processing equipment will lead to shortened equipment life and increased operating costs; the production site environment is complex, with a lot of ore sand dust and frequent strong vibrations, which make the sensor severely interfered with in this scenario, and the signal reception is inaccurate and cannot be used; and the geographical location of the production site has poor lighting conditions and frequent changes in sunlight, which makes image recognition difficult, thereby affecting the performance of the visual servo drive.

[0005] The related conventional technology provides a translation scraper type cleaning device (publication number CN201521139955), which can clean the vibrating screen without any obstruction on the plane. However, in the production site with strong vibration, the existing translation type cleaning device is not suitable for the multi-stage anti-deformation composite structure vibrating screen with anti-deformation rods in the middle, and the purpose of prying the ore cannot be achieved. Summary of the Invention

[0006] Based on this, it is necessary to propose an intelligent processing device and method for agglomerated ore based on an enhanced visual recognition trigger algorithm that can adapt to the production process of various vibrating screens with different shapes and structures in order to solve the above problems. The device and method can process agglomerated ore on a multi-stage anti-deformation composite structure vibrating screen based on an enhanced visual recognition trigger algorithm. The device and method can be used to process agglomerated ore on a multi-stage anti-deformation composite structure vibrating screen, thereby solving the problems of blockage of large-sized agglomerated ore on the surface of a multi-stage anti-deformation composite structure vibrating screen in the background technology, difficulty in cleaning the screen, high health risks for workers, and affected production cycles.

[0007] The technical solution adopted by the present invention is: an intelligent processing device for agglomerated ore based on an enhanced visual recognition trigger algorithm, comprising an ore disassembling and prying device 1 and a vibrating screen, wherein the ore disassembling and prying device 1 is arranged above the vibrating screen through a supporting device, and the ore disassembling and prying device 1 comprises a power drive shaft assembly device and a driven shaft assembly device, wherein the power drive shaft assembly device and the driven shaft assembly device are connected through a transmission chain 8.12, an industrial camera 7.10 and a controller 7.9 are installed on the outer surface of the power drive shaft assembly device, and a ring hook is installed on the outer surface of the driven shaft assembly device. 4.8, a cutting blade 4.9, a variable speed motor and a signal processor 8.11 are installed on the support device, the industrial camera 7.10 is facing the vibrating screen on the vibrating screen, and the controller 7.9 includes an image processing module, a data analysis module, and a motion control module. The input end of the data analysis module is connected to the image processing module, and the output end is connected to the motion control module. The image processing module is connected to the industrial camera 7.10. The input end of the signal processor 8.11 is connected to the motion control module, and the output end is connected to the variable speed motor, and the variable speed motor is connected to the power drive shaft assembly device.

[0008] Specifically, the power drive shaft assembly device includes a drive support stepped shaft 6.3, a transmission sprocket I 6.1, a hollow rod sleeve 6.2, a transmission sprocket II 6.4, a bearing 6.5, a shaft elastic retaining ring 6.8, and a hook-head wedge key 6.6; the hollow rod sleeve 6.2 is loosely sleeved on the outside of the drive support stepped shaft 6.3 and is fixed thereto; the drive support stepped shaft 6.3 is connected to the transmission sprocket II 6.4 and the transmission sprocket I 6.1 at both ends respectively through the hook-head wedge key 6.6; the fixing plate 6.7 is fixedly connected to the support device. Two sets of bearings 6.5 are installed between the fixed plate 6.7 and the driving support stepped shaft 6.3. One end of each bearing 6.5 is rotatably connected to the step of the fixed plate 6.7 through the step of the supporting stepped shaft 6.3, and the other end is rotatably connected to the shaft by an elastic retaining ring 6.8. A stepped shaft pulley is installed at one end of the driving support stepped shaft 6.3. The stepped shaft pulley is connected to the transmission pulley on the output shaft of the variable speed motor through a belt drive. The transmission sprocket I 6.1 and the transmission sprocket II 6.4 are respectively connected to the driven shaft assembly device through a transmission chain 8.12.

[0009] Specifically, the driven shaft assembly device includes a driven support shaft 4.6, a driven sprocket I 4.1, a driven sprocket II 4.4, a prying roller 4.2, a self-locking fixed connecting rod 4.3, a shaft elastic circlip II 4.5, a driven support rod 4.7, a ring hook 4.8, and a cutting blade 4.9. The prying roller 4.2 is sleeved on the driven support shaft 4.6, and the driven sprocket I 4.1 and the driven sprocket II 4.4 are installed at both ends of the driven support shaft 4.6 through the shaft elastic circlip 4.5. I 4.1 and driven sprocket II 4.4 are connected to the prying roller 4.2 via a self-locking fixed connecting rod 4.3. A ring hook 4.8 is mounted on the prying roller 4.2. A cutting blade 4.9 is mounted in a slot of the prying roller 4.2. Both ends of the driven support shaft 4.6 are fixedly connected to the driven support rod 4.7. The driven support rod 4.7 is fixedly connected to the supporting device. The driven sprocket I 4.1 and the driven sprocket II 4.4 are respectively connected to the power drive shaft assembly via a transmission chain 8.12.

[0010] Preferably, the prying roller 4.2 is circular, and cutting blade mounting grooves and overhanging cantilevers are evenly provided on the circumference of the prying roller 4.2. A cutting blade 4.9 is fixed in the cutting blade mounting groove, and a group of overhanging cantilevers are fixed between the two blade mounting grooves, and a ring hook 4.8 is fixed at the end of the overhanging cantilever.

[0011] More preferably, three cutting blade mounting grooves and three groups of overhanging cantilevers are evenly arranged on the circumference of the prying roller 4.2, the circumferential angles of the three cutting blades 4.9 differ by 120 degrees from each other, the axial spacing of the three groups of overhanging cantilevers is 1 / 3 of the width of the vibrating screen, the circumferential angle difference is 120 degrees, the vibrating screen has 15 rows of slots, each group of overhanging cantilevers includes five overhanging cantilevers, and a ring hook mounting position is provided at the end of each overhanging cantilever. The angle difference between the three groups of overhanging cantilevers is 120°, and the lateral coordinate difference is 325mm and 650mm respectively. The ring hook 4.8 is a long strip with wide sides and narrow in the middle.

[0012] A processing method of an intelligent processing device for agglomerated ore based on an enhanced visual recognition trigger algorithm comprises the following steps:

[0013] Step 1: Perform the following processing through the image processing module in controller 7.9:

[0014] Step 1.1: Use the industrial camera 7.10 to capture live images and mark the target area of ​​the image:

[0015] Step 1.2: Use the industrial camera 7.10 to capture and process the on-site image. The marked area is processed into a transparent hollow area, and the hollow area image is fused with the actual collected image to ensure that the pixel values ​​of the transparent hollow area are from the actual collected image. The change in pixel values ​​is used to determine whether there is agglomerated ore.

[0016] Step 1.3: Perform image enhancement processing on the fused image by performing target texture restoration through the frequency domain analysis filtering denoising method based on neural network;

[0017] Step 1.4: De-noise the enhanced image using a bidirectional denoising diffusion model improved based on the dual attention mechanism, and then perform digital preprocessing on the denoised image.

[0018] Step 2: The pre-processed image is processed as follows by the data analysis module in the controller 7.9:

[0019] Step 2.1: Use the threshold segmentation method to segment the preprocessed image;

[0020] Step 2.2: Calculate the area ratio of the agglomerated ore in the image at the current moment and the area ratio of the agglomerated ore in the image at the previous moment for the segmented regions that meet the target conditions in the segmented image; based on the calculated area ratio of the two adjacent moments and the current speed, generate control signals for the speed adjustment corresponding to different gears;

[0021] Step 3: The motion control module in the controller 7.9 transmits the control signal generated by the data analysis module to the signal processor 8.11;

[0022] Step 4: Signal processor 8.11 adjusts the speed of the variable speed motor according to the control signal. The variable speed motor drives the power drive shaft assembly to rotate. The power drive shaft assembly drives the driven shaft assembly to rotate. When the driven shaft assembly rotates, it drives the ring hook 4.8 and the cutting blade 4.9 to pry and / or cut the ore blocks with unqualified sizes on the surface of the vibrating screen.

[0023] The details of step 1.1 are as follows:

[0024] Step 1.1.1: Determine the installation location of the industrial camera 7.10 according to the actual situation and install the industrial camera 7.10;

[0025] Step 1.1.2: Use industrial camera 7.10 to take pictures of the upper surface of the vibrating screen;

[0026] Step 1.1.3: Based on the actual production situation and on-site cleaning experience, a normal distribution function is constructed, and the probability distribution function is used to estimate the location of the area where the agglomerated ore may fall on the vibrating screen;

[0027] Step 1.1.4: Calibrate and divide the target cleaning area of ​​the image read by the camera, that is, the position of the vibrating screen holes to clarify the image target area;

[0028] The step 1.3 is as follows: First, a physical guided temporal neural network and real-time image acquisition are used to construct the light incidence model S of the production workshop on sunny and cloudy days. s (x,y) and S c (x, y) and the actual production live acquisition model P(x, y), and then use the live acquisition model P(x, y) and the weather light incidence model S s (x,y) and S c (x,y) Perform light compensation, polarization stripe noise removal, and image restoration on the fused image under complex environments to obtain an enhanced clear image of the target under normal and stable conditions;

[0029] The step 1.4 is specifically as follows: first, the enhanced fused image is subjected to hierarchical modular fractal through a modular fractal module and a regional attention mechanism, so as to distinguish between clear areas without noise and unclear areas with much noise; noise is continuously added to different areas of the enhanced fused image in varying degrees through forward diffusion and a position attention mechanism; the forward diffusion process is then reversed through the PDF parameters in the forward process predicted by deep learning; and finally, a clear image is restored from the noisy image, thereby achieving a first denoising process for the enhanced fused image; and finally, the denoised image is subjected to digital preprocessing.

[0030] The step 2.2 is as follows:

[0031] Step 2.2.1: After image segmentation, the image gradient of the calibrated hollowed-out area in the fused image is calculated to detect changes in pixel values. The area with a larger gradient amplitude is the edge of the image. Then, a binary edge image is output. The foreground edge contour is approximated by the Monte Carlo method, and the area of ​​the irregular shape is calculated to approximate the area of ​​the contour area enclosed by the edge features after the image segmentation method:

[0032] Area calculation:

[0033]

[0034] Where n is the number of contour areas enclosed by the edge features after image segmentation processing at the current time t, i is the i-th contour area enclosed by the edge features after image segmentation processing, i = 1, 2, 3, ... n, b i with a i They are the upper and lower bounds of the independent variable x in the i-th contour area surrounded by the edge features of the irregular area after image segmentation at the current time t, x p is in the interval [a i ,b i] is a random number uniformly distributed on the image, N is the i-th contour area surrounded by the edge features of the irregular area after image segmentation at the current moment t in the interval [a i ,b i ], m is the number of contour areas enclosed by the edge features processed at the previous moment (t-1), j is the jth contour area enclosed by the edge features after image segmentation processing, j = 1, 2, 3, ... m, b j with a j They are the upper and lower bounds of the independent variable x in the j-th contour area surrounded by the edge features of the irregular area processed at the previous moment (t-1), x q is in the interval [a j ,b j ] is a random number uniformly distributed on the surface, M is the jth contour area surrounded by the edge features of the irregular area processed at the previous moment (t-1) in the interval [a j ,b j ] the number of random sampling points, s i is the approximate area of ​​the i-th contour region enclosed by the edge features processed at the current time t, s j is the approximate area of ​​the jth contour region enclosed by the edge features after segmentation at the previous moment (t-1), S (t) Extract the feature area at the current moment t, S (t-1) is the feature area extracted at the previous moment (t-1);

[0035] Step 2.2.2: The calculated different area ratios correspond to different gear adjustment speeds. The gear change is determined based on the trigger conditions:

[0036] The ratio of the area at the current moment to the previous moment:

[0037]

[0038] Calculation of the area ratio of the contour area in the entire area at the previous moment:

[0039]

[0040] Where: S * is the ratio of the area at the current moment to the area at the previous moment, A is the area of ​​the entire image, Calculate the area ratio of the contour area at the current moment to the entire image area. Calculate the area ratio of the contour area at the current moment to the entire image area. Calculate the proportion of the contour area at the previous moment to the entire image area;

[0041] 1) If S is calculated *The range is: The equipment runs at the first gear speed, y=1;

[0042] 2) If S is calculated * The range is: The equipment runs at the second speed, y=2;

[0043] 3) If S is calculated * The range is: The equipment runs at the third speed, y=3;

[0044] Area ratio S * The conversion formula corresponding to the theoretical speed v is:

[0045]

[0046] Where: v max(y) For S * The theoretical maximum speed value of the corresponding gear within the change range, v min(y) For S * The theoretical minimum speed value of the corresponding gear within the change range, v (t) Calculate the area ratio of the contour area at the current moment to the entire image area and the ratio of the area at the current moment to the previous moment S * The corresponding theoretical value of velocity is: For S * The maximum value within the variation range, For S * The minimum value within the change range, y = 1, 2, 3 means the equipment gear is adjusted to gear 1, gear 2, gear 3, ε1 = v min(2) -v max(1) ,ε2=v min(3) -v max(2) , ε1>0,ε2>0, ε1 and ε2 are both constants;

[0047] In the inequality, The threshold value of the state variable for equipment gear switching is automatically selected by using a genetic algorithm based on historical data and field production experience.

[0048] The Step 4 specifically includes:

[0049] Step 4.1: The signal processor 8.11 sends a control signal to the variable speed motor. If the variable speed motor is not in working state, it is powered on and the speed is the gear speed generated by the controller. If the variable speed motor is in working state, it is determined whether the current speed gear is consistent with the gear speed required by the controller. If they are consistent, it continues to work. If not, it is adjusted to be consistent.

[0050] The power of the variable speed motor is transmitted to the driving support stepped shaft 6.3 through the transmission of the driving pulley and the stepped shaft pulley. The driving support stepped shaft 6.3 is connected with the driving sprocket I 6.1 and the driving sprocket II 6.4 by a hook-head wedge key 6.6 to realize the coaxial rotation of the driving support stepped shaft 6.3, the driving sprocket I 6.1 and the driving sprocket II 6.4. Then, the driven sprocket II 4.4 and the driven sprocket I 4.1 on the driven shaft rotate following the rotation of the driving sprocket II 6.4 and the driving sprocket 6.1 on the driving support shaft 6.3 through the transmission chain 8.12. Lock the fixed connecting rod 4.3 to connect the prying roller 4.2 with the driven sprocket II 4.4 and the driven sprocket I 4.1, so that the prying roller 4.2 and the driven sprocket II 4.4 and the driven sprocket I 4.1 on the driven shaft can rotate coaxially. Then determine the rotation direction and hang the ring hook 4.8 on the outward cantilever of the prying roller 4.2. As the prying roller 4.2 rotates, the ring hook 4.8 rotates to push out the ore blocks in the screen. If large pieces of agglomerated ore are encountered, the large pieces of agglomerated ore are first dismantled by rotating the cutting blade 4.9, and then the ore blocks in the screen are pushed out by rotating the ring hook 4.8.

[0051] Compared with the prior art, the beneficial effects of the present invention are as follows: the present invention proposes an intelligent processing device and method for agglomerated ore based on an enhanced visual recognition trigger algorithm. Compared with the original manual operation process, it can effectively improve production efficiency and reduce the labor intensity of workers. In addition, the device is equipped with an intelligent visual speed control system, which calibrates the upper surface image of the vibrating screen in advance, and uses the hollowed-out area after calibration as the identification target area. The speed of the hardware structure equipment can be controlled according to the real-time working conditions, effectively maintaining the life of the equipment. Moreover, this device can be installed independently of the production line, which can effectively avoid stopping work and production to improve the site, and can further improve the overall benefits on the basis of ensuring the existing benefits. In addition, the application of the mineral material crushing structure device provides a new direction for production and processing, and the device is simple to install, reliable in principle, and low in cost. It can make the mineral blocks fully crushed and then transported back, greatly reducing labor intensity. By controlling the operating equipment with reference to the gear speed, the operating and maintenance costs are greatly reduced. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] Figure 1 This is a schematic diagram of the external structure of the vibrating screen of the present invention;

[0053] Figure 2 Schematic diagram of the application scenario of the present invention;

[0054] Figure 3 This is an overall schematic diagram of the combination of the ore disassembling and prying device and the vibrating screen of the present invention;

[0055] Figure 4This is a schematic diagram of the overall structure of the driven shaft assembly device in the ore disassembling and prying device of the present invention;

[0056] Figure 5 It is a structural diagram of the ring hook in the present invention;

[0057] Figure 6 It is a schematic structural diagram of the cutting blade in the present invention;

[0058] Figure 7 This is a schematic diagram of the installation position of the ring hook and cutting blade on the prying roller of the present invention;

[0059] Figure 8 This is a schematic half-section diagram of the overall structure of the power drive shaft assembly device of the present invention;

[0060] Figure 9 This is the hardware installation location of the industrial camera and controller in the visual part of the power drive shaft assembly structure of the present invention;

[0061] Figure 10 This is the hardware installation location of the visual signal processor in the structure of the power drive shaft assembly device of the present invention;

[0062] Figure 11 Schematic diagram of the control principle of the software adaptive visual speed regulation of the present invention;

[0063] Figure 12 It is the image area required for recognition by the industrial camera in the present invention, where A is the camera recognition area and B is the marking area.

[0064] The numbers in the figure are: 1 - ore disassembling and prying device, 2 - ore conveying and crushing device, 3 - ore particle transport and recovery device, 4.1 - driven sprocket I, 4.2 - prying roller, 4.3 - self-locking fixed connecting rod, 4.4 - driven sprocket II, 4.5 - shaft elastic ring, 4.6 - driven support shaft, 4.7 - driven support rod, 4.8 - hook, 4.9 - cutting blade, 4.8-1 - first hook installation position, 4.8-2 - second hook installation position, 4.8-3 - third hook installation position. Installation position, 4.9-1—first cutting blade installation slot, 4.9-2—second cutting blade installation slot, 4.9-3—third cutting blade installation slot, 6.1—drive sprocket I, 6.2—hollow rod sleeve, 6.3—drive support stepped shaft, 6.4—drive sprocket II, 6.5—bearing, 6.6—hook-type wedge key, 6.7—fixing plate, 6.8—elastic retaining ring for shaft, 7.9—controller, 7.10—industrial camera, 8.11—signal processor, 8.12—drive chain. DETAILED DESCRIPTION

[0065] The invention will be further described below with reference to the accompanying drawings and embodiments, but the content of the present invention is not limited to the scope of the drawings.

[0066] Example 1: Figure 1-12 As shown, an intelligent processing device for agglomerated ore based on enhanced visual recognition triggering algorithm includes an ore disassembly and prying device 1, a vibrating screen (such as Figure 1 As shown, the ore dismantling and prying device 1 is arranged above the vibrating screen through a supporting device (it should be noted that, according to the size requirements of the site layout, attention should be paid to the distance between the ore dismantling and prying device 1 and the vibrating screen). The ore dismantling and prying device 1 includes a power drive shaft assembly and a driven shaft assembly. The power drive shaft assembly and the driven shaft assembly are connected by a transmission chain 8.12. The industrial camera 7.10 and the controller 7.9 are mounted on the outer surface of the power drive shaft assembly. The ring hook 4.8 and the cutting blade 4.9 are mounted on the outer surface of the driven shaft assembly. A variable speed motor and a signal processor 8.11 are mounted on the supporting device. The industrial camera 7.10 is directly facing the vibrating screen on the vibrating screen. The controller 7.9 includes an image processing module, a data analysis module, and a motion control module. The input end of the data analysis module is connected to the image processing module, and the output end is connected to the motion control module. The image processing module is connected to the industrial camera 7.10. The input end of the signal processor 8.11 is connected to the motion control module, and the output end is connected to the variable speed motor. The variable speed motor is connected to the power drive shaft assembly.

[0067] The device is driven by a variable-speed motor, utilizing belt and chain drives for power transmission. An external structural frame provides fixed load-bearing support, ensuring stable operation of the ore levering device without being affected by strong vibrations. The device's speed is controlled by visually identifying actual production conditions. This invention addresses the processing of agglomerated ores using a multi-stage, anti-deformation composite vibrating screen. An intelligent visual speed control system then performs on-site image capture, data analysis, and control output, ultimately enabling intelligent speed regulation of the structural equipment.

[0068] like Figure 2 As shown, in actual application, the overall production line includes an ore dismantling and prying device 1, an ore conveying and crushing device 2, and a mineral particle recovery device 3; the ore dismantling and prying device 1 is fixed with a support device to bear the weight, ensuring the operational stability of the ore dismantling and prying device 1, the ore conveying and crushing device 2 adopts a serrated rolling element extrusion mechanism to achieve progressive crushing of the ore, and the mineral particle recovery device 3 is equipped with an efficient conveying mechanism to accurately return the processed mineral particles to the screening system.

[0069] The ore disassembly and prying device 1 of the present invention can be installed independently of the production line. It can pry the rollers in a circular motion, driving the ring hook 4.8 or cutting blade 4.9 to remove or cut agglomerated ore from the vibrating screen, thereby cleaning large multi-stage, deformation-resistant composite structure vibrating screens under complex working conditions. This can effectively avoid downtime for on-site improvements, further improving overall profitability while maintaining existing benefits. The ore disassembly and prying device 1 of the present invention has a wide range of target users and enhanced functionality. It can be used for the disassembly and prying of various types of mineral particles. The mineral particle sizes can be divided into three types. Small particles (length and width ≤ 45mm) can pass directly through the vibrating screen without the need for prying and crushing; large agglomerated mineral particles (45mm ≤ length and width ≤ 150mm) are retained on the surface of the vibrating screen and need to be forcibly peeled off by the ore disassembly and prying device 1 and transported to the crushing process for particle size optimization; super-large agglomerated mineral particles (150mm ≤ length and width ≤ 300mm) need to be cut and then plucked off for crushing because the anti-deformation fixing rod in the middle of the vibrating screen has a fixed small distance from the vibrating screen surface. Compared with the traditional vibrating screen cleaning structure, the ore disassembly and prying device 1 of the present invention eliminates the influence of the strong vibration of the vibrating screen on the ore disassembly and prying effect, and at the same time achieves seamless connection with the production line, significantly improving the ore processing effect while ensuring continuous production. In addition, the application of the ore disassembly and prying device 1 provides a new direction for production and processing. The device is simple to install, has a reliable principle, and is low in cost. It can fully crush the ore blocks and then transport them back, greatly reducing labor intensity. By controlling the operating equipment with reference to the gear speed, the operating and maintenance costs are greatly reduced.

[0070] Furthermore, the power drive shaft assembly comprises a drive support stepped shaft 6.3, a drive sprocket I 6.1, a hollow rod sleeve 6.2, a drive sprocket II 6.4, a bearing 6.5, a shaft circlip 6.8, and a hook-shaped wedge key 6.6. The hollow rod sleeve 6.2 is loosely fitted over the drive support stepped shaft 6.3 and remains stationary. The hollow rod sleeve 6.2 primarily houses the industrial camera 7.10 and controller 7.9 and does not rotate with the drive support stepped shaft 6.3. To ensure that the camera 7.10 is properly aligned with the vibrating screen, counterweights are placed symmetrically around the controller to balance the weight of the industrial camera and controller. The driving support stepped shaft 6.3 is connected to the transmission sprocket II 6.4 and the transmission sprocket I 6.1 at both ends respectively through the hook-shaped wedge key 6.6, thereby ensuring that the driving support stepped shaft 6.3 drives the transmission sprocket II 6.4 and the transmission sprocket I 6.1 to rotate coaxially. The fixed plate 6.7 is fixedly connected to the support device. The purpose of supporting the power drive shaft assembly device is achieved by the fixed connection between the fixed plate 6.7 and the support device. Two sets of bearings are installed between the fixed plate 6.7 and the driving support stepped shaft 6.3 6.5, one end of each bearing 6.5 is rotatably connected to the step of the supporting stepped shaft 6.3 and the step of the fixed plate 6.7, and the other end is rotatably connected to the shaft by an elastic retaining ring 6.8, thereby preventing the bearing 6.5 from moving left and right. A stepped shaft pulley is installed at one end of the driving supporting stepped shaft 6.3, and the stepped shaft pulley is connected to the transmission pulley on the output shaft of the variable speed motor through a belt drive. The transmission sprocket I 6.1 and the transmission sprocket II 6.4 are respectively connected to the driven shaft assembly device through a transmission chain 8.12.

[0071] Furthermore, the driven shaft assembly device includes a driven support shaft 4.6, a driven sprocket I 4.1, a driven sprocket II 4.4, a prying roller 4.2, a self-locking fixed connecting rod 4.3, a shaft elastic circlip II 4.5, a driven support rod 4.7, a ring hook 4.8, and a cutting blade 4.9. The prying roller 4.2 is sleeved on the driven support shaft 4.6, and the driven sprocket I 4.1 and the driven sprocket II 4.4 are installed at both ends of the driven support shaft 4.6 through the shaft elastic circlip 4.5 to prevent the driven sprocket I 4.1 and the driven sprocket II 4.4 moves left and right, driven sprocket I 4.1 and driven sprocket II 4.4 are connected to the prying roller 4.2 through a self-locking fixed connecting rod 4.3, a ring hook 4.8 is installed on the prying roller 4.2, and a cutting blade 4.9 is installed in the slot position of the prying roller 4.2. The two ends of the driven support shaft 4.6 are fixedly connected to the driven support rod 4.7, and the driven support rod 4.7 is fixedly connected to the supporting device. The driven sprocket I 4.1 and the driven sprocket II 4.4 are respectively connected to the power drive shaft assembly device through a transmission chain 8.12.

[0072] Furthermore, if Figure 4-7As shown, the prying roller 4.2 is circular, and cutting blade mounting grooves and overhanging cantilevers are evenly provided on the circumference of the prying roller 4.2. A cutting blade 4.9 is fixed in the cutting blade mounting groove, and a group of overhanging cantilevers is fixed between the two blade mounting grooves, and a ring hook 4.8 is fixed at the end of the overhanging cantilever.

[0073] Furthermore, if Figure 7 As shown, three cutting blade mounting grooves (respectively, the first cutting blade mounting groove 4.9-1, the second cutting blade mounting groove 4.9-2 and the third cutting blade mounting groove 4.9-3) and three groups of outriggers are evenly arranged on the circumference of the prying roller 4.2. The circumferential angles between the three cutting blades 4.9 differ by 120 degrees. The axial spacing between the three groups of outriggers is 1 / 3 of the width of the vibrating screen, and the circumferential angle difference is 120 degrees. The vibrating screen has 15 slots. There are five outriggers in a row. Each group of outriggers includes five outriggers. The end of each outrigger is provided with a hook installation position (respectively, the first hook installation position 4.8-1, the second hook installation position 4.8-2 and the third hook installation position 4.8-3). The angle difference between the three groups of outriggers is 120°, and the lateral coordinate differences are 325mm and 650mm respectively. Compared with the distribution of the entire row, this arrangement can ensure that the ore that could have fallen freely will not be blocked. The hook 4.8 is a long strip with wide sides and a narrow middle. The intermittent segmented arrangement of the outriggers and the narrow inside and wide outside hook 4.8 is more conducive to industrial production. The purpose of the wide sides and narrow middle design is that when the roller 4.2 is pried to rotate, the hook 4.8 will be restricted in its rotation angle after removing the ore, and will not hit the anti-deformation fixing rod in the middle of the multi-stage anti-deformation composite structure vibrating screen due to inertia. Compared to designs with uniform width, this design, with wider sides and a narrower center, avoids harsh noise. The purpose of the hook ring is to ensure that slightly larger ore can pass smoothly under the anti-deformation bar of the vibrating screen while still effectively prying stuck ore. This design achieves an optimal balance between screening efficiency and ore prying function by optimizing the distance between the centerline of the drive sprocket sleeve and the screen surface. The size of the 4.9-inch cutting blade is designed based on the position, shape, and size of the sprocket slot, and the cutting blade is heat-treated to improve its hardness, toughness, wear resistance, and other properties.

[0074] A processing method of an intelligent processing device for agglomerated ore based on an enhanced visual recognition trigger algorithm comprises the following steps:

[0075] Step 1: Perform the following processing through the image processing module in controller 7.9:

[0076] Step 1.1: Use the industrial camera 7.10 to capture images of the actual scene and mark the target area of ​​the image;

[0077] Step 1.2: Use the industrial camera 7.10 to capture and process the actual on-site image. The marked area is processed into a transparent hollow area, and the hollow area image is fused with the actual collected image to ensure that the pixel values ​​of the transparent hollow area come from the actual collected image. The change in pixel values ​​is used to determine whether there is agglomerated ore.

[0078] Step 1.3: Perform image enhancement processing on the fused image by performing target texture restoration through the frequency domain analysis filtering denoising method based on neural network;

[0079] Step 1.4: The enhanced image is denoised using an improved bidirectional denoising diffusion model based on the dual attention mechanism, and then the denoised image is digitally preprocessed.

[0080] The Industrial Camera 7.10 primarily identifies the number of agglomerated ore within the calibrated, demarcated hollow areas of the fused image. This number serves as a key control variable. Visual recognition technology is then used to monitor on-site conditions and capture real-time images. These images are then pre-processed to enhance subsequent processing.

[0081] Step 2: The pre-processed image is processed as follows by the data analysis module in the controller 7.9:

[0082] Step 2.1: Use the threshold segmentation method to segment the preprocessed image to distinguish the foreground and background for further analysis;

[0083] Step 2.2: Calculate the area ratio of the agglomerated ore in the image at the current moment and the area ratio of the agglomerated ore in the image at the previous moment for the segmented regions that meet the target conditions in the segmented image; based on the calculated area ratio of the two adjacent moments and the current speed, generate control signals for the speed adjustment corresponding to different gears;

[0084] Step 3: The motion control module in the controller 7.9 transmits the control signal generated by the data analysis module to the signal processor 8.11;

[0085] Step 4: Signal processor 8.11 adjusts the speed of the variable speed motor based on the control signal. The variable speed motor drives the power drive shaft assembly, which in turn drives the driven shaft assembly. The rotation of the driven shaft assembly drives the ring hook 4.8 and cutting blade 4.9, thereby prying and / or cutting off unqualified ore blocks on the surface of the vibrating screen. This ultimately achieves intelligent control of the agglomerated ore processing process, effectively resolving the clogging problem that can occur with multi-stage, anti-deformation composite vibrating screens when processing large agglomerated ore. This also reduces production costs and improves ore processing efficiency.

[0086] Furthermore, the step 1.1 is as follows:

[0087] Step 1.1.1: Based on the designed structure of the machine for processing agglomerated ores, combined with recognition accuracy, recognition efficiency, and installation convenience, confirm the installation point of the industrial camera 7.10 and install the industrial camera 7.10.

[0088] Step 1.1.2: Take a photo of the top surface of the multi-stage anti-deformation vibration screen structure at the locked installation position of industrial camera 7.10. Demarcate the target cleaning area (the location of the vibration screen holes) on the image obtained by the photo. The unmarked area in the marked and divided image is regarded as the background area.

[0089] Step 1.1.3: Based on the actual production situation and on-site cleaning experience, a normal distribution function is constructed, and the probability distribution function is used to estimate the location of the area where the agglomerated ore may fall on the vibrating screen;

[0090] Step 1.1.4: Calibrate and divide the target cleaning area of ​​the image read by the camera, that is, the position of the vibration screen holes, and clarify the target area of ​​the image.

[0091] Furthermore, the step 1.3 is specifically as follows: First, a light incident model S of the production workshop on sunny and cloudy days is constructed using a physical guided temporal neural network and real-time collected images. s (x,y) and S c (x, y) and the actual production live acquisition model P(x, y), and then use the live acquisition model P(x, y) and the weather light incidence model S s (x,y) and S c (x,y) Perform light compensation, polarization stripe noise removal, and image restoration on the fused image under complex environments to obtain an enhanced clear image of the target under normal and stable conditions;

[0092] The step 1.4 is specifically as follows: first, the enhanced fused image is subjected to hierarchical modular fractal through a modular fractal module and a regional attention mechanism, so as to distinguish between clear areas without noise and unclear areas with much noise; noise is continuously added to different areas of the enhanced fused image in varying degrees through forward diffusion and a position attention mechanism; the forward diffusion process is then reversed through the PDF parameters in the forward process predicted by deep learning; and finally, a clear image is restored from the noisy image, thereby achieving a first denoising process for the enhanced fused image; and finally, the denoised image is subjected to digital preprocessing.

[0093] Furthermore, the step 2.2 is as follows:

[0094] Step 2.2.1: After image segmentation, the image gradient of the calibrated hollowed-out area in the fused image is calculated to detect changes in pixel values. The area with a larger gradient amplitude is the edge of the image. Then, a binary edge image is output. The foreground edge contour is approximated by the Monte Carlo method, and the area of ​​the irregular shape is calculated to approximate the area of ​​the contour area enclosed by the edge features after the image segmentation method:

[0095] Area calculation:

[0096]

[0097] Where n is the number of contour areas enclosed by the edge features after image segmentation processing at the current time t, i is the i-th contour area enclosed by the edge features after image segmentation processing, i = 1, 2, 3, ... n, b i with a i They are the upper and lower bounds of the independent variable x in the i-th contour area surrounded by the edge features of the irregular area after image segmentation at the current time t, x p is in the interval [a i ,b i ] is a random number uniformly distributed on the image, N is the i-th contour area surrounded by the edge features of the irregular area after image segmentation at the current moment t in the interval [a i ,b i ], m is the number of contour areas enclosed by the edge features processed at the previous moment (t-1), j is the jth contour area enclosed by the edge features after image segmentation processing, j = 1, 2, 3, ... m, b j with a j They are the upper and lower bounds of the independent variable x in the j-th contour area surrounded by the edge features of the irregular area processed at the previous moment (t-1), x q is in the interval [a j ,b j ] is a random number uniformly distributed on the surface, M is the jth contour area surrounded by the edge features of the irregular area processed at the previous moment (t-1) in the interval [a j ,b j ] the number of random sampling points, s i is the approximate area of ​​the i-th contour region enclosed by the edge features processed at the current time t, s j is the approximate area of ​​the jth contour region enclosed by the edge features after segmentation at the previous moment (t-1), S (t) Extract the feature area at the current moment t, S (t-1) is the feature area extracted at the previous moment (t-1);

[0098] Step 2.2.2: The calculated different area ratios correspond to different gear adjustment speeds. The gear change is determined based on the trigger conditions:

[0099] The ratio of the area at the current moment to the previous moment:

[0100]

[0101] Calculation of the area ratio of the contour area in the entire area at the previous moment:

[0102]

[0103] Where: S * is the ratio of the area at the current moment to the area at the previous moment, A is the area of ​​the entire image, Calculate the area ratio of the contour area at the current moment to the entire image area. Calculate the area ratio of the contour area at the current moment to the entire image area. Calculate the proportion of the contour area at the previous moment to the entire image area;

[0104] 1) If S is calculated * The range is: The equipment runs at the first gear speed, y=1;

[0105] 2) If S is calculated * The range is: The equipment runs at the second speed, y=2;

[0106] 3) If S is calculated * The range is: The equipment runs at the third speed, y=3;

[0107] Area ratio S * The conversion formula corresponding to the theoretical speed v is:

[0108]

[0109]

[0110] Where: v max(y) For S * The theoretical maximum speed value of the corresponding gear within the change range, v min(y) For S * The theoretical minimum speed value of the corresponding gear within the change range, v (t) Calculate the area ratio of the contour area at the current moment to the entire image area and the ratio of the area at the current moment to the previous moment S * The corresponding theoretical value of velocity is: For S *The maximum value within the variation range, For S * The minimum value within the change range, y = 1, 2, 3 means the equipment gear is adjusted to gear 1, gear 2, gear 3, ε1 = v min(2) -v max(1) ,ε2=v min(3) -v max(2) , ε1>0,ε2>0, ε1 and ε2 are both constants;

[0111] In the inequality, The threshold value of the state variable for equipment gear switching is automatically selected by using a genetic algorithm based on historical data and field production experience.

[0112] Furthermore, the Step 4 specifically includes:

[0113] Step 4.1: The signal processor 8.11 sends a control signal to the variable speed motor. If the variable speed motor is not in working state, it is powered on and the speed is the gear speed generated by the controller. If the variable speed motor is in working state, it is determined whether the current speed gear is consistent with the gear speed required by the controller. If they are consistent, it continues to work. If not, it is adjusted to be consistent.

[0114] The power of the variable speed motor is transmitted to the driving support stepped shaft 6.3 through the transmission of the driving pulley and the stepped shaft pulley. The driving support stepped shaft 6.3 is connected with the driving sprocket I 6.1 and the driving sprocket II 6.4 by a hook-head wedge key 6.6 to realize the coaxial rotation of the driving support stepped shaft 6.3, the driving sprocket I 6.1 and the driving sprocket II 6.4. Then, the transmission chain 8.12 is used to drive the driven sprocket II 4.4 and the driven sprocket I 4.1 on the driven shaft to rotate following the rotation of the driving sprocket II 6.4 and the driving sprocket I 6.1 on the driving support shaft 6.3. Then, the self-propelled gear is automatically Lock the fixed connecting rod 4.3 to connect the prying roller 4.2 with the driven sprocket II 4.4 and the driven sprocket I 4.1, so that the prying roller 4.2 and the driven sprocket II 4.4 and the driven sprocket I 4.1 on the driven shaft can rotate coaxially. Then determine the rotation direction and hang the ring hook 4.8 on the outward cantilever of the prying roller 4.2. As the prying roller 4.2 rotates, the ring hook 4.8 rotates to push out the ore blocks in the screen. If large pieces of agglomerated ore are encountered, the large pieces of agglomerated ore are first dismantled by rotating the cutting blade 4.9, and then the ore blocks in the screen are pushed out by rotating the ring hook 4.8.

[0115] The working principle of the present invention is as follows: the number of agglomerated ores identified by the industrial camera 7.10 is used as the core parameter, the image processing module and the data analysis module in the controller 7.9 generate the optimal speed gear in real time, the motion control module gives the control signal of the speed gear signal to the signal processor 8.11, the signal processor 8.11 transmits the speed gear signal to the variable speed motor, and the variable speed motor thereby regulates the running speed of the agglomerated ores intelligent processing device based on the enhanced visual recognition trigger algorithm, and the ring hook 4.8 with wide ends and narrow middle is suspended by the cantilever part of the prying roller 4.2, and the ring hook 4.8 and the cutting blade 4.9 rotate with the prying roller 4.2, thereby removing or cutting and disassembling the ore, and performing prying processing on the ore to ensure that the ore block prying device puts the previous working section on the multi-stage anti-deformation composite structure vibration screen ( Figure 1 ) The unqualified ore blocks on the surface are pried or cut so that they can slide off the screen smoothly, realizing precise and intelligent control.

[0116] The outstanding features of the present invention are described as follows:

[0117] 1. The ore disassembly and prying device for agglomerated ore is independent of the vibrating screen and can be loaded by the support device. This reduces the interference of strong vibration on the ore disassembly and ore prying functions of the structural device. The main working object of this invention is the multi-stage anti-deformation composite structure vibrating screen in a confined space, and the actual working conditions are relatively complex.

[0118] 2. The device of the present invention adopts visual recognition and adjusts the speed according to the actual situation on site, which is more conducive to extending the service life of the equipment.

[0119] 3. Compared with the guide rail movement mode, the ore dismantling and prying device 1 of the present invention adopts a rotation mode and chain drive, which reduces the subsequent maintenance cost in actual application.

[0120] 4. The 4.8 ring hook used for prying ore adopts a shape that is wide on both sides and narrow in the middle, which can limit the rotation angle of the ring hook and effectively avoid the generation of harsh noise.

[0121] 5. The cutting blade 4.9 for oversized ores is mounted on the prying roller 4.2, which can simultaneously achieve the two functions of screen ore prying and ore cutting, thus enhancing functionality.

[0122] 6. In the present invention, the target area of ​​the recognition image is marked in advance, which simplifies the calculation process and improves the calculation speed when the target is recognized in the later image processing.

[0123] 7. The present invention performs image enhancement processing for target texture restoration on an image by constructing a frequency domain analysis filtering and denoising method based on a neural network.

[0124] 8. The present invention uses an improved bidirectional denoising diffusion model to perform adaptive denoising on the enhanced fused image.

[0125] 9. Compared with the structural equipment that operates at a constant high speed, the present invention can visually identify the amount of agglomerated ore and adjust the speed according to the current production situation.

[0126] The specific embodiments of the present invention are described in detail above with reference to the accompanying drawings. However, the present invention is not limited to the above embodiments. Various changes can be made within the knowledge of ordinary technicians in this field without departing from the scope of the present invention.

Claims

1. An intelligent processing device for agglomerated ore based on an enhanced visual recognition trigger algorithm, characterized by: The invention comprises an ore dismantling and prying device (1) and a vibrating screen. The ore dismantling and prying device (1) is arranged above the vibrating screen through a supporting device. The ore dismantling and prying device (1) comprises a power drive shaft assembly device and a driven shaft assembly device. The power drive shaft assembly device and the driven shaft assembly device are connected through a transmission chain (8.12). An industrial camera (7.10) and a controller (7.9) are installed on the outer surface of the power drive shaft assembly device. A ring hook (4.8) and a cutting blade (4.9) are installed on the outer surface of the driven shaft assembly device. The supporting device A variable speed motor and a signal processor (8.11) are installed on the controller (7.9). The industrial camera (7.10) faces the vibrating screen on the vibrating screen. The controller (7.9) includes an image processing module, a data analysis module, and a motion control module. The input end of the data analysis module is connected to the image processing module, and the output end is connected to the motion control module. The image processing module is connected to the industrial camera (7.10). The input end of the signal processor (8.11) is connected to the motion control module, and the output end is connected to the variable speed motor. The variable speed motor is connected to the power drive shaft assembly device.

2. The intelligent processing device for agglomerated ore based on enhanced visual recognition triggering algorithm according to claim 1, characterized in that: The power drive shaft assembly device comprises a drive support stepped shaft (6.3), a transmission sprocket I (6.1), a hollow rod sleeve (6.2), a transmission sprocket II (6.4), a bearing (6.5), a shaft elastic retaining ring (6.8), and a hook-shaped wedge key (6.6); the hollow rod sleeve (6.2) is loosely sleeved on the outside of the drive support stepped shaft (6.3) and is fixed thereto; the drive support stepped shaft (6.3) is respectively connected to the transmission sprocket II (6.4) and the transmission sprocket I (6.1) at both ends via the hook-shaped wedge key (6.6); and the fixing plate (6.7) is fixedly connected to the support device. Two sets of bearings (6.5) are installed between the fixed plate (6.7) and the driving support stepped shaft (6.3). One end of each bearing (6.5) is rotatably connected to the step of the fixed plate (6.7) through the step of the supporting stepped shaft (6.3), and the other end is rotatably connected to the shaft with an elastic retaining ring (6.8). A stepped shaft pulley is installed at one end of the driving support stepped shaft (6.3). The stepped shaft pulley is connected to the transmission pulley on the output shaft of the variable speed motor through a belt drive. The driving sprocket I (6.1) and the driving sprocket II (6.4) are respectively connected to the driven shaft assembly device through a transmission chain (8.12).

3. The intelligent processing device for agglomerated ore based on enhanced visual recognition triggering algorithm according to claim 1, characterized in that: The driven shaft assembly device comprises a driven support shaft (4.6), a driven sprocket I (4.1), a driven sprocket II (4.4), a prying roller (4.2), a self-locking fastening connecting rod (4.3), a shaft elastic circlip II (4.5), a driven support rod (4.7), a ring hook (4.8), and a cutting blade (4.9). The prying roller (4.2) is sleeved on the driven support shaft (4.6). The driven sprocket I (4.1) and the driven sprocket II (4.4) are installed at both ends of the driven support shaft (4.6) through the shaft elastic circlip (4.5). The driven sprocket I (4.1) and the driven sprocket II (4.4) are connected to the prying roller (4.2) through a self-locking fixed connecting rod (4.3), the ring hook (4.8) is installed on the prying roller (4.2), the cutting blade (4.9) is installed at the slot position of the prying roller (4.2), the two ends of the driven support shaft (4.6) are fixedly connected to the driven support rod (4.7), the driven support rod (4.7) is fixedly connected to the supporting device, and the driven sprocket I (4.1) and the driven sprocket II (4.4) are respectively connected to the power drive shaft assembly device through a transmission chain (8.12).

4. The intelligent processing device for agglomerated ore based on enhanced visual recognition triggering algorithm according to claim 3 is characterized in that: The prying roller (4.2) is circular, and cutting blade mounting grooves and outwardly extending cantilever arms are evenly arranged on the circumference of the prying roller (4.2). A cutting blade (4.9) is fixed in the cutting blade mounting groove, and a group of outwardly extending cantilever arms is fixed between the two blade mounting grooves. The ends of the outwardly extending cantilever arms are fixed with ring hooks (4.8).

5. The intelligent processing device for agglomerated ore based on enhanced visual recognition triggering algorithm according to claim 4 is characterized in that: The prying roller (4.2) is evenly provided with three cutting blade mounting grooves and three groups of outriggers on its circumference. The circumferential angles of the three cutting blades (4.9) differ by 120 degrees from each other. The axial spacing of the three groups of outriggers is 1 / 3 of the width of the vibrating screen, and the circumferential angle difference is 120 degrees. The vibrating screen has 15 rows of slots. Each group of outriggers includes five outriggers. A ring hook mounting position is provided at the end of each outrigger. The angle difference between the three groups of outriggers is 120 degrees, and the transverse coordinate differences are 325 mm and 650 mm respectively.

6. The intelligent processing device for agglomerated ore based on enhanced visual recognition triggering algorithm according to claim 1, characterized in that: The ring hook (4.8) is in the shape of a long strip with wide sides and narrow in the middle.

7. An intelligent processing method for agglomerated ore based on an enhanced visual recognition trigger algorithm, characterized by: The steps include: Step 1: The image processing module in the controller (7.9) performs the following processing: Step 1.1: Use an industrial camera (7.10) to capture live images and mark the target area of ​​the image: Step 1.2: Use the industrial camera (7.10) to capture and process the on-site image. The marked area is processed into a transparent hollow area. The hollow area image is fused with the actual collected image to ensure that the pixel values ​​of the transparent hollow area are from the actual collected image. The change in pixel values ​​is used to determine whether there is agglomerated ore. Step 1.3: Perform image enhancement processing on the fused image by performing target texture restoration through the frequency domain analysis filtering denoising method based on neural network; Step 1.4: De-noise the enhanced image using a bidirectional denoising diffusion model improved based on the dual attention mechanism, and then perform digital preprocessing on the denoised image. Step 2: The pre-processed image is processed as follows by the data analysis module in the controller (7.9): Step 2.1: Use the threshold segmentation method to segment the preprocessed image; Step 2.2: Calculate the area ratio of the agglomerated ore in the image at the current moment and the area ratio of the agglomerated ore in the image at the previous moment for the segmented regions that meet the target conditions in the segmented image; Based on the calculated ratio of the area sizes at two adjacent moments and the current speed, a control signal is generated corresponding to the speed adjustment of different gears; Step 3: The motion control module in the controller (7.9) transmits the control signal generated by the data analysis module to the signal processor (8.11); Step 4: The signal processor (8.11) adjusts the speed of the variable speed motor according to the control signal. The variable speed motor drives the power drive shaft assembly to rotate. The power drive shaft assembly drives the driven shaft assembly to rotate. When the driven shaft assembly rotates, it drives the ring hook (4.8) and the cutting blade (4.9) to rotate, thereby prying and / or cutting the ore blocks with unqualified sizes on the surface of the vibrating screen.

8. The method for intelligently processing agglomerated ores based on an enhanced visual recognition trigger algorithm according to claim 7, characterized in that: The details of step 1.1 are as follows: Step 1.1.1: Determine the installation location of the industrial camera (7.10) according to the actual situation and install the industrial camera (7.10); Step 1.1.2: Use the industrial camera (7.10) to take pictures of the upper surface of the vibrating screen. Step 1.1.3: Based on the actual production situation and on-site cleaning experience, a normal distribution function is constructed, and the probability distribution function is used to estimate the location of the area where the agglomerated ore may fall on the vibrating screen; Step 1.1.4: Calibrate and divide the target cleaning area of ​​the image read by the camera, that is, the position of the vibrating screen holes to clarify the image target area; The step 1.3 is as follows: First, a physical guided temporal neural network and real-time image acquisition are used to construct the light incidence model S of the production workshop on sunny and cloudy days. s (x,y) and S c (x, y) and the actual production live acquisition model P(x, y), and then use the live acquisition model P(x, y) and the weather light incidence model S s (x,y) and S c (x,y) Perform light compensation, polarization stripe noise removal, and image restoration on the fused image under complex environments to obtain an enhanced clear image of the target under normal and stable conditions; The step 1.4 is specifically as follows: first, the enhanced fused image is subjected to hierarchical modular fractal through a modular fractal module and a regional attention mechanism, so as to distinguish between clear areas without noise and unclear areas with much noise; noise is continuously added to different areas of the enhanced fused image in varying degrees through forward diffusion and a position attention mechanism; the forward diffusion process is then reversed through the PDF parameters in the forward process predicted by deep learning; and finally, a clear image is restored from the noisy image, thereby achieving a first denoising process for the enhanced fused image; and finally, the denoised image is subjected to digital preprocessing.

9. The method for intelligently processing agglomerated ores based on an enhanced visual recognition trigger algorithm according to claim 7, characterized in that: The step 2.2 is as follows: Step 2.2.1: After image segmentation, the image gradient of the calibrated hollowed-out area in the fused image is calculated to detect changes in pixel values. The area with a larger gradient amplitude is the edge of the image. Then, a binary edge image is output. The foreground edge contour is approximated by the Monte Carlo method, and the area of ​​the irregular shape is calculated to approximate the area of ​​the contour area enclosed by the edge features after the image segmentation method: Area calculation: Where n is the number of contour areas enclosed by the edge features after image segmentation processing at the current time t, i is the i-th contour area enclosed by the edge features after image segmentation processing, i = 1, 2, 3, ... n, b i with a i They are the upper and lower bounds of the independent variable x in the i-th contour area surrounded by the edge features of the irregular area after image segmentation at the current time t, x p is in the interval [a i ,b i ] is a random number uniformly distributed on the image, N is the i-th contour area surrounded by the edge features of the irregular area after image segmentation at the current moment t in the interval [a i ,b i ], m is the number of contour areas enclosed by the edge features processed at the previous moment (t-1), j is the jth contour area enclosed by the edge features after image segmentation processing, j = 1, 2, 3, ... m, b j with a j They are the upper and lower bounds of the independent variable x in the j-th contour area surrounded by the edge features of the irregular area processed at the previous moment (t-1), x q is in the interval [a j ,b j ] is a random number uniformly distributed on the surface, M is the jth contour area surrounded by the edge features of the irregular area processed at the previous moment (t-1) in the interval [a j ,b j ] the number of random sampling points, s i is the approximate area of ​​the i-th contour region enclosed by the edge features processed at the current time t, s j is the approximate area of ​​the jth contour region enclosed by the edge features after segmentation at the previous moment (t-1), S (t) Extract the feature area at the current moment t, S (t-1) is the feature area extracted at the previous moment (t-1); Step 2.2.2: The calculated different area ratios correspond to different gear adjustment speeds. The gear change is determined based on the trigger conditions: The ratio of the area at the current moment to the previous moment: Calculation of the area ratio of the contour area in the entire area at the previous moment: Where: S * is the ratio of the area at the current moment to the area at the previous moment, A is the area of ​​the entire image, Calculate the area ratio of the contour area at the current moment to the entire image area. Calculate the area ratio of the contour area at the current moment to the entire image area. Calculate the proportion of the contour area at the previous moment to the entire image area; 1) If S is calculated * The range is: The equipment runs at the first gear speed, y=1; 2) If S is calculated * The range is: The equipment runs at the second speed, y=2; 3) If S is calculated * The range is: The equipment runs at the third speed, y=3; Area ratio S * The conversion formula corresponding to the theoretical speed v is: Where: v max(y) For S * The theoretical maximum speed value of the corresponding gear within the change range, v min(y) For S * The theoretical minimum speed value of the corresponding gear within the change range, v (t) Calculate the area ratio of the contour area at the current moment to the entire image area and the ratio of the area at the current moment to the previous moment S * The corresponding theoretical value of the velocity is, For S * The maximum value within the variation range, For S * The minimum value within the change range, y = 1, 2, 3 means the equipment gear is adjusted to gear 1, gear 2, gear 3, ε1 = v min(2) -v max(1) ,ε2=v min(3) -v max(2) , ε1>0,ε2>0, ε1 and ε2 are both constants; In the inequality, The threshold value of the state variable for equipment gear switching is automatically selected by using a genetic algorithm based on historical data and field production experience.

10. The method for intelligently processing agglomerated ores based on an enhanced visual recognition trigger algorithm according to claim 7, characterized in that: The Step 4 specifically includes: Step 4.1: The signal processor (8.11) sends a control signal to the variable speed motor. If the variable speed motor is not in operation, it is powered on and the speed is the gear speed generated by the controller. If the variable speed motor is in operation, it is determined whether the current speed gear is consistent with the gear speed required by the controller. If they are consistent, it continues to work. If not, it is adjusted to be consistent. The power of the variable speed motor is transmitted to the driving support stepped shaft (6.3) through the transmission of the driving pulley and the stepped shaft pulley. The driving support stepped shaft (6.3) is connected to the driving sprocket I (6.1) and the driving sprocket II (6.4) by a hook-shaped wedge key (6.6) to realize the coaxial rotation of the driving support stepped shaft (6.3) and the driving sprocket I (6.1) and the driving sprocket II (6.4). Then, the driven sprocket II (4.4) and the driven sprocket I (4.1) on the driven shaft rotate along with the rotation of the driving sprocket II (6.4) and the driving sprocket (6.1) on the driving support shaft (6.3). Then, the driven sprocket II (4.4) and the driven sprocket I (4.1) on the driven shaft rotate by the rotation of the driving sprocket II (6.4) and the driving sprocket (6.1) on the driving support shaft (6.3). A self-locking fixed connecting rod (4.3) connects the prying roller (4.2) with the driven sprocket II (4.4) and the driven sprocket I (4.1) to achieve coaxial rotation of the prying roller (4.2) and the driven sprocket II (4.4) and the driven sprocket I (4.1) on the driven shaft. After that, the rotation direction is determined and a ring hook (4.8) is hung on the outward cantilever of the prying roller (4.2). As the prying roller (4.2) rotates, the ring hook (4.8) rotates to push out the ore blocks in the screen. If large pieces of agglomerated ore are encountered, the large pieces of agglomerated ore are first dismantled by rotating the cutting blade (4.9), and then the ore blocks in the screen are pushed out by rotating the ring hook (4.8).

Images