Method for monitoring raw materials of reaction kettle

Abstract

The invention relates to a method for monitoring raw materials of a reaction kettle, which comprises the following steps of: performing vibration type crushing treatment on the raw materials by adopting a set vibration frequency to obtain corresponding crushed powder of each part; and performing sharpening processing on the frequency domain enhanced image by using a preferred sharpening algorithm to obtain a dynamic sharpened image, and determining a set vibration frequency positively associated with a maximum value based on the maximum value in the horizontal cross-sectional area of each entity. The intelligent control level can be improved.

Description

Technical Field

[0001] The invention relates to the field of reactors, and in particular to a monitoring method for raw materials of a reactor. Background Art

[0002] Mines include coal mines, metal mines, non-metallic mines, building materials mines, chemical mines, etc. The scale of a mine (also known as production capacity) is usually expressed in annual output or daily output. Annual output refers to the amount of ore produced by a mine each year. According to the size of the output, it is divided into three types: large, medium and small. The size of the mine should be consistent with the economically reasonable service life of the mine. Only in this way can we save infrastructure costs and reduce costs. In the mining production process, mining operations consume the most manpower and material resources, occupy the most funds, and are the production link with the greatest potential for reducing mining costs. The main way to reduce mining costs is to improve labor productivity and product quality and reduce material consumption.

[0003] Before the ore is processed as raw material, for example, before it is put into a reactor for processing, the ore needs to be crushed. However, when the ore is crushed, the size of the ore is uncontrollable, and it is easy to be too large or too small. At this time, if the same vibration frequency is used to crush the stone, it obviously cannot match the actual scene. At the same time, the content of the mining scene is constantly changing. The same sharpening algorithm is used to perform the same sharpening processing on each frame of the mining scene corresponding to different times, and the image quality results are also unsatisfactory.

[0004] CN118072311A discloses a method, device, system and medium for ore image segmentation, including: collecting image data of the ore to be segmented; locating and cropping each ore position in the ore image data through a target detection algorithm to obtain image data of each ore; inputting the image data of the ore into a segmentation model for segmentation to obtain a mask of each ore; cropping the image data according to the ore mask to obtain a mask map of each ore, and identifying according to the ore mask map to obtain the result of each ore.

[0005] CN117994318A discloses a particle size detection method and system in an ore conveying process, the method comprising: graying the collected ore image to obtain a gray image, performing edge detection on the gray image according to an improved Sobel operator, obtaining the edge response value of a target pixel, and then obtaining the edge response values ​​of all pixels in the gray image, comparing the edge response values ​​of all pixels with a set threshold value, distinguishing edge pixels from non-edge pixels, obtaining an enhanced image, detecting the ore particle size according to the enhanced image, and obtaining a prediction result of the ore particle size.

[0006] CN117975340A discloses a video recognition system and method for ore particle size. The video recognition system for ore particle size includes: a belt conveyor, a video acquisition module and a data processing system. The video acquisition module is used to collect ore image video information, including a gantry bracket, a camera and an LED explosion-proof floodlight. The data processing system includes an image preprocessing module, an ore recognition module, an ore counting and statistics module, an incoming ore particle qualified rate calculation module, a human-computer interaction module and a statistical analysis module. The system solves the problem that the recognition results vary greatly under a single threshold value due to different light intensities in space and time, optimizes the shape of the stone by using a convex hull algorithm, optimizes and counts the shape, size and quantity of the stone by using a breadth-first search method and ratio conversion, and calculates the qualified rate in a timely and accurate manner according to the required stone threshold. Summary of the invention

[0007] In order to solve the above problems, the present invention provides a monitoring method for raw materials in a reactor, wherein a binary identifier of a set sharpening algorithm, various image information of a frequency domain enhanced image, and a resolution of the frequency domain enhanced image are input into a feedforward neural network model after multiple learnings, and the feedforward neural network model after multiple learnings is executed to obtain the frequency domain enhanced image output by the model to complete the image signal-to-noise ratio after the set sharpening algorithm, and the sharpening algorithm corresponding to the image signal-to-noise ratio with the largest value among the various image signal-to-noise ratios corresponding to various sharpening algorithms is output as a preferred sharpening algorithm, and the frequency domain enhanced image is sharpened using the preferred sharpening algorithm to obtain a dynamic sharpened image, and at the same time, based on the imaging areas corresponding to each piece of ore in the identified dynamic sharpened image, the corresponding physical horizontal cross-sectional areas of each piece of ore are judged, and based on the maximum value of the physical horizontal cross-sectional areas, the set vibration frequency positively associated with the maximum value is determined to be sent to the crushing actuator, thereby improving the intelligent control level of the crushing actuator.

[0008] According to the present invention, a method for monitoring raw materials in a reactor is provided, the method comprising: The crushing actuator is arranged at the mine outlet of the mine, and is used to perform vibration crushing processing on each piece of ore conveyed from the mine outlet at a set vibration frequency to obtain corresponding portions of crushed powder; The ore-exit collection mechanism is arranged just above the ore-exit port, and is used to collect image data of the ore-exit port once every certain period of time, so as to obtain and output corresponding interval collection images; Connecting a data interpolation device to the mining collection mechanism to perform nearest neighbor interpolation processing on the received interval collection images to obtain and output corresponding nearest neighbor interpolation images; Connecting a dynamic processing device to the data interpolation device to perform white balance processing based on a dynamic threshold on the received nearest neighbor interpolation image to obtain and output a corresponding dynamic processing image; Connecting the enhancement operation device to the dynamic processing device to perform enhancement processing based on the frequency domain of the image content on the received dynamic processing image to obtain and output a corresponding frequency domain enhanced image; Connecting a sharpening prediction device to the enhancement operation device to receive the frequency domain enhanced image, for obtaining various image information of the frequency domain enhanced image, wherein the various image information of the frequency domain enhanced image includes the signal-to-noise ratio, maximum noise amplitude, number of noise types and contrast of the frequency domain enhanced image, inputting a binary identifier of a set sharpening algorithm, various image information of the frequency domain enhanced image and the resolution of the frequency domain enhanced image into a feedforward neural network model after completing multiple learning, executing the feedforward neural network model after completing multiple learning to obtain the image signal-to-noise ratio of the frequency domain enhanced image outputted by the model after completing the setting of the sharpening algorithm, and outputting the sharpening algorithm corresponding to the image signal-to-noise ratio with the largest value among the various image signal-to-noise ratios corresponding to the various sharpening algorithms as the preferred sharpening algorithm; Connecting a parameter correction device to the pulverizing actuator to perform sharpening processing on the frequency domain enhanced image using the preferred sharpening algorithm to obtain a dynamic sharpened image, and judging the respective physical horizontal cross-sectional areas corresponding to the respective pieces of ore based on the respective imaging areas corresponding to the respective pieces of ore in the identified dynamic sharpened image, and determining a set vibration frequency positively associated with the maximum value among the respective physical horizontal cross-sectional areas based on the maximum value to send to the pulverizing actuator; Among them, based on the identified imaging areas corresponding to each piece of ore in the dynamic sharpened image, the physical horizontal cross-sectional areas corresponding to each piece of ore are judged, and based on the maximum value of each physical horizontal cross-sectional area, a set vibration frequency positively associated with the maximum value is determined to be sent to the crushing actuator, including: identifying the imaging areas corresponding to each piece of ore in the dynamic sharpened image based on the grayscale imaging features corresponding to the ore.

[0009] It can be seen that the present invention has at least the following four important inventive concepts: Invention concept 1: receiving a frequency domain enhanced image, and obtaining various image information of the frequency domain enhanced image, wherein the various image information of the frequency domain enhanced image includes a signal-to-noise ratio, a maximum noise amplitude, a number of noise types, and a contrast of the frequency domain enhanced image, thereby providing reliable multiple basic data for the optimization of a sharpening algorithm; Invention concept 2: Inputting a binary flag of a set sharpening algorithm, various image information of a frequency domain enhanced image, and the resolution of the frequency domain enhanced image into a feedforward neural network model that has completed multiple learnings, and executing the feedforward neural network model that has completed multiple learnings to obtain the image signal-to-noise ratio of the frequency domain enhanced image output by the model after the set sharpening algorithm is completed; Inventive concept three: outputting the sharpening algorithm corresponding to the image signal-to-noise ratio with the largest value among the various image signal-to-noise ratios corresponding to the various sharpening algorithms as the preferred sharpening algorithm; Inventive concept four: Use a preferred sharpening algorithm to perform sharpening processing on the frequency domain enhanced image to obtain a dynamic sharpened image, judge the physical horizontal cross-sectional areas corresponding to each piece of ore based on the imaging areas corresponding to each piece of ore in the identified dynamic sharpened image, and determine the set vibration frequency positively correlated with the maximum value based on the maximum value of the physical horizontal cross-sectional areas to send to the crushing actuator, thereby improving the intelligent control level of the crushing actuator. DETAILED DESCRIPTION

[0010] The following is a detailed description of an embodiment of the method for monitoring raw materials in a reactor of the present invention.

[0011] Example 1

[0012] Example 1 is a method for monitoring raw materials in a reactor according to the first embodiment of the present invention, and the method comprises: The crushing actuator is arranged at the mine outlet of the mine, and is used to perform vibration crushing processing on each piece of ore conveyed from the mine outlet at a set vibration frequency to obtain corresponding portions of crushed powder; For example, a pulverizing actuator is arranged at a mine outlet of a mine, and is used to perform a vibration-type pulverizing process on each piece of ore conveyed from the mine outlet by using a set vibration frequency to obtain corresponding portions of pulverized powder, including: the pulverizing actuator includes a separation processing unit, an ore processing unit, and a powder output unit, which are used to cooperate to realize a process of performing a vibration-type pulverizing process on each piece of ore conveyed from the mine outlet by using a set vibration frequency to obtain corresponding portions of pulverized powder; The ore-exit collection mechanism is arranged just above the ore-exit port, and is used to collect image data of the ore-exit port once every certain period of time, so as to obtain and output corresponding interval collection images; Connecting a data interpolation device to the mining collection mechanism to perform nearest neighbor interpolation processing on the received interval collection images to obtain and output corresponding nearest neighbor interpolation images; Connecting a dynamic processing device to the data interpolation device to perform white balance processing based on a dynamic threshold on the received nearest neighbor interpolation image to obtain and output a corresponding dynamic processing image; Connecting the enhancement operation device to the dynamic processing device to perform enhancement processing based on the frequency domain of the image content on the received dynamic processing image to obtain and output a corresponding frequency domain enhanced image; Connecting a sharpening prediction device to the enhancement operation device to receive the frequency domain enhanced image, for obtaining various image information of the frequency domain enhanced image, wherein the various image information of the frequency domain enhanced image includes the signal-to-noise ratio, maximum noise amplitude, number of noise types and contrast of the frequency domain enhanced image, inputting a binary identifier of a set sharpening algorithm, various image information of the frequency domain enhanced image and the resolution of the frequency domain enhanced image into a feedforward neural network model after completing multiple learning, executing the feedforward neural network model after completing multiple learning to obtain the image signal-to-noise ratio of the frequency domain enhanced image outputted by the model after completing the setting of the sharpening algorithm, and outputting the sharpening algorithm corresponding to the image signal-to-noise ratio with the largest value among the various image signal-to-noise ratios corresponding to the various sharpening algorithms as the preferred sharpening algorithm; Connecting a parameter correction device to the pulverizing actuator to perform sharpening processing on the frequency domain enhanced image using the preferred sharpening algorithm to obtain a dynamic sharpened image, and judging the respective physical horizontal cross-sectional areas corresponding to the respective pieces of ore based on the respective imaging areas corresponding to the respective pieces of ore in the identified dynamic sharpened image, and determining a set vibration frequency positively associated with the maximum value among the respective physical horizontal cross-sectional areas based on the maximum value to send to the pulverizing actuator; Wherein, judging the respective physical horizontal cross-sectional areas corresponding to the respective pieces of ore based on the respective imaging areas corresponding to the respective pieces of ore in the identified dynamic sharpened image, and determining the set vibration frequency positively associated with the maximum value among the respective physical horizontal cross-sectional areas to send to the crushing execution mechanism comprises: identifying the respective imaging areas corresponding to the respective pieces of ore in the dynamic sharpened image based on the grayscale imaging features corresponding to the ore; Among them, based on the imaging areas corresponding to each piece of ore in the identified dynamic sharpened image, the physical horizontal cross-sectional areas corresponding to each piece of ore are judged, and based on the maximum value of each piece of physical horizontal cross-sectional area, a set vibration frequency positively associated with the maximum value is determined to be sent to the crushing execution mechanism, which also includes: for each piece of ore in the dynamic sharpened image, the more pixels it occupies in the imaging area, the larger its corresponding physical horizontal cross-sectional area.

[0013] Example 2

[0014] Embodiment 2 is a method for monitoring raw materials in a reactor according to the second embodiment of the present invention.

[0015] Different from Example 1, the monitoring method of the reactor raw materials in Example 2 may also include: Using a data monitoring device, the data monitoring device includes a plurality of monitoring cameras facing the data interpolation device, the dynamic processing device, the enhanced operation device and the sharpening prediction device respectively, and is used for time-sharing monitoring of the real-time status of the data interpolation device, the dynamic processing device, the enhanced operation device and the sharpening prediction device; Wherein, a data monitoring device is used, and the data monitoring device includes a plurality of monitoring cameras facing the data interpolation device, the dynamic processing device, the enhanced operation device and the sharpening prediction device respectively, and is used for time-sharing monitoring of the real-time status of the data interpolation device, the dynamic processing device, the enhanced operation device and the sharpening prediction device, including: the data interpolation device, the dynamic processing device, the enhanced operation device and the sharpening prediction device are respectively connected to the data monitoring device using different CSI interfaces; Wherein, a data monitoring device is used, the data monitoring device comprises a plurality of monitoring cameras facing the data interpolation device, the dynamic processing device, the enhanced operation device and the sharpening prediction device respectively, and is used for time-sharing monitoring of the real-time status of the data interpolation device, the dynamic processing device, the enhanced operation device and the sharpening prediction device, and further comprises: the plurality of monitoring cameras facing the data interpolation device, the dynamic processing device, the enhanced operation device and the sharpening prediction device respectively have the same structure; Wherein, a data monitoring device is used, and the data monitoring device includes a plurality of monitoring cameras facing the data interpolation device, the dynamic processing device, the enhanced operation device and the sharpening prediction device respectively, and is used for time-sharing monitoring of the real-time status of the data interpolation device, the dynamic processing device, the enhanced operation device and the sharpening prediction device, and further includes: a plurality of monitoring cameras with built-in image sensing units facing the data interpolation device, the dynamic processing device, the enhanced operation device and the sharpening prediction device respectively; Among them, a plurality of surveillance cameras have built-in image sensing units respectively facing the data interpolation device, the dynamic processing device, the enhancement operation device and the sharpening prediction device: the image sensing unit is a CMOS sensing unit or a CCD sensing unit.

[0016] Example 3

[0017] Example 3 is a method for monitoring raw materials in a reactor according to an embodiment of the present invention.

[0018] Different from Example 1, the monitoring method of the reactor raw materials in Example 3 may also include: The content display device is connected to the data interpolation device, the dynamic processing device, the enhancement operation device and the sharpening prediction device respectively, and is used to simultaneously display various status information of the data interpolation device, the dynamic processing device, the enhancement operation device and the sharpening prediction device, and further includes: the content display device is a liquid crystal display screen or an LED display array; Wherein, a quartz oscillator device is connected to the data interpolation device, the dynamic processing device, the enhanced operation device and the sharpening prediction device respectively, so as to simultaneously provide the data interpolation device, the dynamic processing device, the enhanced operation device and the sharpening prediction device with the working clock pulses required by each of them; Among them, the quartz oscillator device is respectively connected to the data interpolation device, the dynamic processing device, the enhanced operation device and the sharpening prediction device to simultaneously provide the data interpolation device, the dynamic processing device, the enhanced operation device and the sharpening prediction device with the required working clock pulses, including: the required working clock pulses provided to the data interpolation device, the dynamic processing device, the enhanced operation device and the sharpening prediction device simultaneously are all square wave waveforms.

[0019] In addition, in the monitoring method of the raw materials in the reactor, the physical horizontal cross-sectional areas corresponding to each piece of ore are judged based on the imaging areas corresponding to each piece of ore in the identified dynamic sharpened image, and the set vibration frequency positively associated with the maximum value among the physical horizontal cross-sectional areas is determined based on the maximum value to be sent to the crushing actuator, which also includes: the grayscale imaging feature corresponding to the ore is the grayscale value distribution range corresponding to the ore.

[0020] The monitoring method of the reactor raw materials of the present invention can solve the technical problems in the prior art that the time-sharing image optimization effect of the ore outlet is poor and the single stone crushing strategy cannot match different ore outlet scenes. At the same time, it can intelligently select the best sharpening algorithm for the ore outlet image at different times, and can also determine the positively correlated set vibration frequency based on the maximum physical horizontal cross-sectional area of ​​the current ore, thereby improving the intelligent control level of the crushing actuator.

[0021] It will be appreciated by those skilled in the art that various modifications may be made to the embodiments described above and other embodiments of the present invention without departing from the scope of the present invention. It is therefore understood that the present invention is not limited to the specific embodiments disclosed above, but covers any changes, modifications or substitutions within the scope defined by the claims of the present application.

Claims

1. A method for monitoring raw materials in a reactor, characterized in that: The method comprises: The crushing actuator is arranged at the mine outlet of the mine, and is used to perform vibration crushing processing on each piece of ore conveyed from the mine outlet at a set vibration frequency to obtain corresponding portions of crushed powder; The ore-exit collection mechanism is arranged just above the ore-exit port, and is used to collect image data of the ore-exit port once every certain period of time, so as to obtain and output corresponding interval collection images; Connecting a data interpolation device to the mining collection mechanism to perform nearest neighbor interpolation processing on the received interval collection images to obtain and output corresponding nearest neighbor interpolation images; Connecting a dynamic processing device to the data interpolation device to perform white balance processing based on a dynamic threshold on the received nearest neighbor interpolation image to obtain and output a corresponding dynamic processing image; Connecting the enhancement operation device to the dynamic processing device to perform enhancement processing based on the frequency domain of the image content on the received dynamic processing image to obtain and output a corresponding frequency domain enhanced image; Connecting a sharpening prediction device to the enhancement operation device to receive the frequency domain enhanced image, for obtaining various image information of the frequency domain enhanced image, wherein the various image information of the frequency domain enhanced image includes the signal-to-noise ratio, maximum noise amplitude, number of noise types and contrast of the frequency domain enhanced image, inputting a binary identifier of a set sharpening algorithm, various image information of the frequency domain enhanced image and the resolution of the frequency domain enhanced image into a feedforward neural network model after completing multiple learning, executing the feedforward neural network model after completing multiple learning to obtain the image signal-to-noise ratio of the frequency domain enhanced image outputted by the model after completing the setting of the sharpening algorithm, and outputting the sharpening algorithm corresponding to the image signal-to-noise ratio with the largest value among the various image signal-to-noise ratios corresponding to the various sharpening algorithms as the preferred sharpening algorithm; Connecting a parameter correction device to the pulverizing actuator to perform sharpening processing on the frequency domain enhanced image using the preferred sharpening algorithm to obtain a dynamic sharpened image, and judging the respective physical horizontal cross-sectional areas corresponding to the respective pieces of ore based on the respective imaging areas corresponding to the respective pieces of ore in the identified dynamic sharpened image, and determining a set vibration frequency positively associated with the maximum value among the respective physical horizontal cross-sectional areas based on the maximum value to send to the pulverizing actuator; Among them, based on the identified imaging areas corresponding to each piece of ore in the dynamic sharpened image, the physical horizontal cross-sectional areas corresponding to each piece of ore are judged, and based on the maximum value of each physical horizontal cross-sectional area, a set vibration frequency positively associated with the maximum value is determined to be sent to the crushing actuator, including: identifying the imaging areas corresponding to each piece of ore in the dynamic sharpened image based on the grayscale imaging features corresponding to the ore.

2. The method for monitoring the raw materials of a reactor according to claim 1, characterized in that: Based on the imaging areas corresponding to the respective pieces of ore in the identified dynamic sharpened image, the respective physical horizontal cross-sectional areas corresponding to the respective pieces of ore are judged; based on the maximum value of the respective physical horizontal cross-sectional areas, a set vibration frequency positively associated with the maximum value is determined to be sent to the crushing actuator, and the set vibration frequency also includes: for each piece of ore in the dynamic sharpened image, the more pixels in the imaging area it occupies, the larger the corresponding physical horizontal cross-sectional area.

3. The method for monitoring the raw materials of a reactor according to claim 2, characterized in that: The method further comprises: Using a data monitoring device, the data monitoring device includes a plurality of monitoring cameras facing the data interpolation device, the dynamic processing device, the enhanced operation device and the sharpening prediction device respectively, and is used for time-sharing monitoring of the real-time status of the data interpolation device, the dynamic processing device, the enhanced operation device and the sharpening prediction device; Among them, a data monitoring device is used, and the data monitoring device includes multiple monitoring cameras facing the data interpolation device, the dynamic processing device, the enhanced operation device and the sharpening prediction device respectively, and is used to time-share monitor the real-time status of the data interpolation device, the dynamic processing device, the enhanced operation device and the sharpening prediction device. The real-time status includes: the data interpolation device, the dynamic processing device, the enhanced operation device and the sharpening prediction device are respectively connected to the data monitoring device using different CSI interfaces.

4. The method for monitoring the raw materials of a reactor according to claim 3, characterized in that: A data monitoring device is used, and the data monitoring device includes multiple monitoring cameras respectively facing the data interpolation device, the dynamic processing device, the enhanced operation device and the sharpening prediction device, and is used for time-sharing monitoring of the real-time status of the data interpolation device, the dynamic processing device, the enhanced operation device and the sharpening prediction device. The device also includes: the structures of the multiple monitoring cameras respectively facing the data interpolation device, the dynamic processing device, the enhanced operation device and the sharpening prediction device are the same.

5. The method for monitoring the raw materials of a reactor according to claim 4, characterized in that: A data monitoring device is used, and the data monitoring device includes multiple monitoring cameras respectively facing the data interpolation device, the dynamic processing device, the enhanced operation device and the sharpening prediction device, and is used for time-sharing monitoring of the real-time status of the data interpolation device, the dynamic processing device, the enhanced operation device and the sharpening prediction device. It also includes: multiple monitoring cameras with built-in image sensing units respectively facing the data interpolation device, the dynamic processing device, the enhanced operation device and the sharpening prediction device.

6. The method for monitoring the raw materials of a reactor according to claim 5, characterized in that: A plurality of surveillance camera built-in image sensing units respectively facing the data interpolation device, the dynamic processing device, the enhancement operation device and the sharpening prediction device: the image sensing unit is a CMOS sensing unit or a CCD sensing unit.

7. The method for monitoring the raw materials of a reactor according to claim 2, characterized in that: The method further comprises: The content display device is respectively connected to the data interpolation device, the dynamic processing device, the enhanced operation device and the sharpening prediction device to simultaneously display various status information of the data interpolation device, the dynamic processing device, the enhanced operation device and the sharpening prediction device, and also includes: the content display device is a liquid crystal display screen or an LED display array.

8. The method for monitoring the raw materials of a reactor according to claim 7, characterized in that: A quartz oscillator device is respectively connected to the data interpolation device, the dynamic processing device, the enhanced operation device and the sharpening prediction device to simultaneously provide the data interpolation device, the dynamic processing device, the enhanced operation device and the sharpening prediction device with the working clock pulses they need.

9. The method for monitoring the raw materials of a reactor according to claim 8, characterized in that: The quartz oscillator device is respectively connected to the data interpolation device, the dynamic processing device, the enhanced operation device and the sharpening prediction device to simultaneously provide the data interpolation device, the dynamic processing device, the enhanced operation device and the sharpening prediction device with the required working clock pulses, including: the required working clock pulses provided to the data interpolation device, the dynamic processing device, the enhanced operation device and the sharpening prediction device simultaneously are all square wave waveforms.