An image sharpening device and method for a dusty environment

Abstract

This application discloses an image sharpening processing device and method for dusty environments, belonging to the fields of industrial video surveillance and image recognition processing. The device includes an image acquisition module, a sound field generator, a sharpness analysis module electrically connected to the image acquisition module, a controller electrically connected to the sharpness analysis module and the sound field generator, and a power supply module. This device can generate a dynamically optimized frequency-tuned sound field, causing suspended dust particles in the environment to collide and agglomerate, increasing particle size, reducing the number of particles, and partially settling, thus physically enhancing environmental visibility. It can emit a highly directional sound field towards target equipment, improving the clarity of the video image input from the image acquisition source. It has the function of combining general improvement of clarity in the monitored area and enhancement of clarity in key areas. Furthermore, it combines algorithms to achieve a dual collaborative image recognition and data processing process of physical enhancement and algorithm optimization, improving video clarity and thus improving industrial control accuracy and operational safety.

Description

Technical Field

[0001] This application belongs to the fields of industrial video surveillance, image recognition and image processing technology, and specifically relates to an image sharpening processing device and method for dusty environments. Background Technology

[0002] In industrial production processes, video monitoring equipment is widely used in safety monitoring, production scheduling, and equipment status monitoring. When such video monitoring equipment is placed in a dusty environment, the scattering and absorption of light by suspended dust particles can lead to blurred images, reduced contrast, and loss of detail, affecting the monitoring effect on the production process. Therefore, how to acquire images in dusty environments is an important issue.

[0003] Currently, video image acquisition in dusty environments typically employs physical protection methods and dehazing image processing methods. Physical protection methods safeguard the video monitoring equipment to prevent dust intrusion, ensuring high-reliability video image acquisition. Dehazing image processing methods utilize image enhancement, image restoration, fusion-based, and deep learning-based image processing algorithms on the raw images acquired by the video monitoring equipment to obtain the final video image.

[0004] However, among the two methods mentioned above for video image acquisition in dusty environments, physical protection methods can only prevent dust from entering the equipment but cannot remove the obstruction of the optical path by dust. Therefore, the clarity of the acquired video image is limited by the obstruction of the optical path, resulting in insufficient image clarity. Dehazing image processing methods cannot eliminate physically present dust, and when the dust concentration is too high, the transmittance is extremely low, resulting in poor clarity of the acquired original image. Consequently, it is impossible to effectively restore image details based on the original image, also leading to insufficient image clarity. Therefore, there is a need to provide an image processing device and method that can effectively improve image clarity in dusty environments. Summary of the Invention

[0005] The purpose of this application is to provide an image sharpening processing device and method for dusty environments, which can solve the problem of insufficient image sharpness in related technologies.

[0006] To solve the above-mentioned technical problems, this application is implemented as follows: In a first aspect, this application provides an image sharpening processing apparatus for dusty environments, comprising: The image acquisition module is used to acquire target video images of the target area to be monitored. A sound field generator is used to emit a sound field toward the target area according to the instructions of the controller. The sound field generator emits sound in the direction of the target area, and the distance between the sound field generator and the image acquisition module, as well as the angle between the sound field generator and the optical axis of the image acquisition module, make the sound field generated by the sound field generator cover the target area. A sharpness analysis module, electrically connected to the image acquisition module, is used to acquire the target video image and determine the dust concentration value of the target area and the sharpness value of the target video image based on the target video image and the dark channel prior algorithm. The controller is electrically connected to the clarity analysis module and the sound field generator respectively, and is used to receive the dust concentration value and clarity value determined by the clarity analysis module, and control the operating frequency and / or output power of the sound field generator based on the received dust concentration value, clarity value and preset dust concentration threshold and clarity threshold. The power supply module is used to supply power to the image acquisition module, the sharpness analysis module, and the controller.

[0007] Optionally, the sound field generator is an adjustable directional sound field generator, used to emit a sound field to the target area and a strongly directional sound field to the target device in the target area according to the instructions of the controller.

[0008] Optionally, the sound field generator includes a compression driver and an exponentially type cylinder adapted to the compression driver.

[0009] Optionally, the sharpness analysis module is further configured to perform image enhancement processing on the target video image to obtain the target video image after image enhancement processing.

[0010] Optionally, the device further includes a control switch electrically connected to the controller for controlling the opening and closing of the controller.

[0011] Optionally, the device further includes a control switch, which is embedded in the housing of the image acquisition module and electrically connected to the controller for controlling the opening and closing of the controller.

[0012] Optionally, the image acquisition module has a protection rating of at least IP67, and the housing of the image acquisition module integrates a heat sink and an air-cooling channel.

[0013] Secondly, this application provides an image sharpening processing method for dusty environments, the method comprising: Acquire target video images of the target area to be monitored; Based on the target video image and the dark channel prior algorithm, the dust concentration value of the target area and the sharpness value of the target video image are determined. When the dust concentration value is greater than or equal to a preset dust concentration threshold, and / or the clarity value of the target video image is less than or equal to a preset clarity threshold, a directional sound field is emitted toward the target area until the dust concentration value is less than the preset dust concentration threshold and the clarity value of the target video image is greater than the preset clarity threshold.

[0014] Optionally, determining the dust concentration value of the target region and the sharpness value of the target video image based on the target video image and the dark channel prior algorithm includes: The dark channel prior algorithm is used to calculate the dark channel image corresponding to the target video image; Based on a first pixel in the dark channel image at a preset ratio, a second pixel in the target video image corresponding to the position of the first pixel is determined, and the atmospheric light value is determined based on the pixel value of the second pixel. The first pixel is the pixel with the highest fog concentration. Based on the atmospheric light value, the transmittance of the target area is calculated using an atmospheric scattering model; The dust concentration value of the target area is determined based on the transmittance of the target area. The sharpness value of the target video image is calculated using the Brenner gradient function based on the grayscale values ​​of two pixels that differ by two units in the target video image.

[0015] Optionally, the step of emitting a directional sound field toward the target area when the dust concentration value is greater than or equal to a preset dust concentration threshold, and / or the sharpness value of the target video image is less than or equal to a preset sharpness threshold, until the dust concentration value is less than the preset dust concentration threshold and the sharpness value of the target video image is greater than the preset sharpness threshold, includes: When the dust concentration value is greater than or equal to a preset dust concentration threshold, and / or the clarity value of the target video image is less than or equal to a preset clarity threshold, a directional sound field is emitted toward the target area based on a first operating frequency and a first output power. The first operating frequency and the first output power are initial operating frequencies and initial output power set according to the dust concentration value and clarity value at the monitoring time. The target area is frequency scanned based on a preset frequency range and a preset step size. The second operating frequency of the directional sound field is determined based on the rate of increase in the sharpness value during the frequency scan. The second output power of the directional sound field is determined based on the determined second operating frequency and the rate of increase in the sharpness value. Based on the second operating frequency and the second output power, a directional sound field is emitted towards the target area until the dust concentration value is less than a preset dust concentration threshold and the clarity value of the target video image is greater than a preset clarity threshold, at which point the directional sound field emission towards the target area stops.

[0016] Optionally, the step of emitting a directional sound field toward the target area when the dust concentration value is greater than or equal to a preset dust concentration threshold, and / or the sharpness value of the target video image is less than or equal to a preset sharpness threshold, until the dust concentration value is less than the preset dust concentration threshold and the sharpness value of the target video image is greater than the preset sharpness threshold, includes: When the dust concentration value is greater than or equal to a preset dust concentration threshold, and / or the clarity value of the target video image is less than or equal to a preset clarity threshold, a directional sound field is emitted toward the target area based on a first operating frequency and a first output power. The first operating frequency and the first output power are initial operating frequencies and initial output power set according to the dust concentration value and clarity value at the monitoring time. The target area is frequency scanned based on a preset frequency range and a preset step size. The second operating frequency of the directional sound field is determined based on the rate of increase in the sharpness value during the frequency scan. The second output power of the directional sound field is determined based on the determined second operating frequency and the rate of increase in the sharpness value. The directional sound field is then emitted toward the target area based on the second operating frequency and the second output power. The target area is cyclically scanned within a preset frequency range with a preset step size. The second operating frequency is updated based on the rate of increase in sharpness value during the frequency scan. The second output power is updated based on the updated second operating frequency and the rate of increase in sharpness value. A directional sound field is emitted toward the target area based on the updated second operating frequency and the updated second output power. The directional sound field is emitted toward the target area until the dust concentration value is less than a preset dust concentration threshold and the sharpness value of the target video image is greater than a preset sharpness threshold, at which point the emission of the directional sound field toward the target area stops.

[0017] Optionally, the method further includes emitting a directional sound field towards the target area until the dust concentration value is less than a preset dust concentration threshold and the sharpness value of the target video image is greater than a preset sharpness threshold: The target video image is subjected to image enhancement processing to obtain the target video image after image enhancement processing; Video recognition processing is performed on the target video image after image enhancement.

[0018] The technical solution provided in this application may include the following beneficial effects: This application provides an image sharpening processing device for dusty environments. The device acquires a target video image of the area to be monitored via an image acquisition module, then transmits it to a sharpness analysis module. The sharpness analysis module directly estimates the transmittance of the dusty environment based on the target video image using a dark channel prior algorithm, and determines the dust concentration and sharpness value of the target video image based on the transmittance. This eliminates the need for an additional dust concentration sensor, effectively reducing the complexity and implementation cost of the image sharpening processing device. A sound field generator causes suspended dust particles in the dusty environment to collide and agglomerate, increasing particle size, reducing the number of particles, and partially settling them, physically enhancing environmental visibility and ensuring that the target video image acquired by the image acquisition module maintains consistently high sharpness. The control module adjusts the operating frequency and output power of the sound field generator according to the dust concentration and image sharpness value, quickly matching the optimal sound field parameters and achieving timely and precise control of the sound field adjustment. It can quickly re-optimize when dust concentration changes, greatly reducing the impact of dust on image sharpness and improving industrial control accuracy and operational safety. Furthermore, after the sound field generator effectively improves the environmental visibility from a physical level under the adjustment of the controller, the image is then enhanced by the dark channel prior algorithm of the sharpness analysis module. This enables image processing based on a clearer original image, achieving a dual collaborative image recognition and data processing effect of physical preprocessing enhancement and further algorithm optimization, which is beneficial to further improve the sharpness of video images.

[0019] This application provides an image sharpening method for dusty environments. This method utilizes a dark channel prior algorithm to determine the dust concentration and sharpness of the target video image, avoiding the use of parameter acquisition devices such as dust concentration sensors, thus effectively improving data acquisition efficiency and reducing data acquisition costs. By timely monitoring of dust concentration and sharpness values, and emitting a directional sound field towards the target area based on the monitoring results, sound field coagulation technology can be used to cause suspended dust particles in the dusty environment to collide, agglomerate, and quickly settle, thereby promptly removing suspended dust, improving environmental visibility, and ensuring that the acquired video image maintains a consistently high level of sharpness. Furthermore, it facilitates subsequent image processing based on a clearer original image, achieving a dual synergistic effect of physical preprocessing enhancement and further algorithm optimization in image recognition and data processing, thereby further improving the sharpness of video images in dusty environments. Attached Figure Description

[0020] Figure 1 This is a schematic diagram of the structure of an image sharpening processing device for a dusty environment disclosed in an embodiment of this application; Figure 2 This is a schematic flowchart of an image sharpening method for dusty environments disclosed in an embodiment of this application; Figure 3 This is a schematic flowchart of another image sharpening method for dusty environments disclosed in the embodiments of this application.

[0021] Figure labeling: 110: Image acquisition module, 120: Sound field generator, 130: Sharpness analysis module, 140: Controller, 150: Power supply module. Detailed Implementation

[0022] To make the objectives, technical solutions, and advantages of this application clearer, the technical solutions of this application will be clearly and completely described below in conjunction with specific embodiments and corresponding drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of them. All other embodiments obtained by those skilled in the art based on the embodiments of this application without creative effort are within the scope of protection of this application.

[0023] The technical solutions disclosed in the various embodiments of this application are described in detail below with reference to the accompanying drawings.

[0024] Example 1 like Figure 1 As shown in the figure, this specification provides an image sharpening processing device for dusty environments. The device includes: an image acquisition module 110, a sound field generator 120, a sharpness analysis module 130, a controller 140, and a power supply module 150. Wherein: The image acquisition module 110 is used to acquire target video images of the target area to be monitored.

[0025] In this application, "dust environment" refers to industrial environments and sites with dust, such as: iron and steel enterprise mineral powder storage and transportation sites, stone crushing and cement plant sintering sites, coal mine underground mining faces, and mine crushing stations. The target area corresponding to the aforementioned industrial video monitoring scenarios involving dust interference in steel, mining, cement, and chemical industries can be automated production lines, conventional factory video monitoring areas, etc., and this target area can be an open space or a closed space.

[0026] To obtain clearer video images in the aforementioned dusty environment, this application designs an image sharpening processing device. The image acquisition module can employ a camera, industrial camera, 3D vision sensor, etc. This application uses a camera as an example for description; other image acquisition modules operate on the same principle as the camera.

[0027] In implementation, the resolution of the image acquisition module must meet the basic clarity requirements of industrial video monitoring, such as a resolution greater than or equal to 1920×1080 and a video frame rate greater than or equal to 25fps. The image acquisition module can be installed above the target area or on a support column. The installation height is determined according to the size of the target area, usually 3-5 meters, with a downward viewing angle of 15°-30°. The installation position, height, and angle of the image acquisition module should ensure that it covers the target area.

[0028] The sound field generator 120 is used to emit a sound field towards the target area according to the instructions of the controller. The sound field generator emits sound towards the target area (i.e., the monitoring area of ​​the image acquisition module). The distance between the sound field generator and the image acquisition module, as well as the angle between the optical axis of the sound field generator and the image acquisition module, make the sound field generated by the sound field generator cover the target area, and the image acquisition model can acquire the target video image of the target area.

[0029] The directional sound field sent by the sound field generator can be a normal sound field pointing in all directions toward the target area, or a highly directional sound field targeting specific equipment (such as a robotic arm, a key piece of equipment, etc.) in the target area.

[0030] The sound field generator is installed near the image acquisition module, specifically above, below, or to the side of the image acquisition module. The distance between the sound field generator and the image acquisition module, as well as the angle between the sound field generator and the optical axis direction (or the main line of sight) of the image acquisition module, ensure that the sound field generated by the sound field generator can cover the target area, while the image acquisition model can acquire the target video image of the target area.

[0031] In practice, taking a camera as an example, the angle between the sound field generator's sound emission direction and the camera's optical axis can be 0°-15°, and the operating frequency is adjustable within the range of 500Hz-8kHz. The sound pressure level generated at the center of the monitoring area is no less than 130dB. The sound field generated by this sound field generator can cause micron and submicron suspended particles (i.e., dust) in the air of the target area to undergo violent relative motion, collide with each other, and agglomerate and adhere to each other under the action of van der Waals forces and electrostatic forces, forming large-diameter agglomerates. These large-diameter agglomerates settle under the action of gravity, thereby reducing the number of suspended particles in the environment and improving the environmental visibility of the target area.

[0032] In practice, depending on the driving method, an electric sound field generator or a pneumatic sound field generator can be used. The pneumatic sound field generator also needs to be equipped with a corresponding air source and pipeline. A piezoelectric sound field generator can also be used. This application does not limit the use of such generators.

[0033] See also Figure 1It is known that the device also includes a sharpness analysis module 130, which is electrically connected to the image acquisition module 110. It is used to acquire the target video image and determine the dust concentration value of the target area and the sharpness value (also known as sharpness score) of the target video image based on the target video image and the dark channel prior algorithm.

[0034] By electrically connecting to the image acquisition module, the sharpness analysis module can acquire the target video image from the image acquisition module. Then, the sharpness analysis module uses a dark channel prior algorithm to determine the dust concentration value of the target area and the sharpness value of the target video image. In this application, the sharpness analysis module first uses a dark channel prior algorithm to determine the transmittance of the target area, then determines the dust concentration value based on the transmittance and the transmittance-dust concentration relationship, and finally calculates the sharpness value of the target video image using the Brenner gradient function. This method avoids the need for a separate dust concentration sensor in the device, effectively reducing the complexity of the device and also helping to reduce device costs.

[0035] Because the image sharpening processing device is equipped with a sound field generator, the acquired target video image is clearer. The sharpness analysis module can then use the dark channel prior algorithm to calculate the dust concentration value and sharpness value based on the clearer target video image. This is equivalent to optimizing the input data of the dark channel prior algorithm through the sound field generator. Therefore, the device of this application can greatly improve the sharpness of video images in dusty environments.

[0036] Furthermore, when acquiring the target video image, the clarity analysis module in this application can determine whether to acquire the target video image in real time or at set intervals, depending on different application scenarios and different requirements for video image clarity. No limitation is made here.

[0037] In practice, the clarity analysis module can use an industrial-grade image processing chip, and the video clarity analysis module can be fixedly installed inside the housing of the image acquisition module, further simplifying the structure of the image clarity processing device.

[0038] In implementation, the process by which the sharpness analysis module acquires the target video image and determines the dust concentration value of the target area and the sharpness value of the target video image based on the dark channel prior algorithm is as follows: 1) Calculate the dark channel image corresponding to the target video image using the dark channel prior algorithm; 2) Based on the first pixel in the dark channel image with a preset ratio, determine the second pixel in the target video image corresponding to the position of the first pixel, and determine the atmospheric light value based on the pixel value of the second pixel. The first pixel is the pixel with the highest fog concentration. 3) Based on atmospheric light values, the transmittance of the target area is calculated using an atmospheric scattering model; 4) Determine the dust concentration value of the target area based on the transmittance of the target area and the empirical formula of transmittance-dust concentration; 5) Based on the gray values ​​of two pixels that differ by two units in the target video image, the sharpness value of the target video image is calculated using the Brenner gradient function.

[0039] See also Figure 1 It is understood that the device also includes a controller 140, which is electrically connected to the clarity analysis module and the sound field generator, respectively. The controller receives the dust concentration value and clarity value determined by the clarity analysis module, and controls the operating frequency and / or output power of the sound field generator based on the received dust concentration value, clarity value, and preset dust concentration threshold and clarity threshold. In practice, an industrial-grade microcontroller can be used.

[0040] The preset dust concentration threshold and resolution threshold can be set according to the specific requirements of video surveillance of the target area in different application scenarios. For example, the dust concentration threshold can be set to 50. The sharpness threshold is set to the average sharpness value under the original dust-free conditions. 90% (the average value of the normalized results of the sharpness evaluation function values ​​in the same dust-free environment) As the original reference for sharpness scoring, As a threshold).

[0041] In implementation, when the dust concentration value obtained by the controller from the clarity analysis module is greater than or equal to the preset dust concentration threshold, and the clarity value is less than or equal to the preset clarity threshold, or when the dust concentration value of the target area is greater than or equal to the preset dust concentration threshold and the clarity value of the target video image is less than or equal to the preset clarity threshold, the controller controls the sound field generator to start. During the operation of the sound field generator, the controller adjusts the operating frequency of the sound field generator and the output power of the sound field generator according to the changes in the dust concentration value and clarity value in the environment, or adjusts the operating frequency and output power of the sound field generator simultaneously, thereby rapidly improving the transmittance and visibility of the environment.

[0042] See also Figure 1 It is understood that the device also includes a power supply module 150, which is used to power the image acquisition module, the sharpness analysis module and the controller.

[0043] In practice, if the sound field generator is an electric sound field generator, the power supply module is also electrically connected to the electric sound field generator to provide it with power.

[0044] In implementation, a power module can be integrated into the controller. The power module uses a rechargeable lithium battery and is equipped with a USB fast charging interface. The power module has overvoltage, overcurrent and overtemperature protection circuits to improve the safety of the device.

[0045] In this embodiment, the sound field generator can be implemented in various ways. The following provides an optional structure. The sound field generator is an adjustable directional sound field generator, used to emit a sound field to the target area and emit a strong directional sound field to the target device in the target area according to the controller's instructions.

[0046] The target device can be a specific device or component in the target area that needs to be monitored. Correspondingly, the clarity requirements for the video monitoring images of the target device are also higher. In this case, a strong directional sound field (i.e., a high-intensity directional sound field) can be emitted towards the target device using an adjustable directional sound field generator. To emit a strong directional sound field, the operating frequency range of the directional sound field generator can be 500Hz-8000Hz, the directional angle of the directional sound field generator can be 20°-45°, and the sound pressure level generated by the directional sound field generator at the center of the target area should be greater than or equal to 130dB, preferably 150dB.

[0047] The aforementioned adjustable directional sound field generator may include a sound-generating component and an acoustic waveguide structure (i.e., a horn) disposed in front of the sound-generating component. The acoustic waveguide structure is used to constrain and guide the sound field to be emitted in a preset direction. In this embodiment, "adjustable" refers to the ability to adjust the position of the acoustic waveguide structure, such as allowing it to rotate or focus, thereby adjusting the sound field and emitting sound fields with different operating frequencies or output powers toward the target device in the target area. The directional sound field generator has the function of emitting a strong directional sound field toward a selected key area in the target area (i.e., the area where the target device is located). It can generate a dynamically optimized frequency-tuned sound field, thereby further accelerating particulate matter settling, rapidly reducing the number of particulate matter, and thus rapidly improving the environmental visibility of the area where a specific device or component is located. This improves the clarity of the target video image from the image acquisition source, achieving the effect of generally improving the clarity of the target area and enhancing the video clarity of the key area where the target device is located.

[0048] In this embodiment, the sound field generator can be implemented in various ways. Another optional structure is provided below, in which the sound field generator includes a compression driver and an exponential cylinder adapted to the compression driver.

[0049] The aforementioned sound field generator is an electrically powered sound source. In implementation, the rated power of the compression driver can be set to 20W, the operating frequency range can be set to 500Hz-8000Hz, and the maximum sound pressure level can reach 150dB. The index horn can be made of aluminum alloy, with the inner wall polished to further reduce energy loss during sound wave transmission. The small port diameter of the index horn can be set to 20mm for seamless connection with the vibrating end of the compression driver, while the large port diameter can be set to 80mm, and the directivity angle can be set to 30°, thereby converting the point sound source into a planar directive sound field.

[0050] The sound field generator using an electric sound source can make full use of the power supply in the image sharpening processing device, eliminating the need for air sources, pipelines, and other structures, thus simplifying the device and making it easier to promote and use.

[0051] Furthermore, in this embodiment, the sharpness analysis module is also used to perform image enhancement processing on the target video image to obtain the target video image after image enhancement processing.

[0052] Specifically, according to the atmospheric scattering model It can be obtained According to the formula Image enhancement processing is performed on the target video image to restore image details, thereby improving image clarity. These are observed images with fog. It is a fog-free image to be recovered. It is the atmospheric light value. Transmittance indicates the proportion of light that passes through fog and reaches the camera. It is the lower limit threshold of the image, which is usually set to 0.1 to avoid noise amplification caused by a denominator that is too small.

[0053] Furthermore, the image sharpening processing device in this embodiment also includes a control switch, which is electrically connected to the controller and used to control the controller's on and off states. By controlling the controller's on and off states, the sound field generator is also controlled to turn on and off.

[0054] By setting a control switch, the controller can be manually turned on and off in emergency situations, which is convenient and helps to further improve the stability and reliability of the image sharpening processing device.

[0055] Furthermore, the image sharpening processing device in this embodiment also includes a control switch, which is embedded in the housing of the image acquisition module and electrically connected to the controller for controlling the controller to turn on and off.

[0056] By embedding the control switch into the housing of the image acquisition module, the structure of the image sharpening processing device can be further simplified, and it is also convenient to realize the electrical connection between the control switch and the controller.

[0057] Furthermore, in this embodiment, the image acquisition module has a protection level of at least IP67, and the housing of the image acquisition module integrates a heat sink and an air-cooling channel.

[0058] In practice, taking a camera as an example, the camera's protection rating can be IP67 or IP68, effectively protecting the image acquisition module, further improving the clarity of video image acquisition, and also helping to extend the device's service life. By integrating heat sinks and air-cooling channels into the housing of the image acquisition module, heat dissipation can be accelerated when the image acquisition module is used in high-temperature scenarios, improving the device's performance.

[0059] Furthermore, sealing rings and waterproof sleeves can be added to the camera's housing interface to further prevent dust and liquids from entering the device, thus providing protection. The camera's lens barrel can be made of stainless steel or coated with a corrosion-resistant film, effectively preventing environmental corrosion when the device is used in acidic or alkaline environments. The camera's lens assembly can be sealed with sealant to prevent moisture penetration. The camera can also have a built-in rubber buffer bracket to absorb mechanical vibration energy, thereby improving the clarity of the video image.

[0060] The following sections will take video monitoring scenarios of underground coal mining faces and video monitoring scenarios of mine crushing stations as examples to detail the implementation process of the image enhancement processing device in practical applications.

[0061] 1) Image enhancement processing device for coal dust environment in underground coal mining In accordance with standards such as the "Specifications for Video Installation and Network Access in Coal Mine Industry", cameras are installed at the support of the coal mining face, and the image enhancement processing device is modified to be explosion-proof, which meets the requirements of coal mine safety regulations and has obtained the MA mark.

[0062] The camera is a mine-grade explosion-proof camera with an IP68 protection rating. The housing is made of 304 stainless steel with a thickness of ≥3mm. The sound generator uses a mine-grade explosion-proof compression driver and an exponential generator, with an adjustable operating frequency of 500Hz-8kHz, a maximum sound pressure level of 150dB, and an explosion-proof rating of Ex d I Mb (where "d" represents the explosion-proof housing type, "I" represents Class I coal mine underground equipment, and "Mb" represents the equipment protection level). The clarity analysis module optimizes the dark channel algorithm parameters for low-light underground environments: the local window size is 20×20 pixels. = 0.9.

[0063] The working process of the image enhancement processing device used in underground coal mining dust environments is as follows: A1: The coal mining machine generates dust during coal cutting. The camera captures real-time video images of the target at the coal face. For example, the clarity analysis module detects the average transmittance. =0.52, estimated average dust concentration .

[0064] A2: The controller determines the average dust concentration c> The sound field generator is activated to emit a highly directional sound field toward the target area.

[0065] A3: The controller performs a frequency scan in 200Hz steps within the 500Hz-8kHz frequency range, operating for 15 seconds at each frequency point, and recording the transmittance improvement rate corresponding to each frequency point. The scan shows that the transmittance improvement rate is fastest at x1kHz, therefore x1kHz is selected as the operating frequency.

[0066] A4: The controller adjusts the sound source output power in 5% increments at a working frequency of x1kHz, corresponding to a sound pressure level gradually increasing from 130dB to 150dB. The rate of improvement in intelligibility at each power point is recorded. Testing showed that 20W power, corresponding to a sound pressure level of 150dB, resulted in the fastest improvement in intelligibility; this power value was selected.

[0067] A5: A highly directional sound field acts on the target area, causing suspended coal dust particles to aggregate and settle in the same direction. After 20 seconds of sound field application, the average transmittance increases to [value missing]. =0.78, estimated dust concentration decreased to .

[0068] A6: The controller stops the sound field generator and enters standby monitoring mode.

[0069] A7: The sharpness analysis module enhances the target video image based on the dark channel prior algorithm, further improving the sharpness value. The video footage was restored to clear, and the video recognition system can accurately identify the equipment operating status, personnel operation status, etc. The recognition results are uploaded to the integrated management and control platform through the mining network.

[0070] 2) Image enhancement processing device for dusty environments in mining crushing stations According to the "Guideline for Intelligent Construction of Metal and Non-metal Mines (2025 Edition)," the crushing station is an important location for the installation of electromechanical equipment and requires video surveillance. The camera used is a high-definition network camera with IP68 protection, installed on a steel structure support above the crushing station at a height of 4 meters, providing a top-down view of the crusher feed inlet and belt conveyor area. The sound field generator uses a compression driver and an exponential converter, with an adjustable operating frequency of 500Hz-8kHz, a maximum sound pressure level of 150dB, and a directivity angle of 30°. It is installed 0.5 meters above camera 1, with the sound direction at a 5° angle to the camera's optical axis. The dark channel algorithm parameters for the sharpness analysis module are set as follows: local window size 20×20 pixels. = 0.95, guided filter window radius 60 pixels.

[0071] The working process of the image enhancement processing device for dusty environments in mining crushing plants described above is as follows: B1: The crushing operation generates high concentrations of dust. A camera captures real-time video images of the crushing station. The clarity analysis module detects an average transmittance of 0.50 and estimates the average dust concentration. .

[0072] B2: The controller determines that the average dust concentration c > 50. Start the sound field generator.

[0073] B3: The controller performs a frequency scan in 200Hz steps within the 500Hz-8kHz frequency range, operating for 15 seconds at each frequency point, and recording the transmittance improvement rate corresponding to each frequency point. The scan revealed the fastest transmittance improvement rate at x2kHz, therefore x2kHz was selected as the operating frequency.

[0074] B4: The controller adjusts the sound source output power in 5% increments at a working frequency of x2kHz, corresponding to a sound pressure level gradually increasing from 130dB to 150dB. The rate of improvement in clarity score at each power point is recorded. Testing showed that 16W power, corresponding to a sound pressure level of 142dB, resulted in the fastest improvement in clarity; this power value was selected.

[0075] B5: A highly directional sound field acts on the target area, causing suspended ore dust particles to agglomerate and settle in the same direction. After 20 seconds of sound field application, the average transmittance increases to [value missing]. =0.76, estimated dust concentration decreased to .

[0076] B6: The controller stops the sound field generator and enters standby monitoring mode.

[0077] B7: The clarity analysis module enhances the target video image based on the dark channel prior algorithm, further improving the clarity value and restoring environmental visibility. The recognition system based on the monitoring video detects issues such as crusher belt misalignment and personnel entering dangerous areas, automatically alarms and takes corresponding measures.

[0078] B8: The system will upload the work log (start-up time, working frequency x2kHz, working power 16W, average transmittance change, etc.) to the integrated management and control platform via a wired network.

[0079] Example 2 The above are the image sharpening processing apparatus for dusty environments provided in the embodiments of this specification. Based on the same idea, the embodiments of this specification also provide an image sharpening processing apparatus and method for dusty environments, such as... Figure 2 As shown.

[0080] In step S1, target video images of the target area to be monitored are acquired.

[0081] In practice, target video images can be acquired using image acquisition modules such as cameras, industrial cameras, and 3D vision sensors. The target area can be an open space environment with dust or a closed space environment with dust. Dust environments can be industrial environments with dust interference, such as the steel industry, mining industry, cement industry, and chemical industry.

[0082] In step S2, based on the target video image and the dark channel prior algorithm, the dust concentration value of the target area and the sharpness value of the target video image are determined.

[0083] The dark channel prior algorithm is based on empirical observations of outdoor fog-free images: in most local regions other than the sky (such as a large tree background, a wall, etc.), there exist some pixels in at least one color channel with very low intensity values, that is, the dark channel value of a fog-free image approaches 0.

[0084] Step S2 above may include the following process: S21: Calculate the dark channel image corresponding to the target video image using the dark channel prior algorithm.

[0085] In practice, the physical process of image formation in a dusty environment is first described based on an atmospheric scattering model: ,in, These are observed images with fog. It is a fog-free image to be recovered. It is the atmospheric light value. Transmittance is the percentage of light that passes through fog and reaches the camera.

[0086] Then, the dark channel prior algorithm is used to calculate the dark channel image. The formula for the dark channel prior algorithm is: ,in, This represents the three color channels: RGB. Represented in pixels The local region centered on the object, which is typically sized as... The window, It is a positive integer, and The value ranges from 15 to 20, for example, a local region size of 15×15 pixels. According to the dark channel prior theory, the dark channel value of a hazy image approaches 0, i.e. The inner min operation takes the minimum value of each channel within the local region, while the outer min operation takes the minimum value among the three minimum values ​​of the channels.

[0087] S22: Based on the first pixel in the dark channel image with a preset ratio, determine the second pixel in the target video image corresponding to the position of the first pixel, and determine the atmospheric light value based on the pixel value of the second pixel.

[0088] The first pixel represents the pixel with the highest fog density, which is also the brightest pixel in the dark channel image. The preset ratio can be determined based on the specific application scenario. The atmospheric light value can be determined based on the pixel value of the second pixel by selecting the maximum value from the RGB three channels as the atmospheric light value A.

[0089] In the dark channel image, select the brightest 0.1% pixel (i.e., the first pixel), and then map the 0.1% pixel to the same position pixel in the original image (i.e., the target video image being captured) (i.e., the second pixel). In the second pixel of the original image, select the largest value among the three RGB channels as the atmospheric light value A.

[0090] S23: The transmittance of the target area is calculated using an atmospheric scattering model based on atmospheric light values.

[0091] In practice, by calculating the dark channel of a foggy image and combining it with an atmospheric scattering model, the following formula can be used: Estimate the initial transmittance of the target region, where, It is a factor that retains a small amount of fog. The value can be 0.95. Transmittance is also called light transmittance.

[0092] The refined average transmittance is determined based on the initial transmittance. The average transmittance is used as the transmittance of the target area.

[0093] S24: Determine the dust concentration value of the target area based on the transmittance of the target area.

[0094] The transmittance reflects the dust concentration in a dusty environment. Based on the transmittance of the target area estimated using the dark channel prior algorithm, and combined with the empirical formula of transmittance minus dust concentration... The average dust concentration in the target area can be calculated, where Average dust concentration, in units of , The average projection rate of the target area is used, and the calculated average dust concentration is used as the dust concentration value of the target area.

[0095] S25: Based on the gray values ​​of two pixels that differ by two units in the target video image, the sharpness value of the target video image is calculated using the Brenner gradient function.

[0096] The formula for the Brenner gradient function is as follows: , and This refers to the grayscale values ​​of two pixels that differ by two units.

[0097] This function evaluates the sharpness of a target video image by calculating the mean squared error between two pixels. The calculated sharpness evaluation function value is then normalized, and the average of the normalized values ​​is taken from the sharpness evaluation function values ​​under the same dust-free environment. As the original reference for the sharpness score (i.e., sharpness value), the average value at time t is used. Compare the clarity of the target video image at different times or at different sound field frequencies.

[0098] Continue to participate Figure 2 It can be seen that in step S3, when the dust concentration value is greater than or equal to the preset dust concentration threshold, and / or the clarity value of the target video image is less than or equal to the preset clarity threshold, a directional sound field is emitted towards the target area until the dust concentration value is less than the preset dust concentration threshold and the clarity value of the target video image is greater than the preset clarity threshold.

[0099] According to step S3, the dust concentration value of the target area and the clarity value of the target video image are monitored. When the dust concentration value is greater than or equal to the preset dust concentration threshold and the clarity value of the target video image is less than or equal to the preset clarity threshold, or when the dust concentration value is greater than or equal to the preset dust concentration threshold and the clarity value of the target video image is less than or equal to the preset clarity threshold, a directional sound field is emitted towards the target area, thereby effectively reducing suspended dust in the dusty environment and achieving the purpose of improving environmental visibility.

[0100] In practice, a directional sound field can be emitted towards the target area using a sound field generator. When emitting the directional sound field, the operating frequency and output power of the directional sound field transmitter can be set according to the dust concentration and image sharpness values ​​at the monitoring time, until the dust concentration is less than a preset dust concentration threshold and the image sharpness of the target video image is greater than a preset image sharpness threshold. Alternatively, the operating frequency and output power of the sound field generator can be adjusted during operation based on changes in the dust concentration and image sharpness values ​​in the environment, until the dust concentration is less than a preset dust concentration threshold and the image sharpness of the target video image is greater than a preset image sharpness threshold.

[0101] In addition, if it is necessary to monitor target equipment (such as key components, robotic arms, etc.) in the target area, a highly directional sound field can be sent to the area where the target equipment is located in the target area.

[0102] Furthermore, the processing in step S3 above can take many forms. The following is one optional processing method, which can be found in the following steps S31-S33: In step S31, when the dust concentration value is greater than or equal to a preset dust concentration threshold, and / or the clarity value of the target video image is less than or equal to a preset clarity threshold, a directional sound field is emitted toward the target area based on a first operating frequency and a first output power, wherein the first operating frequency and the first output power are initial operating frequencies and initial output power set according to the dust concentration value and clarity value at the monitoring time.

[0103] In step S32, the target area is frequency scanned with a preset step size based on a preset frequency range. The second operating frequency of the directional sound field is determined based on the rate of increase of the sharpness value during the frequency scan. The second output power of the directional sound field is determined based on the determined second operating frequency and the rate of increase of the sharpness value. The directional sound field is then emitted to the target area based on the second operating frequency and the second output power.

[0104] The preset frequency range can be 500Hz-8kHz, and the preset step size can be 200Hz. When performing a frequency scan, you can work for 10-20 seconds at each frequency point to ensure more accurate scan results.

[0105] In practice, according to the above steps S31 and S32, the target area can be frequency scanned after a directional sound field is emitted to the target area based on the first operating frequency and the first output power, or the target area can be frequency scanned at the same time as the directional sound field is emitted to the target area based on the first operating frequency and the first output power.

[0106] In step S33, a directional sound field is emitted toward the target area based on the second operating frequency and the second output power until the dust concentration value is less than a preset dust concentration threshold and the clarity value of the target video image is greater than a preset clarity threshold, at which point the emission of the directional sound field toward the target area stops.

[0107] By performing frequency scanning on the target area to determine the operating frequency and output power that better match the particle size of the dust in the dusty environment, the acoustic agglomeration effect of suspended particles can be further accelerated, thereby further and rapidly improving the visibility of the dusty environment.

[0108] Furthermore, the processing in step S3 above can take many forms. The following provides another optional processing method, which can be found in steps S34-S36: In step S34, when the dust concentration value is greater than or equal to a preset dust concentration threshold, and / or the clarity value of the target video image is less than or equal to a preset clarity threshold, a directional sound field is emitted toward the target area based on a first operating frequency and a first output power, wherein the first operating frequency and the first output power are initial operating frequencies and initial output power set according to the dust concentration value and clarity value at the monitoring time.

[0109] In step S35, the target area is frequency scanned with a preset step size based on a preset frequency range. The second operating frequency of the directional sound field is determined based on the rate of increase of the sharpness value during the frequency scan. The second output power of the directional sound field is determined based on the determined second operating frequency and the rate of increase of the sharpness value. The directional sound field is then emitted to the target area based on the second operating frequency and the second output power.

[0110] The processing of steps S34-S35 above can be referred to the processing of steps S31-S32, and will not be repeated here.

[0111] In step S36, the target area is cyclically scanned with a preset step size based on a preset frequency range. The second operating frequency is updated based on the rate of increase in the sharpness value during the frequency scanning process. The second output power is updated based on the updated second operating frequency and the rate of increase in the sharpness value. A directional sound field is emitted towards the target area based on the updated second operating frequency and the updated second output power until the dust concentration value is less than a preset dust concentration threshold and the sharpness value of the target video image is greater than a preset sharpness threshold, at which point the emission of the directional sound field towards the target area stops.

[0112] Step S36 allows for dynamic adjustment of the second operating frequency and the second output power at any time, achieving a dynamic optimization process. This method enables the operating frequency of the sound field generator to match the current dust particle size in a timely manner, thereby further accelerating the acoustic agglomeration effect of suspended particles and rapidly improving the visibility of the dusty environment. The entire adjustment process can be completed within 20 seconds, and the environmental visibility is significantly improved within 15-30 seconds, resulting in a significant improvement in the clarity of the acquired video image.

[0113] Furthermore, such as Figure 3 As shown, after step S3, the image sharpening processing method in this application further includes: Step S4: Perform image enhancement processing on the target video image to obtain the enhanced target video image.

[0114] In practice, deep neural networks such as 3D convolutional networks or diffusion models can be used to enhance the target video image, deblurring and denoising it, thereby improving the clarity of the target video image.

[0115] Alternatively, you can choose a neural network with lower computational cost, such as TensorRT, OpenVINO, or TNN, to perform image enhancement processing on the target video image.

[0116] The methods for enhancing target video images using deep neural networks, diffusion models, and computationally less computationally intensive neural networks can be implemented using existing methods, and will not be elaborated upon here.

[0117] Furthermore, the target video image can be enhanced based on the dark channel prior algorithm to further restore image details, thereby obtaining a clearer target video image. Specifically, based on the atmospheric scattering model... It can be obtained According to the formula Image enhancement processing is performed on the target video image to restore image details, thereby improving image clarity. It is the image enhancement threshold, which can be set to 0.1 to avoid noise amplification caused by a denominator that is too small.

[0118] Step S5: Perform video recognition processing based on the target video image after image enhancement processing.

[0119] After acquiring the target video image after image enhancement processing, it is sent to the background video recognition system. Based on the acquired target video image, the background video recognition system can more accurately monitor the operating status of equipment, the working status of personnel, etc., thereby improving industrial production efficiency.

[0120] For the parts of the image sharpening processing method for dusty environments in this application embodiment that are not described in detail, please refer to the image sharpening processing device for dusty environments in Embodiment 1. The two embodiments can be referred to each other, and will not be repeated here.

[0121] The embodiments of this application have been described above with reference to the accompanying drawings. However, this application is not limited to the specific embodiments described above. The specific embodiments described above are merely illustrative and not restrictive. Those skilled in the art can make many other forms under the guidance of this application without departing from the spirit and scope of the claims, and all of these forms are within the protection scope of this application.

Claims

1. An image sharpening processing device for dusty environments, characterized in that, The device includes: The image acquisition module is used to acquire target video images of the target area to be monitored. A sound field generator is used to emit a sound field toward the target area according to the instructions of the controller. The sound field generator emits sound in the direction of the target area, and the distance between the sound field generator and the image acquisition module, as well as the angle between the sound field generator and the optical axis of the image acquisition module, make the sound field generated by the sound field generator cover the target area. A sharpness analysis module, electrically connected to the image acquisition module, is used to acquire the target video image and determine the dust concentration value of the target area and the sharpness value of the target video image based on the target video image and the dark channel prior algorithm. The controller is electrically connected to the clarity analysis module and the sound field generator respectively, and is used to receive the dust concentration value and clarity value determined by the clarity analysis module, and control the operating frequency and / or output power of the sound field generator based on the received dust concentration value, clarity value and preset dust concentration threshold and clarity threshold. The power supply module is used to supply power to the image acquisition module, the sharpness analysis module, and the controller.

2. The image sharpening processing device for dusty environments according to claim 1, characterized in that, The sound field generator is an adjustable directional sound field generator, used to emit a sound field to the target area and a strongly directional sound field to the target device in the target area according to the controller's instructions.

3. The image sharpening processing device for dusty environments according to claim 1, characterized in that, The sound field generator includes a compression driver and an exponentially type cylinder adapted to the compression driver.

4. The image sharpening processing device for dusty environments according to claim 1, characterized in that, The sharpness analysis module is also used to perform image enhancement processing on the target video image to obtain the target video image after image enhancement processing.

5. The image sharpening processing device for dusty environments according to claim 1, characterized in that, The device also includes a control switch, which is embedded in the housing of the image acquisition module and electrically connected to the controller for controlling the controller to turn on and off.

6. The image sharpening processing apparatus for dusty environments according to claim 1, characterized in that, The image acquisition module has a protection rating of at least IP67, and the housing of the image acquisition module integrates heat sinks and air-cooling channels.

7. An image sharpening method for dusty environments, characterized in that, The method includes: Acquire target video images of the target area to be monitored; Based on the target video image and the dark channel prior algorithm, the dust concentration value of the target area and the sharpness value of the target video image are determined. When the dust concentration value is greater than or equal to a preset dust concentration threshold, and / or the clarity value of the target video image is less than or equal to a preset clarity threshold, a directional sound field is emitted toward the target area until the dust concentration value is less than the preset dust concentration threshold and the clarity value of the target video image is greater than the preset clarity threshold.

8. The image sharpening method for dusty environments according to claim 7, characterized in that, The step of determining the dust concentration value of the target region and the sharpness value of the target video image based on the target video image and the dark channel prior algorithm includes: The dark channel prior algorithm is used to calculate the dark channel image corresponding to the target video image; Based on a first pixel in the dark channel image at a preset ratio, a second pixel in the target video image corresponding to the position of the first pixel is determined, and the atmospheric light value is determined based on the pixel value of the second pixel. The first pixel is the pixel with the highest fog concentration. Based on the atmospheric light value, the transmittance of the target area is calculated using an atmospheric scattering model; The dust concentration value of the target area is determined based on the transmittance of the target area. The sharpness value of the target video image is calculated using the Brenner gradient function based on the grayscale values ​​of two pixels that differ by two units in the target video image.

9. The image sharpening method for dusty environments according to claim 7, characterized in that, The step of emitting a directional sound field toward the target area when the dust concentration value is greater than or equal to a preset dust concentration threshold, and / or the clarity value of the target video image is less than or equal to a preset clarity threshold, until the dust concentration value is less than the preset dust concentration threshold and the clarity value of the target video image is greater than the preset clarity threshold, includes: When the dust concentration value is greater than or equal to a preset dust concentration threshold, and / or the clarity value of the target video image is less than or equal to a preset clarity threshold, a directional sound field is emitted toward the target area based on a first operating frequency and a first output power. The first operating frequency and the first output power are initial operating frequencies and initial output power set according to the dust concentration value and clarity value at the monitoring time. The target area is frequency scanned based on a preset frequency range and a preset step size. The second operating frequency of the directional sound field is determined based on the rate of increase in the sharpness value during the frequency scan. The second output power of the directional sound field is determined based on the determined second operating frequency and the rate of increase in the sharpness value. The directional sound field is then emitted toward the target area based on the second operating frequency and the second output power. The target area is cyclically scanned with a preset frequency range and a preset step size. The second operating frequency is updated based on the rate of increase in the sharpness value during the frequency scan. The second output power is updated based on the updated second operating frequency and the rate of increase in the sharpness value. A directional sound field is emitted towards the target area based on the updated second operating frequency and the updated second output power. The directional sound field is emitted towards the target area until the dust concentration value is less than a preset dust concentration threshold and the sharpness value of the target video image is greater than a preset sharpness threshold, at which point the emission of the directional sound field towards the target area stops.

10. The image sharpening method for dusty environments according to claim 7, characterized in that, The method further includes emitting a directional sound field towards the target area until the dust concentration value is less than a preset dust concentration threshold and the sharpness value of the target video image is greater than a preset sharpness threshold. The target video image is subjected to image enhancement processing to obtain the target video image after image enhancement processing; Video recognition processing is performed on the target video image after image enhancement.

Images