A control method for linking dust image analysis and fog gun pointing

Abstract

This invention discloses a control method for pointing a fog cannon based on dust image analysis, relating to the fields of dust image analysis and fog cannon dust suppression control. The method includes: acquiring a continuous sequence of dust images; identifying and associating dust regions; determining the current center position of the dust region, the direction of dust movement, and the dust movement speed; determining an initial target angle based on the detector attitude and angle mapping relationship; obtaining the fog cannon response delay based on the angle difference between the initial target angle and the current angle of the fog cannon; predicting the dust position based on the dust movement direction, dust movement speed, and fog cannon response delay; correcting the initial target angle based on the position offset to obtain the target horizontal angle and target pitch angle; controlling the fog cannon rotation; and controlling the fog cannon spray when the real-time angle meets a preset angle deviation condition. This invention can compensate for pointing deviations caused by dust movement and fog cannon response delay, improving the pointing accuracy of the fog cannon in dynamic dust regions.

Description

Technical Field

[0001] The present invention relates to the technical field of dust image analysis and fog cannon dust suppression control, in particular to a control method for linking fog cannon pointing based on dust image analysis. Background Art

[0002] With the continuous improvement of requirements for dust pollution control in scenarios such as industrial production, port loading and unloading, mining, building construction, and material storage yards, dust monitoring and dust suppression equipment has gradually evolved from the traditional fixed-point monitoring and manual start-stop mode to the direction of image perception, automatic identification and linkage control. Existing dust control systems usually use dust sensors, visible light imaging devices or binocular vision devices to obtain the dust concentration, spatial distribution or image features in the operation area, and control dust suppression equipment such as fog cannons and water mist generators to perform spraying according to the detection results. Compared with the continuous spraying mode in a fixed direction, adjusting the spraying area of the fog cannon based on visual information can reduce invalid spraying and improve the utilization rate of water resources and energy. However, in the actual open-air operation environment, the dust area usually continues to undergo spatial migration due to the influence of wind field, material movement and equipment operation disturbance, and it takes a certain mechanical response time for the fog cannon to rotate from the current posture to the target posture. If the pointing direction of the fog cannon is determined only according to the current position of the dust identified at the image acquisition moment, when the fog cannon completes the steering, the dust area has already changed its position, which easily leads to a spatial deviation between the actual spraying direction of the fog cannon and the actual position of the dust. Especially when the dust moves at a relatively high speed or the rotation angle of the fog cannon is relatively large, this type of dynamic pointing error caused by the combination of dust movement and equipment response delay is more obvious. Therefore, how to correlate the motion information in the dust image, the steering response characteristics of the fog cannon, and the target position at the spraying moment is a problem that needs to be further solved to improve the accuracy of vision-linked dust suppression control.

[0003] CN111530203A discloses an intelligent dust monitoring and control system and its dust suppression method. The system includes a sensing module, a vision module, a decision-making module, a control module, and an execution module. The sensing module acquires environmental parameters of the area to be monitored and controlled, as well as dust concentration values ​​at measurement points. The execution module suppresses dust in the area. The control module controls the operation of each dust suppression device in the execution module. The vision module acquires multi-view images of various locations within the area. The decision-making module, connected to the vision module, grids the area to be monitored and controlled to obtain a three-dimensional gridded field of view. The decision-making module, connected to the sensing module, obtains the dust concentration distribution in the area and thus the control strategy for the execution module. The decision-making module, connected to the control module, enables the control module to control the operation of each dust suppression device according to the control strategy. This system achieves gridded monitoring and intelligent control of dust, precise treatment, and reduced water and electricity consumption. This technology can formulate corresponding dust suppression control strategies based on dust concentration distribution. However, its key features lie in regional gridding, dust concentration distribution acquisition, and coordinated control of multiple dust suppression devices. It does not further address the inter-frame motion state of dust regions in continuous dust images, nor does it incorporate the response delay caused by the fog cannon rotating from its current angle to the target angle to predict the dust position when the fog cannon completes its turn. Therefore, when the dust region continuously moves, the control position directly determined based on the current dust distribution may deviate from the dust position at the subsequent actual spraying moment.

[0004] CN118115440B discloses a binocular vision-based precision dust suppression system using a fog cannon, relating to the field of binocular vision imaging technology. The system includes an environmental monitoring module that monitors and acquires the distribution of dust in the environment in real time and generates a dust distribution map; a data integration module that processes and analyzes the image based on the dust distribution map and generates analysis results; and an intelligent control module that generates control strategies and performs regulation. It also includes a water mist generator and an electrostatic generator placed around it; the electrostatic generator generates a strong electric field to acquire charged water droplets and charged particles; a charged dust-water droplet complex is formed, and the sedimentation of the dust-water droplet complex is controlled by controlling the strong electric field. This invention achieves precise and intelligent fog cannon spraying by real-time monitoring of dust concentration and distribution in the environment, using image processing technology for precise analysis, and then generating control strategies; effectively reducing the amount of fog cannon sprayed, improving dust suppression efficiency while reducing energy consumption. While this technology utilizes visual analysis results to generate fog cannon control strategies, it primarily focuses on dust distribution detection and the electrostatic settling process between charged water droplets and charged particles. It lacks further correlation between the direction and speed of dust movement across consecutive image frames, the response delay corresponding to the fog cannon's turning process, and the positional migration of the dust region during the response delay. Furthermore, when there is a time interval between the detected position and the actual spraying time, the spraying direction determined based on the current dust distribution may still lag behind the actual dust movement position. Summary of the Invention

[0005] The purpose of this section is to outline some aspects of the embodiments of the present invention and to briefly introduce some preferred embodiments. Simplifications or omissions may be made in this section, as well as in the abstract and title of this application. For example, a control method for the pointing of a fog cannon based on dust image analysis is provided, which includes... The system acquires a sequence of dust images continuously collected by an image-type dust detector, identifies and correlates dust regions in the image sequence, and determines the current center position of the dust region as well as the corresponding dust movement direction and speed. The horizontal and vertical angles of the image-type dust detector are obtained. Based on the current center position of the dust area, the horizontal and vertical angles of the detector, and the pre-calibrated angle mapping relationship between the image-type dust detector and the fog cannon, the initial target horizontal and vertical angles are determined. The fog cannon response delay is obtained based on the angle difference between the initial target horizontal and vertical angles and the current angle of the fog cannon. The predicted dust position corresponding to the dust area is predicted based on the dust movement direction, dust movement speed and fog cannon response delay. The initial target horizontal angle and initial target pitch angle are corrected based on the position offset of the predicted dust position relative to the current dust area center position and the correspondence between the preset position offset and angle correction amount, so as to obtain the target horizontal angle and target pitch angle. Control the fog cannon to rotate towards the target at the horizontal and vertical angles, and acquire the real-time horizontal and vertical angles of the fog cannon; when the real-time horizontal and vertical angles meet the preset angle deviation conditions of the corresponding target angles, control the fog cannon to spray.

[0006] Compared with existing technologies, the advantages of this invention are as follows: By utilizing inter-frame correlation recognition of image sequences, not only is the current center position of the dust area captured, but its direction of motion and speed are also extracted, upgrading the fog cannon control from static fixed-point spraying to dynamic tracking spraying with predictive capabilities, fundamentally overcoming the target drift problem caused by dust migration; Based on this, by combining the detector installation angle with the pre-calibrated spatial mapping relationship, the image coordinates are converted into the initial target angle of the fog cannon, and the fog cannon response delay is dynamically calculated based on the difference between the target angle and the current angle, quantifying the inertial lag of the actuator into a precise time reference, avoiding system deviations caused by fixed empirical values; Subsequently, the dust motion vector and the delay are spatiotemporally fused to predict the dust's movement in the fog cannon. The actual drift position of the fog cannon during rotation is determined, and a linear conversion relationship between image domain pixel offset and angle correction is established to accurately compensate for the initial target angle. This eliminates system-level aiming errors caused by time delay at the mechanistic level, ensuring that the actual landing point of the fog cannon matches the actual position of the dust in the same domain. Finally, real-time angle feedback and preset deviation condition judgment are introduced during the fog cannon rotation process to ensure that the spray command is triggered only when the attitude converges to the neighborhood of the target angle. This not only guarantees the spray hit rate but also takes into account the timeliness of response and the system's anti-interference ability through a tolerance mechanism. This invention significantly improves the pointing accuracy and dust suppression efficiency of fog cannons in industrial dust scenarios with large lag and high dynamics, and achieves synergistic optimization of control predictability, execution accuracy, and system robustness. Attached Figure Description

[0007] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort. Wherein: Figure 1 This is a flowchart of a control method for directing fog cannons based on dust image analysis. Detailed Implementation

[0008] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments.

[0009] Based on the embodiments of this invention, all other embodiments obtained by those skilled in the art without inventive effort should fall within the scope of protection of this invention.

[0010] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the spirit of the invention. Therefore, the invention is not limited to the specific embodiments disclosed below.

[0011] As mentioned in the background section, while existing technologies can determine dust suppression areas through image acquisition and dust distribution analysis, and control fog cannons accordingly, they typically rely primarily on the current dust position to generate control strategies. They do not adequately consider the positional shift of the dust area during fog cannon turning, nor do they correlate the dust movement direction and speed with the fog cannon's response delay. Therefore, when the dust is in continuous motion or the fog cannon rotates at a large angle, the target direction can easily become inconsistent with the actual dust position. To address these issues, this invention provides a control method for fog cannon pointing based on dust image analysis.

[0012] Reference Figure 1 , Figure 1 This is a flowchart illustrating a control method for directing a fog cannon based on dust image analysis, according to an embodiment of the present invention. Figure 1 As shown, a control method for directing a fog cannon based on dust image analysis includes: S1: Acquire a sequence of dust images continuously collected by the image-type dust detector, identify and correlate dust regions in the dust image sequence, and determine the current center position of the dust region as well as the corresponding dust movement direction and speed.

[0013] S1.1: Acquire dust images collected by the image-type dust detector at multiple acquisition times, arrange the dust images in chronological order of acquisition times to obtain a dust image sequence; each frame of the dust image sequence carries a corresponding acquisition time tag.

[0014] S1.2: Perform grayscale processing and edge enhancement processing on each frame of the dust image sequence to obtain a preprocessed dust image; perform threshold segmentation on the preprocessed dust image, and extract connected regions with grayscale values ​​greater than a preset grayscale threshold and connected region areas greater than a preset area threshold as candidate dust regions.

[0015] Specifically, the method for obtaining the preprocessed dust image is as follows: the red, green, and blue color channel components of the dust image are weighted and summed according to preset weights to obtain the grayscale image corresponding to the dust image; a preset edge enhancement operator is used to perform convolution processing on the grayscale image to increase the local variation amplitude of pixel grayscale values ​​in the grayscale image, thus obtaining the preprocessed dust image.

[0016] Preferably, a preset edge enhancement operator is used to perform convolution processing on the grayscale image. Specifically, a preset Laplacian operator is obtained as the edge enhancement operator, wherein the edge enhancement operator is a three-row, three-column matrix (the matrix adopts a four-neighborhood form with a positive center coefficient, i.e., [[0, -1, 0], [-1, 4, -1], [0, -1, 0]]); the edge enhancement operator is slid across the grayscale image with a step size of 1 pixel; for each pixel in the grayscale image, the grayscale value of the three-row, three-column neighborhood pixels centered on that pixel is taken, and multiplied and summed with the value at the corresponding position of the edge enhancement operator to obtain the convolution result value corresponding to that pixel; the original grayscale value corresponding to each pixel in the grayscale image is added to the convolution result value corresponding to that pixel to obtain the enhanced grayscale value corresponding to that pixel; the enhanced grayscale values ​​of each pixel constitute the preprocessed dust image.

[0017] Furthermore, the method for extracting connected regions is as follows: pixels with gray values ​​greater than a preset gray threshold in the preprocessed dust image are marked to obtain a binary image; the marked pixels in the binary image are traversed and merged according to the eight-neighborhood connectivity rule, and regions that are adjacent and belong to the same marked pixels are grouped into the same connected region, and the number of pixels contained in each connected region is counted as the area of ​​the corresponding connected region.

[0018] Specifically, the marked pixels in the binary image are traversed and merged according to the eight-neighbor connectivity rule. This involves traversing each marked pixel in the binary image from left to right and top to bottom. For each pixel, it is determined whether there are any marked pixels already merged into a connected region within its eight-neighbor domain. If so, the current pixel is merged into that connected region. If multiple marked pixels have been merged into different connected regions within its eight-neighbor domain, these regions are merged into a single connected region, and the current pixel is merged into the merged region. If no pixel has been merged into any connected region within its eight-neighbor domain, a new connected region is created, and the current pixel is merged into this new region. After traversing all marked pixels in the binary image, multiple connected regions are obtained.

[0019] It should be noted that the preset grayscale threshold is determined based on the mean grayscale value and standard deviation of multiple background images acquired by the image-type dust detector at the same installation location under dust-free conditions. The sum of the mean grayscale value and the standard deviation of the grayscale value at a preset multiple is used as the preset grayscale threshold, which is adjusted according to changes in ambient lighting conditions. The preset area threshold is calculated based on the actual physical size corresponding to the focal length, pixel size, and preset minimum dust cloud image area of ​​the image-type dust detector, according to imaging geometry. The preset area threshold is recalibrated according to changes in the installation height and field of view of the image-type dust detector. The preset multiple ranges from 1.5 to 3.0. When ambient lighting changes drastically, the upper limit of the preset multiple is used to improve anti-interference capability; when ambient lighting is stable, the lower limit of the preset multiple is used to improve the ability to capture weak dust signals.

[0020] S1.3: For candidate dust regions, calculate the aspect ratio, edge diffusion, and the difference between the mean gray value inside the candidate dust region and the mean gray value of the external background. Based on the matching results of the aspect ratio, edge diffusion, and difference with the preset dust feature conditions, select candidate dust regions that meet the preset dust feature conditions from the candidate dust regions and use them as dust regions. Each frame of dust image in the dust image sequence corresponds to at least one dust region.

[0021] Furthermore, the matching result between the difference and the preset dust feature conditions is as follows: when the aspect ratio of the candidate dust region is within the aspect ratio range corresponding to the preset dust feature conditions, and the edge diffusion of the candidate dust region is greater than the edge diffusion threshold corresponding to the preset dust feature conditions, and the difference between the mean gray value inside the candidate dust region and the mean gray value of the external background is within the difference range corresponding to the preset dust feature conditions, then the candidate dust region is determined to meet the preset dust feature conditions; otherwise, the candidate dust region is removed.

[0022] It should be noted that aspect ratio, edge diffusion, and difference represent the physical characteristics of dust clumps from three independent dimensions: the outline shape of the candidate dust region, the boundary transition characteristics, and the internal and external light transmittance. There is no inclusion or substitution relationship among the three conditions. Only when the candidate dust region meets all three conditions can other non-dust interference targets that only meet one or two characteristic conditions be excluded. Therefore, the matching result adopts the judgment method of meeting all three conditions at the same time.

[0023] Preferably, the preset dust feature conditions are as follows: the aspect ratio range is statistically obtained based on the irregular elongated outlines of dust clumps affected by the airflow diffusion direction in historical dust images, used to distinguish fixed-shape targets such as vehicles or people with an aspect ratio close to 1 and regular boundaries; the edge diffusion threshold is determined based on the historical statistical distribution of the gray-level gradient of the boundary pixels of the candidate dust region, and the edge diffusion is measured by the pixel width of the gradual change in gray-level value at the boundary of the candidate dust region, used to distinguish fixed-shape targets with clear and sharp boundaries; the difference range is determined based on the historical statistical distribution of gray-level difference values ​​corresponding to the semi-transparent characteristics of dust clumps, used to distinguish opaque objects with gray-level difference values ​​greater than the upper limit of the difference range and background noise with gray-level difference values ​​less than the lower limit of the difference range.

[0024] It should be noted that the dust region identification in S1.3 is the result of processing a single frame of dust image in the dust image sequence. The corresponding dust regions in different frames of dust image in the dust image sequence have not yet been established. Since the dust region is a non-rigid region whose edge contour changes over time in a drifting and diffusion state, it has recognition differences from a fixed-shape target. Directly using the nearest neighbor method to establish a correspondence between dust regions in different frames of dust images is prone to association errors due to changes in the shape of the dust region. Therefore, based on the dust region identification in S1.3, through steps S1.4 to S1.7, a comprehensive judgment is made on the position, area, and grayscale distribution of the dust region to establish an inter-frame association relationship between dust regions between two adjacent frames of dust image in the dust image sequence.

[0025] S1.4: For two adjacent dust image frames in a dust image sequence, calculate the center position, area, and grayscale histogram of the dust region in the previous dust image, and the center position, area, and grayscale histogram of each dust region in the current dust image; based on the center position of the dust region in the previous dust image and a preset search radius, determine the search window range in the current dust image; within the search window range, filter out dust regions whose center position distance from the dust region in the previous dust image is less than a preset distance threshold, whose area difference is less than a preset area difference threshold, and whose grayscale histogram similarity is greater than a preset similarity threshold, as associated dust regions of the dust region in the previous dust image in the current dust image; when there is no dust region within the search window range that meets the above three conditions, it is determined that there is no associated dust region of the dust region in the previous dust image in the current dust image.

[0026] It should be noted that the grayscale histogram similarity is calculated using Bach distance or histogram intersection to reduce the interference of illumination changes on the inter-frame correlation results. The preset search radius is determined based on the product of the acquisition time difference between two adjacent dust image frames in the dust image sequence and the preset maximum dust drift velocity, and a preset redundancy is added to ensure that the preset search radius covers the maximum displacement range of the dust region within the acquisition interval between two adjacent dust image frames. The preset distance threshold is determined based on the statistical distribution of the center position offset of the same dust region between two adjacent dust image frames in the historical dust image sequence, and its value is less than the preset search radius. The preset area difference threshold is determined based on the statistical distribution of the area change of the same dust region between two adjacent dust image frames in the historical dust image sequence. The preset similarity threshold is determined based on the statistical distribution of the grayscale histogram similarity of the same dust region between two adjacent dust image frames in the historical dust image sequence.

[0027] Furthermore, the method for determining the search window range is as follows: using the center position of the dust region in the previous frame dust image as the center and a preset search radius as the radius, a circular area is determined in the pixel coordinate system of the current frame dust image, and this circular area is used as the search window range. The center position of each dust region in the current frame dust image is then checked to see if it falls within the search window range, and dust regions that fall within the search window range are considered candidate association objects. Specifically, if a dust region in the current frame dust image does not fall within the search window range, it is determined that this dust region has no association relationship with the dust region in the previous frame dust image. This dust region is considered a newly appearing dust region in the current frame dust image and is not included in the current inter-frame association result. However, in the association process of subsequent adjacent dust images, this dust region is treated as a new dust region from the previous frame dust image and participates in the association judgment.

[0028] Furthermore, the method for screening associated dust regions is as follows: For each candidate associated object within the search window, calculate the distance between it and the center position of the dust region in the previous frame dust image, the area difference between it and the area of ​​the dust region in the previous frame dust image, and the similarity between it and the grayscale histogram of the dust region in the previous frame dust image. Candidate associated objects that simultaneously meet the three conditions of distance less than a preset distance threshold, area difference less than a preset area difference threshold, and similarity greater than a preset similarity threshold are identified as associated dust regions.

[0029] Specifically, if a candidate associated object within the search window does not meet at least one of the following three conditions: distance less than a preset distance threshold, area difference less than a preset area difference threshold, or similarity greater than a preset similarity threshold, then the candidate associated object is determined to have no association with the dust region in the previous frame dust image. If all candidate associated objects within the search window do not meet the three conditions, then the dust region in the previous frame dust image is determined to have no associated dust region in the current frame dust image. The processing method is the same as in the case where there is no dust region meeting the conditions within the search window mentioned above in this step.

[0030] In an optional embodiment, when there are multiple dust regions that meet the association conditions within the search window, the dust region with the smallest distance from the center position of the dust region in the previous frame dust image is selected as the associated dust region, so as to avoid association ambiguity when multiple dust regions exist in the same dust image at the same time.

[0031] S1.5: Repeat the association process (step S1.4) sequentially according to the arrangement order of two adjacent dust images in the dust image sequence to obtain multiple dust regions belonging to the same dust region in the dust image sequence and their corresponding center positions, as the inter-frame association result.

[0032] It should be noted that when no associated dust region is identified between two adjacent dust images in the dust image sequence in the inter-frame association result, the tracking of the corresponding dust region is deemed interrupted. The dust regions in the inter-frame association result before the tracking interruption and their corresponding center positions are retained and used for the calculation of motion velocity and motion direction components in subsequent S1.6. The dust regions that appear after the tracking interruption are no longer included. When the inter-frame association result does not contain the associated dust region corresponding to the last dust image in the dust image sequence, it is further determined whether there is an unassociated candidate dust region in the last dust image. If there is, the center position of the candidate dust region is determined as the current dust region center position, and the dust motion velocity and dust motion direction are set to the preset system default value or the motion parameters corresponding to the most recent valid inter-frame association result. If there is no dust target, it is determined that there is no dust target at the current time and it does not participate in the subsequent S2 to S4 steps.

[0033] Based on the inter-frame correlation results, the center position of the dust region corresponding to the last dust image in the dust image sequence is determined as the center position of the current dust region.

[0034] S1.6: Based on the center positions of multiple dust regions belonging to the same dust region in the inter-frame association results, calculate the position difference between the center positions of two adjacent dust regions and the acquisition time difference between the acquisition time tags corresponding to the dust regions in two adjacent frames; based on the position difference and acquisition time difference, calculate the motion velocity component and motion direction component of the center position of the dust region between two adjacent dust images.

[0035] It should be noted that when there are only two frames of dust regions belonging to the same dust region in the inter-frame correlation results, the motion velocity component and motion direction component are the only set of components calculated between the two dust regions.

[0036] S1.7: Perform weighted averaging on multiple sets of motion velocity components and motion direction components in the inter-frame correlation results to obtain the dust motion velocity and dust motion direction.

[0037] Preferably, the weights of each set of motion velocity components and motion direction components are determined based on the time interval between the acquisition time of the two adjacent dust images and the acquisition time of the last dust image in the dust image sequence. The smaller the time interval, the greater the corresponding weight. When there is only one set of motion velocity components and motion direction components in the inter-frame association result, the set of motion velocity components and motion direction components are determined as dust motion velocity and dust motion direction, respectively.

[0038] Furthermore, the weighting method ensures that the velocity and direction components closer to the current acquisition time contribute more to the dust velocity and direction, reducing the interference of historical dust region morphology changes on the current dust direction and velocity calculation results.

[0039] S2: Obtain the horizontal and vertical angles of the image-type dust detector. Based on the current center position of the dust area, the horizontal and vertical angles of the detector, and the pre-calibrated angle mapping relationship between the image-type dust detector and the fog cannon, determine the initial target horizontal and vertical angles. Determine the fog cannon response delay based on the angle difference between the initial target horizontal angle, the initial target vertical angle and the current angle of the fog cannon.

[0040] It should be noted that the current center position of the dust area is the center position of the dust area corresponding to the last dust image in the S1 inter-frame correlation result, and the current center position of the dust area is used as the input parameter for the angle conversion in this step.

[0041] S2.1: Obtain the horizontal and vertical angles of the image-type dust detector at the moment when the dust image is captured at the center position of the current dust area.

[0042] Specifically, the image-type dust detector is mounted on a gimbal, wherein the horizontal rotation mechanism and the pitch rotation mechanism of the gimbal are driven by motors. When the image-type dust detector acquires each frame of dust image, the horizontal rotation angle of the gimbal is used as the detector's horizontal angle corresponding to that frame of dust image, and the pitch rotation angle of the gimbal is used as the detector's pitch angle corresponding to that frame of dust image, and is stored in association with that frame of dust image.

[0043] Furthermore, when the image-type dust detector is in cruise mode, the detector's horizontal angle and pitch angle change with the cruise preset position; when the image-type dust detector stops cruise due to detecting a dust area, the detector's horizontal angle and pitch angle corresponding to the moment cruise stops are used as the detector's horizontal angle and pitch angle corresponding to the last dust image in the S1 dust image sequence.

[0044] S2.2: Based on the pixel coordinates of the current dust area center position in the dust image and the preset conversion relationship between pixel coordinates and angle offset, calculate the horizontal angle offset and pitch angle offset corresponding to the current dust area center position; add the horizontal angle offset to the detector horizontal angle to obtain the detector horizontal angle; add the pitch angle offset to the detector pitch angle to obtain the detector pitch angle.

[0045] Furthermore, the conversion relationship between pixel coordinates and angle offset is as follows: multiply the difference between the horizontal coordinate of the current dust area center position and the horizontal center pixel coordinate of the dust image by the horizontal field of view angle corresponding to a unit pixel of the image-type dust detector to obtain the horizontal angle offset; multiply the difference between the vertical coordinate of the current dust area center position and the vertical center pixel coordinate of the dust image by the vertical field of view angle corresponding to a unit pixel of the image-type dust detector to obtain the pitch angle offset.

[0046] It should be noted that the detection horizontal angle and detection pitch angle represent the actual spatial azimuth angle of the current dust area center position relative to the installation position of the image-type dust detector. Both the detection horizontal angle and detection pitch angle change with the changes in the detector's horizontal angle and detector pitch angle, which is different from the conversion method that determines the dust location based solely on pixel coordinates without considering the detector's real-time rotation angle.

[0047] S2.3: Obtain the pre-calibrated angle mapping relationship between the image-type dust detector and the fog cannon, wherein the angle mapping relationship is used to map the detection horizontal angle and the detection pitch angle to the initial target horizontal angle and the initial target pitch angle.

[0048] Specifically, the angle mapping relationship is obtained through on-site calibration: multiple calibration points are selected within the field of view of the image-type dust detector, covering different distances and heights. The detection horizontal angle and detection pitch angle corresponding to each calibration point are recorded. Simultaneously, the fog cannon is driven to point at each calibration point, and the corresponding actual horizontal angle and actual pitch angle of the fog cannon are recorded. The detection horizontal angle and detection pitch angle corresponding to each calibration point are used as calibration data pairs with the actual horizontal angle and actual pitch angle of the fog cannon. A horizontal mapping function between the detection horizontal angle and the actual horizontal angle of the fog cannon, and a pitch mapping function between the detection pitch angle and the actual pitch angle of the fog cannon are established through polynomial fitting. The detection horizontal angle is substituted into the horizontal mapping function to obtain the initial target horizontal angle, and the detection pitch angle is substituted into the pitch mapping function to obtain the initial target pitch angle.

[0049] It should be noted that the on-site calibration method is suitable for on-site conditions where the actual height distribution of dust targets is relatively fixed. In this case, there is a relatively stable correspondence between the detection pitch angle and the dust estimation distance, so that the coupling error introduced by the independent fitting of the horizontal mapping function and the pitch mapping function is within the preset angle deviation threshold. For on-site conditions where the coupling error exceeds the preset angle deviation threshold due to the difference in the height distribution of dust targets, the following mapping relationship establishment method based on spatial coordinate calculation is adopted.

[0050] The preset angle deviation threshold and coupling error are determined as follows: Multiple measured dust points at different heights are selected within the field of view of the image-type dust detector, and the detection pitch angle and the actual pitch angle of the fog cannon corresponding to each measured dust point are recorded. These measured dust points are divided into a calibration subset and a verification subset. After fitting the horizontal mapping function and pitch mapping function using the calibration subset, the deviation between the actual pitch angle of the fog cannon corresponding to each measured dust point in the verification subset and the pitch angle calculated using the pitch mapping function is calculated. The maximum value among the deviation values ​​is taken as the coupling error. The preset angle deviation threshold is pre-set based on the spray coverage angle range of the fog cannon, taking the maximum allowable pointing deviation within the spray coverage angle range. When the coupling error exceeds the preset angle deviation threshold, it is determined that the current working condition does not meet the applicable conditions of the on-site calibration method, and the method of establishing the mapping relationship based on spatial coordinate calculation is switched to.

[0051] Furthermore, in situations where on-site calibration is not feasible, the angle mapping relationship is established as follows: Pre-calibrated installation baseline parameters between the image-type dust detector and the mist cannon are obtained, including the horizontal distance and height difference between them; the spatial coordinates of the current dust area center relative to the image-type dust detector installation position are calculated based on the detection horizontal angle, detection pitch angle, and a preset dust estimation distance; the horizontal pointing angle and pitch pointing angle of the current dust area center relative to the mist cannon installation position are calculated based on the spatial coordinates and installation baseline parameters, serving as the initial target horizontal angle and initial target pitch angle, respectively. The preset dust estimation distance is determined based on the image-type dust detector installation height and the preset average height of the dust area, serving as an alternative to on-site calibration when it is not feasible; the preset average height of the dust area is pre-measured and stored in the system parameter table based on the external dimensions of the equipment at the dust source location (e.g., the height of a belt conveyor or the height of the material discharge port from the ground); this preset value is updated through on-site manual calibration or laser rangefinder-assisted measurement when switching between different operating conditions.

[0052] S2.4: Obtain the current angle of the fog cannon, calculate the horizontal angle difference between the initial target horizontal angle and the current horizontal angle of the fog cannon, and the pitch angle difference between the initial target pitch angle and the current pitch angle of the fog cannon. The current angle of the fog cannon includes the current horizontal angle and the current pitch angle of the fog cannon, which are detected and fed back in real time by the angle sensor.

[0053] It should be noted that the fog cannon's rotation is driven by a motor, and the rotation speed is not constant. The rotation process includes an acceleration phase, a constant speed phase, and a deceleration phase. Furthermore, the fog cannon rotates in a sequential manner, first horizontally and then pitching after reaching the horizontal position. If the fog cannon's response delay is calculated directly using a fixed rotation speed value and the difference between the horizontal and pitch angles, there will be a deviation from the actual rotation time of the fog cannon. Therefore, the fog cannon's response delay is determined through steps S2.5 and S2.6.

[0054] S2.5: Based on the preset fog cannon rotation speed curve, obtain the horizontal rotation time and the pitch rotation time; wherein the fog cannon rotation speed curve includes the angular velocity change relationship during the acceleration phase, the angular velocity value during the uniform speed phase, the angular velocity change relationship during the deceleration phase, as well as the first angle threshold and the second angle threshold; the first angle threshold is the angle value through which the fog cannon rotates from a stationary state to the uniform speed phase angular velocity value; the second angle threshold is the angle value through which the fog cannon rotates from the uniform speed phase angular velocity value to the stationary state.

[0055] Specifically, the horizontal angle difference is compared with the sum of the first angle threshold and the second angle threshold. When the horizontal angle difference is greater than the sum of the first angle threshold and the second angle threshold, it is determined that the corresponding rotation process has experienced a complete acceleration phase, a constant speed phase, and a deceleration phase. Based on the difference between the horizontal angle difference and the sum of the first angle threshold and the second angle threshold, as well as the angular velocity value of the constant speed phase, the rotation duration of the constant speed phase is calculated. The rotation duration of the constant speed phase is added to the acceleration rotation duration corresponding to the first angle threshold and the deceleration rotation duration corresponding to the second angle threshold to obtain the horizontal rotation duration. The acceleration rotation duration and the deceleration rotation duration are obtained by integrating the angular velocity change relationship of the acceleration phase and the angular velocity change relationship of the deceleration phase within the angle intervals corresponding to the first angle threshold and the second angle threshold, respectively.

[0056] When the horizontal angle difference is less than the sum of the first and second angle thresholds, the corresponding rotation process is determined to only experience acceleration and deceleration phases, without a uniform speed phase. In this case, the actual peak angular velocity during the fog cannon's rotation is less than the angular velocity value during the uniform speed phase. The horizontal rotation duration is calculated as follows: Based on the angular velocity change relationship during the acceleration phase, a first functional relationship between the fog cannon's rotation angle and acceleration duration is determined; based on the angular velocity change relationship during the deceleration phase, a second functional relationship between the fog cannon's rotation angle and deceleration duration is determined; using the actual peak angular velocity as the variable to be determined, the acceleration angle through which the fog cannon accelerates to the actual peak angular velocity is calculated according to the first functional relationship; the deceleration angle through which the fog cannon decelerates from the actual peak angular velocity to a stationary state is calculated according to the second functional relationship; an equation is established where the sum of the acceleration angle and the deceleration angle equals the horizontal angle difference, and the actual peak angular velocity is obtained by solving it; the actual peak angular velocity is substituted into the first functional relationship to obtain the acceleration duration; the actual peak angular velocity is substituted into the second functional relationship to obtain the deceleration duration; and the acceleration duration and deceleration duration are added together to obtain the horizontal rotation duration.

[0057] The specific methods for determining the first functional relationship and the second functional relationship include: Specifically, the first functional relationship is obtained by integrating the angular velocity change relationship during the acceleration phase over the acceleration duration interval; the second functional relationship is obtained by integrating the angular velocity change relationship during the deceleration phase over the deceleration duration interval; the angular velocity change relationship during the acceleration phase represents the change of the fog cannon's angular velocity with the acceleration duration during the acceleration phase, and the angular velocity change relationship during the deceleration phase represents the change of the fog cannon's angular velocity with the deceleration duration during the deceleration phase. Preferably, the angular velocity change during the acceleration phase is linear, meaning the angular acceleration of the fog cannon during the acceleration phase is a constant value, denoted as the first angular acceleration. Similarly, the angular velocity change during the deceleration phase is also linear, meaning the angular acceleration of the fog cannon during the deceleration phase is a constant value, denoted as the second angular acceleration. In this case, the first functional relationship is: the angle through which the fog cannon accelerates is equal to half the product of the first angular acceleration and the square of the acceleration duration; the second functional relationship is: the angle through which the fog cannon decelerates is equal to half the product of the second angular acceleration and the square of the deceleration duration. Furthermore, based on the first functional relationship, the acceleration time required for the fog cannon to accelerate to the actual peak angular velocity is equal to the ratio of the actual peak angular velocity to the first angular acceleration, and the corresponding acceleration angle is equal to the ratio of the square of the actual peak angular velocity to twice the product of the first angular acceleration; according to the second functional relationship, the deceleration time required for the fog cannon to decelerate from the actual peak angular velocity to a stationary state is equal to the ratio of the actual peak angular velocity to the second angular acceleration, and the corresponding deceleration angle is equal to the ratio of the square of the actual peak angular velocity to twice the product of the second angular acceleration; Expanding the equation that the sum of the acceleration angle and the deceleration angle equals the difference in horizontal angle, we obtain that the product of the square of the actual peak angular velocity and half the sum of the reciprocals of the first and second angular accelerations equals the difference in horizontal angle. Based on this, we can calculate the actual peak angular velocity. Substituting the actual peak angular velocity into the calculation relationships for acceleration and deceleration durations, we obtain the acceleration and deceleration durations. Adding these two together gives us the duration of horizontal rotation. In an optional embodiment, when the angular velocity change relationships during the acceleration and deceleration phases are nonlinear, the first and second functional relationships are obtained by numerically integrating the angular velocity change relationships during the acceleration and deceleration phases, respectively. The actual peak angular velocity is determined by a numerical iterative solution method, ensuring that the sum of the acceleration and deceleration angles equals the horizontal angle difference. The numerical iterative solution method employs a bisection method: the search interval for the actual peak angular velocity is set to (0, V0], where V0 is the angular velocity value during the uniform velocity phase; the midpoint of the search interval is taken as the current candidate peak angular velocity, and the sum of the corresponding acceleration and deceleration angles is calculated; if the sum of the acceleration and deceleration angles is less than the horizontal angle difference, the lower limit of the search interval is updated to the current candidate peak angular velocity; if the sum of the acceleration and deceleration angles is greater than the horizontal angle difference, the upper limit of the search interval is updated to the current candidate peak angular velocity; this iteration is repeated until the width of the search interval is less than a preset iteration accuracy threshold, and the midpoint of the final search interval is determined as the actual peak angular velocity.

[0058] Furthermore, following the comparison and calculation methods of this step, the pitch rotation duration is calculated based on the comparison result between the pitch angle difference and the sum of the first and second angle thresholds. When the pitch angle difference is greater than the sum of the first and second angle thresholds, the pitch rotation duration is determined according to the calculation method corresponding to the complete trapezoidal velocity curve. When the pitch angle difference is less than or equal to the sum of the first and second angle thresholds, the pitch rotation duration is determined according to the actual peak angular velocity solution method.

[0059] It should be noted that the first and second angle thresholds are obtained based on the historical rotation records of the fog cannon and can be updated periodically.

[0060] S2.6: Add the horizontal rotation time to the pitch rotation time (since the fog cannon first performs a horizontal rotation and then a pitch rotation after the horizontal rotation is in place) to obtain the fog cannon response delay.

[0061] Furthermore, the response delay of the fog cannon varies with the values ​​of the horizontal angle difference and the pitch angle difference, replacing the calculation method of estimating the response delay with a fixed rotation speed value, thus providing a time reference for S3 dust position prediction.

[0062] It should be noted that the rotation speed curve of the fog cannon is individually calibrated for each fog cannon. Different fog cannons correspond to different angular velocity change relationships in the acceleration phase, angular velocity values ​​in the constant speed phase, angular velocity change relationships in the deceleration phase, first angle threshold, and second angle threshold, which are used to correct for the differences in rotational performance caused by mechanical tolerances between fog cannons of the same model.

[0063] Preferably, the rotation speed curve of the fog cannon is continuously updated with the historical rotation records of the fog cannon. The difference between the actual rotation time in the historical rotation records and the rotation time calculated according to the current rotation speed curve of the fog cannon is used to adjust the corresponding parameters in the angular velocity change relationship in the acceleration phase, the angular velocity value in the uniform speed phase, and the angular velocity change relationship in the deceleration phase, so that the rotation speed curve of the fog cannon changes with the mechanical wear generated by the long-term operation of the fog cannon.

[0064] In an optional embodiment, when the horizontal angle difference is zero, it is determined that the current horizontal angle of the fog cannon is consistent with the initial target horizontal angle, and the horizontal rotation duration is zero. At this time, the response delay of the fog cannon is determined only by the pitch rotation duration. The same processing method is used when the pitch rotation duration is zero.

[0065] S3: Based on the direction of dust movement, the speed of dust movement, and the response delay of the fog cannon, predict the predicted dust position corresponding to the dust area. Based on the position offset of the predicted dust position relative to the current center position of the dust area and the correspondence between the preset position offset and the angle correction, correct the initial target horizontal angle and the initial target pitch angle to obtain the target horizontal angle and the target pitch angle.

[0066] S3.1: Based on the dust movement direction, dust movement speed and fog cannon response delay, calculate the predicted displacement of the dust area within the fog cannon response delay; determine the predicted dust position corresponding to the dust area based on the current center position of the dust area and the predicted displacement.

[0067] Specifically, the predicted displacement includes the horizontal displacement of the dust image and the vertical displacement of the dust image. The horizontal displacement is obtained by multiplying the velocity component of the dust movement in the horizontal direction of the dust image by the fog cannon's response delay; the vertical displacement is obtained by multiplying the velocity component of the dust movement in the vertical direction of the dust image by the fog cannon's response delay. The horizontal direction of the dust image refers to the row direction of the pixels; the vertical direction refers to the column direction of the pixels.

[0068] Furthermore, when the dust movement direction only includes the horizontal motion component of the dust image, the predicted vertical displacement of the dust image is zero; when the dust movement direction only includes the vertical motion component of the dust image, the predicted horizontal displacement of the dust image is zero.

[0069] Furthermore, the method for determining the predicted dust location is as follows: taking the pixel coordinates of the current dust region center position in the dust image as a reference, the predicted horizontal displacement and the predicted vertical displacement of the dust image are superimposed on the pixel row coordinates and pixel column coordinates of the current dust region center position, respectively, to obtain the predicted pixel coordinates of the predicted dust location in the dust image.

[0070] It should be noted that the fog cannon response delay is S2.6, the dust movement direction and velocity are S1.7, and the current dust area center position is S1.5. The predicted dust position represents the projected position of the dust area in the dust image after the fog cannon response delay.

[0071] S3.2: Based on the pixel differences between the predicted dust location and the current dust area center location in the horizontal and vertical directions of the dust image, determine the pixel offset in the horizontal and vertical directions of the dust image, respectively.

[0072] Specifically, the difference between the predicted pixel row coordinates of the predicted dust location and the center pixel row coordinates of the current dust region center is calculated, and the difference is used as the horizontal pixel offset of the dust image; the difference between the predicted pixel column coordinates of the predicted dust location and the center pixel column coordinates of the current dust region center is calculated, and the difference is used as the vertical pixel offset of the dust image.

[0073] Furthermore, when the predicted pixel row coordinates are greater than the center pixel row coordinates, the horizontal pixel offset of the dust image is positive, indicating that the predicted dust position has drifted in the positive horizontal direction relative to the center of the current dust region. When the predicted pixel row coordinates are less than the center pixel row coordinates, the horizontal pixel offset of the dust image is negative, indicating that the predicted dust position has drifted in the negative horizontal direction relative to the center of the current dust region. The method for determining the positive or negative value of the vertical pixel offset of the dust image is the same as that for the horizontal pixel offset.

[0074] S3.3: Substitute the horizontal pixel offset of the dust image into the first correspondence between the preset horizontal offset and the horizontal angle correction to calculate the horizontal angle correction; substitute the vertical pixel offset of the dust image into the second correspondence between the preset vertical offset and the pitch angle correction to calculate the pitch angle correction.

[0075] Furthermore, the first correspondence is that the horizontal angle correction is equal to the product of the horizontal pixel offset of the dust image and the horizontal field of view angle corresponding to a unit pixel of the image-type dust detector. The second correspondence is that the pitch angle correction is equal to the product of the vertical pixel offset of the dust image and the vertical field of view angle corresponding to a unit pixel of the image-type dust detector.

[0076] Specifically, the horizontal field of view angle corresponding to a unit pixel of the image-type dust detector is determined by dividing the horizontal field of view angle of the image-type dust detector by the total number of pixels in the horizontal direction of the dust image; the vertical field of view angle corresponding to a unit pixel of the image-type dust detector is determined by dividing the vertical field of view angle of the image-type dust detector by the total number of pixels in the vertical direction of the dust image.

[0077] Furthermore, if the horizontal pixel offset of the dust image is positive, the horizontal angle correction is a positive correction, used to increase the initial target horizontal angle; if the horizontal pixel offset of the dust image is negative, the horizontal angle correction is a negative correction, used to decrease the initial target horizontal angle. The method for determining the sign of the pitch angle correction is the same as that for the horizontal angle correction.

[0078] It should be noted that the first and second correspondences employ a linear conversion method that multiplies the field of view angle per unit pixel by the pixel offset. This conversion method is based on the far-field imaging assumption that the distance between the dust region and the image-type dust detector is much greater than the focal length of the image-type dust detector. When the spatial distance between the dust region and the image-type dust detector satisfies the far-field condition, the spatial angle increment corresponding to the pixel offset between the predicted dust position and the current center position of the dust region is approximately equal to the product of the pixel offset and the field of view angle per unit pixel. If the dust region does not satisfy the above far-field condition, the field of view angle conversion result is nonlinearly corrected based on the ratio of the actual object distance to the focal length. The correction coefficient is obtained through interpolation using a pre-generated lookup table. The far-field condition is that the spatial distance between the dust region and the image-type dust detector is greater than ten times the focal length of the image-type dust detector.

[0079] S3.4: Add the horizontal angle correction to the initial target horizontal angle to obtain the target horizontal angle; add the pitch angle correction to the initial target pitch angle to obtain the target pitch angle.

[0080] It should be noted that the target horizontal angle and target pitch angle are the target angles that the fog cannon needs to point at after correction of the predicted dust position. These target horizontal and pitch angles are used to replace the initial target horizontal and pitch angles as the target angles for controlling the fog cannon's rotation in S4. The initial target horizontal and pitch angles are the same as those in S2.3, used to compensate for the pointing deviation caused by the continued movement of the dust area during the fog cannon's response delay.

[0081] S3.5: Obtain the mechanical limit parameters of the fog cannon, and compare the target horizontal angle with the minimum and maximum horizontal rotation angles of the fog cannon. If the target horizontal angle is greater than the maximum horizontal rotation angle, limit the target horizontal angle to the maximum horizontal rotation angle; if the target horizontal angle is less than the minimum horizontal rotation angle, limit the target horizontal angle to the minimum horizontal rotation angle. Compare the target pitch angle with the minimum and maximum pitch rotation angles of the fog cannon. If the target pitch angle is greater than the maximum pitch rotation angle, limit the target pitch angle to the maximum pitch rotation angle; if the target pitch angle is less than the minimum pitch rotation angle, limit the target pitch angle to the minimum pitch rotation angle.

[0082] It should be noted that the mechanical limit parameters include the minimum horizontal rotation angle of the fog cannon, the maximum horizontal rotation angle of the fog cannon, the minimum pitch rotation angle of the fog cannon, and the maximum pitch rotation angle of the fog cannon.

[0083] Preferably, the target horizontal angle and target pitch angle obtained after being defined are the final angle values ​​used to control the rotation of the fog cannon.

[0084] Furthermore, if the target horizontal angle is limited to the maximum or minimum horizontal rotation angle of the fog cannon, a horizontal angle limiting prompt signal is generated; if the target pitch angle is limited to the maximum or minimum pitch rotation angle of the fog cannon, a pitch angle limiting prompt signal is generated; the horizontal and pitch angle limiting prompt signals are used to indicate that the current angle correction exceeds the mechanical execution range of the fog cannon.

[0085] S3.6: Record the predicted time of the dust location and use the predicted time as the reference for the start time of S4 controlling the rotation of the fog cannon.

[0086] It should be noted that the predicted time is the sum of the acquisition time corresponding to the current center position of the dust area and the response delay of the fog cannon. The predicted time is used to determine the starting time of the fog cannon rotation in S4, so that the rotation action of the fog cannon is aligned with the expected drift position of the dust area in time.

[0087] In an optional embodiment, when the dust movement speed is zero, the predicted dust position coincides with the current center position of the dust area, the horizontal pixel offset of the dust image and the vertical pixel offset of the dust image are both zero, the horizontal angle correction and the pitch angle correction are both zero, the target horizontal angle is equal to the initial target horizontal angle, and the target pitch angle is equal to the initial target pitch angle.

[0088] In an optional embodiment, when the fog cannon response delay is less than a preset delay threshold, the calculation steps of angle correction in steps S3.3 to S3.4 in steps S3.1 to S3.6 are skipped, and the initial target horizontal angle is directly used as the target horizontal angle, and the initial target pitch angle is used as the target pitch angle.

[0089] It should be noted that the preset delay threshold is determined based on the acquisition frame rate of the dust image sequence. When the response delay of the fog cannon is less than the acquisition time interval between two adjacent dust images, the displacement of the dust area within the response delay of the fog cannon is less than the spatial distance corresponding to one pixel. The horizontal angle correction and the pitch angle correction are both less than the angular resolution of the image-type dust detector. The effect of the correction operation on the pointing accuracy of the fog cannon is negligible.

[0090] S4: Control the fog cannon to rotate towards the target horizontal angle and target pitch angle, and obtain the real-time horizontal angle and real-time pitch angle of the fog cannon; when the real-time horizontal angle and real-time pitch angle respectively meet the preset angle deviation conditions of the corresponding target angle, control the fog cannon to spray.

[0091] S4.1: Obtain the target horizontal angle, target pitch angle, current horizontal angle of the fog cannon, current pitch angle of the fog cannon, and prediction time; determine the horizontal rotation angle based on the difference between the target horizontal angle and the current horizontal angle of the fog cannon; determine the pitch rotation angle based on the difference between the target pitch angle and the current pitch angle of the fog cannon.

[0092] Specifically, the target horizontal angle and target pitch angle are the final angle values ​​used to control the rotation of the fog cannon, obtained after amplitude limiting processing in S3.5; the current horizontal angle and current pitch angle of the fog cannon are in S2.4; and the prediction time is in S3.6. The horizontal rotation angle to be obtained is the difference between the target horizontal angle and the current horizontal angle of the fog cannon, and the pitch rotation angle to be obtained is the difference between the target pitch angle and the current pitch angle of the fog cannon.

[0093] Furthermore, when the horizontal rotation angle to be rotated is zero and the pitch rotation angle to be rotated is zero, it is determined that the fog cannon is currently aligned with the target horizontal angle and the target pitch angle, skipping the rotation control steps from S4.2 to S4.5, and directly executing the spray control step from S4.6.

[0094] S4.2: Based on the horizontal and vertical rotation angles to be rotated, generate horizontal rotation and vertical rotation commands for the fog cannon; send the horizontal and vertical rotation commands to the fog cannon's drive mechanism to control the fog cannon to start rotating.

[0095] Furthermore, the fog cannon's horizontal rotation command includes the target horizontal angle as the target value for horizontal rotation; the fog cannon's pitch rotation command includes the target pitch angle as the target value for pitch rotation. The drive mechanism includes a horizontal drive motor and a pitch drive motor. The horizontal drive motor is used to drive the fog cannon to rotate in the horizontal direction, and the pitch drive motor is used to drive the fog cannon to rotate in the pitch direction.

[0096] Specifically, the fog cannon horizontal rotation command and fog cannon pitch rotation command are sent sequentially: first, the fog cannon horizontal rotation command is sent to control the horizontal drive motor to drive the fog cannon to rotate in the horizontal direction; after the fog cannon's horizontal rotation is completed, the fog cannon pitch rotation command is sent to control the pitch drive motor to drive the fog cannon to rotate in the pitch direction.

[0097] In an optional embodiment, when the horizontal rotation and pitch rotation actions of the fog cannon are configured to be executed synchronously, the fog cannon horizontal rotation command and the fog cannon pitch rotation command are sent simultaneously, and the fog cannon response delay is the maximum value of the horizontal rotation duration and the pitch rotation duration.

[0098] S4.3: Record the sending time of the fog cannon horizontal rotation command and the fog cannon pitch rotation command, and take the sending time as the rotation start time; determine the waiting time threshold based on the rotation start time and the prediction time.

[0099] It should be noted that the waiting time threshold is the difference between the predicted time and the start time of rotation. The predicted time is the predicted time in S3.6, which represents the time when the dust area is expected to reach the predicted dust position; the waiting time threshold is used to limit the maximum allowable time for the fog cannon to rotate into place.

[0100] S4.4: During the rotation of the fog cannon, the real-time horizontal angle and real-time pitch angle of the fog cannon are repeatedly acquired according to the preset sampling period.

[0101] Specifically, the preset sampling period is less than the response time constant of the fog cannon, which is the time required for the fog cannon to rotate from a stationary state to reach a steady-state angular velocity. The real-time horizontal angle of the fog cannon is detected by a first angle sensor installed on the horizontal rotation mechanism of the fog cannon; the real-time pitch angle of the fog cannon is detected by a second angle sensor installed on the pitch mechanism of the fog cannon.

[0102] Furthermore, the first angle sensor and the second angle sensor are magnetic encoders or photoelectric encoders, and the angle resolution of the magnetic encoder or photoelectric encoder is less than the angle deviation threshold in the preset angle deviation condition.

[0103] S4.5: For each acquired real-time horizontal angle of the fog cannon, calculate the horizontal angle deviation between the real-time horizontal angle of the fog cannon and the target horizontal angle; for each acquired real-time pitch angle of the fog cannon, calculate the pitch angle deviation between the real-time pitch angle of the fog cannon and the target pitch angle; when the absolute value of the horizontal angle deviation is less than or equal to the preset horizontal angle deviation threshold, and the absolute value of the pitch angle deviation is less than or equal to the preset pitch angle deviation threshold, it is determined that the fog cannon has rotated to the target position.

[0104] Furthermore, the preset horizontal angle deviation threshold and the preset pitch angle deviation threshold are determined based on the spray coverage angle range of the fog cannon. The spray coverage angle range is the cone angle of the water mist coverage area formed by the water mist sprayed by the fog cannon at the target distance. When the deviation between the actual pointing angle of the fog cannon and the target angle is less than half of the spray coverage angle range, the dust area is still within the water mist coverage area, and the dust suppression effect is not significantly reduced due to the pointing deviation.

[0105] Specifically, the method for continuously determining whether the horizontal angle deviation and pitch angle deviation meet the preset angle deviation conditions is as follows: within N consecutive sampling periods, the absolute value of the horizontal angle deviation acquired each time is less than or equal to the preset horizontal angle deviation threshold, and the absolute value of the pitch angle deviation acquired each time is less than or equal to the preset pitch angle deviation threshold, where N is a preset positive integer. The preset positive integer N is determined based on the rotational stability of the fog cannon and is used to filter out transient fluctuations caused by noise from the angle sensor measurement.

[0106] Furthermore, after determining that the fog cannon has rotated to the target position, the current time is recorded as the rotation arrival time, and the difference between the rotation arrival time and the rotation start time is calculated as the actual rotation duration. The actual rotation duration is compared with the waiting time threshold, and the following control logic is executed according to the comparison result: a. If the actual rotation time is less than the waiting time threshold, it is determined that the fog cannon has rotated to the position ahead of schedule; at this time, the spray control is not executed immediately, but the system enters a waiting state and continuously monitors the current time until the predicted time is reached; in the waiting state, the current horizontal angle and pitch angle of the fog cannon are kept unchanged. b. If the actual rotation time is equal to the waiting time threshold, it is determined that the time when the fog cannon rotates to the correct position is consistent with the predicted time, and spray control can be executed directly; c. If the actual rotation time exceeds the waiting time threshold, a rotation timeout warning signal is generated, and it is determined that the fog cannon cannot complete the pointing action before reaching the predicted dust position in the dust area, and the prediction compensation fails. At this time, no fog cannon spray start command is generated, and the host computer or operator decides whether to execute the subsequent spray operation or wait for the next dust image acquisition cycle to re-locate and predict the target.

[0107] It should be noted that the waiting state is used to compensate for the positive deviation between the actual rotation time of the fog cannon and the response delay of the fog cannon (i.e., the actual rotation is faster than the calibration curve), ensuring that the fog cannon's spraying action starts at the predicted time, so that the water mist coverage time is aligned with the time when the dust area reaches the predicted dust position; the rotation timeout warning signal is used to indicate that the fog cannon's rotation time exceeds the time when the dust area is expected to reach the predicted dust position, the dust area may have deviated from the predicted dust position, and in this case, spraying will not start automatically to avoid ineffective spraying that wastes water resources and disturbs the site environment.

[0108] S4.6: Generate a fog cannon spray start command only when the first condition (a) in S4.5 is met and the current time has reached the predicted time, or when the second condition (b) in S4.5 is met; send the fog cannon spray start command to the fog cannon's water supply system and fan system to control the fog cannon to start spraying water mist.

[0109] Preferably, the fog cannon spray start command includes a preset spray duration; after receiving the fog cannon spray start command, the water supply system opens the water supply valve, and after receiving the fog cannon spray start command, the fan system starts the fan motor, so that the fog cannon sprays water mist in the horizontal and vertical directions of the target.

[0110] It should be noted that the preset spray duration is determined based on the drift speed of the dust area and the spray coverage of the fog cannon, so that the fog cannon continuously sprays water mist within the spray duration to cover the dust clumps passing through the predicted dust location.

[0111] S4.7: After the fog cannon starts spraying water mist, continue to acquire the real-time horizontal angle and real-time pitch angle of the fog cannon according to the sampling cycle; for each acquired real-time horizontal angle and real-time pitch angle of the fog cannon, repeat the angle deviation calculation and judgment steps in S4.5; when it is determined that the fog cannon has deviated from the target position during the fog cannon spraying process, generate an angle deviation prompt signal.

[0112] It should be noted that the angle deviation warning signal is used to indicate that the pointing angle of the fog cannon exceeds the preset angle deviation condition due to mechanical vibration or external interference during the spraying process, and the dust area may have exceeded the water mist coverage area.

[0113] In an optional embodiment, after generating an angle deviation warning signal, a fog cannon spray stop command is first generated to control the fog cannon to stop spraying water mist, and then the rotation control steps S4.2 to S4.5 are executed to realign the fog cannon with the target horizontal angle and target pitch angle, and after it is back in position, the spray control step S4.6 is executed.

[0114] S4.8: Based on the preset spray duration, a timer is started after the fog cannon begins spraying water mist. When the spray duration reaches the preset spray duration, a fog cannon spray stop command is generated. The fog cannon spray stop command is sent to the fog cannon's water supply system and fan system to control the fog cannon to stop spraying water mist.

[0115] Specifically, the fog cannon spray stop command includes a command to close the water supply valve and a command to shut down the fan motor; after receiving the fog cannon spray stop command, the water supply system closes the water supply valve, and after receiving the fog cannon spray stop command, the fan system shuts down the fan motor, so that the fog cannon stops spraying water mist.

[0116] In an optional embodiment, when the fog cannon returns to the execution of the rotation control steps S4.2 to S4.5 due to the angle deviation warning signal during the spraying process, it continues to execute the spray control step S4.6 after returning to the correct position, and the preset spray duration is restarted.

[0117] In summary, this invention utilizes inter-frame correlation recognition in image sequences to not only capture the current center position of the dust region but also extract its direction and speed of motion. This upgrades the control of the fog cannon from static fixed-point spraying to dynamic tracking spraying with predictive capabilities, fundamentally overcoming the target drift problem caused by dust migration. Furthermore, by combining the detector installation angle with a pre-calibrated spatial mapping relationship, the image coordinates are converted into the initial target angle of the fog cannon. The fog cannon's response delay is then dynamically calculated based on the difference between this target angle and the current angle, quantifying the inertial lag of the actuator into a precise time reference and avoiding systemic deviations caused by fixed empirical values. Subsequently, the dust motion vector and the delay are spatiotemporally fused to predict the dust's movement during the fog cannon's rotation. The actual drift position is determined, and a linear conversion relationship is established between the pixel offset in the image domain and the angle correction. The initial target angle is accurately compensated, eliminating the system-level aiming error caused by time delay from the mechanism level, so that the actual landing point of the fog cannon and the actual position of the dust are matched in the same domain. Finally, real-time angle feedback and preset deviation condition judgment are introduced during the rotation of the fog cannon to ensure that the spray command is triggered only when the attitude converges to the neighborhood of the target angle. This not only ensures the spray hit rate, but also takes into account the timeliness of response and the system's anti-interference ability through the tolerance mechanism. This invention significantly improves the pointing accuracy and dust suppression efficiency of the fog cannon in the industrial dust scene with large lag and high dynamics, and realizes the synergistic optimization of control predictability, execution accuracy and system robustness.

[0118] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A control method for directing a fog cannon based on dust image analysis, characterized in that, include: The dust image sequence continuously acquired by the image-type dust detector is obtained, the dust region in the dust image sequence is identified and frame-to-frame correlation is performed, and the current center position of the dust region and the corresponding dust movement direction and dust movement speed are determined. The horizontal and vertical angles of the image-type dust detector are obtained. Based on the current center position of the dust area, the horizontal and vertical angles of the detector, and the pre-calibrated angle mapping relationship between the image-type dust detector and the fog cannon, the initial target horizontal and vertical angles are determined. The fog cannon response delay is obtained based on the angle difference between the initial target horizontal angle, the initial target vertical angle, and the current angle of the fog cannon. Based on the dust movement direction, the dust movement speed, and the fog cannon response delay, the predicted dust position corresponding to the dust area is predicted. Based on the position offset of the predicted dust position relative to the current dust area center position and the correspondence between the preset position offset and the angle correction amount, the initial target horizontal angle and the initial target pitch angle are corrected to obtain the target horizontal angle and the target pitch angle. Control the fog cannon to rotate towards the target at the horizontal angle and the target at the pitch angle, and obtain the real-time horizontal angle and real-time pitch angle of the fog cannon; When the real-time horizontal angle and the real-time pitch angle respectively meet the preset angle deviation conditions of the corresponding target angle, the fog cannon is controlled to spray.

2. The control method for pointing a fog cannon based on dust image analysis as described in claim 1, characterized in that, Controlling the fog cannon to rotate towards the target at both the horizontal and vertical angles includes: The horizontal rotation angle is determined based on the difference between the target's horizontal angle and the fog cannon's current horizontal angle; the pitch rotation angle is determined based on the difference between the target's pitch angle and the fog cannon's current pitch angle. Based on the horizontal rotation angle and the pitch rotation angle, a horizontal rotation command and a pitch rotation command for the fog cannon are generated; the horizontal rotation command and the pitch rotation command are sent to the driving mechanism of the fog cannon to control the fog cannon to start rotating.

3. The control method for pointing a fog cannon based on dust image analysis as described in claim 2, characterized in that, The target horizontal angle and target pitch angle include: Based on the pixel differences between the predicted dust location and the current dust region center location in the horizontal and vertical directions of the dust image, the pixel offset in the horizontal direction and the pixel offset in the vertical direction of the dust image are determined respectively. The horizontal pixel offset of the dust image is substituted into a first correspondence between a preset horizontal offset and a horizontal angle correction to calculate the horizontal angle correction; the vertical pixel offset of the dust image is substituted into a second correspondence between a preset vertical offset and a pitch angle correction to calculate the pitch angle correction. The horizontal angle correction is added to the initial target horizontal angle to obtain the target horizontal angle; the pitch angle correction is added to the initial target pitch angle to obtain the target pitch angle.

4. The control method for pointing a fog cannon based on dust image analysis as described in claim 3, characterized in that, The first and second correspondences adopt a linear conversion method that multiplies the field of view angle corresponding to a unit pixel by the pixel offset; the first correspondence is that the horizontal angle correction is equal to the product of the horizontal pixel offset of the dust image and the horizontal field of view angle corresponding to a unit pixel of the image-type dust detector; the second correspondence is that the pitch angle correction is equal to the product of the vertical pixel offset of the dust image and the vertical field of view angle corresponding to a unit pixel of the image-type dust detector.

5. The control method for pointing a fog cannon based on dust image analysis as described in claim 3, characterized in that, The prediction of dust location includes: Based on the dust movement direction, dust movement speed, and the fog cannon response delay, the predicted displacement of the dust area within the fog cannon response delay is calculated; the predicted displacement includes the predicted displacement of the dust image in the horizontal direction and the predicted displacement of the dust image in the vertical direction. The predicted dust location corresponding to the dust area is determined based on the current center position of the dust area and the predicted displacement.

6. The control method for pointing a fog cannon based on dust image analysis as described in claim 5, characterized in that, The fog cannon response delay includes: Based on a preset fog cannon rotation speed curve, the horizontal rotation time and the pitch rotation time are obtained; wherein the fog cannon rotation speed curve includes the angular velocity change relationship during the acceleration phase, the angular velocity value during the uniform speed phase, the angular velocity change relationship during the deceleration phase, and a first angle threshold and a second angle threshold; the first angle threshold is the angle value through which the fog cannon rotates from a stationary state to the uniform speed phase angular velocity value; the second angle threshold is the angle value through which the fog cannon decelerates from the uniform speed phase angular velocity value to a stationary state; The response delay of the fog cannon is obtained by adding the horizontal rotation time to the pitch rotation time.

7. The control method for pointing a fog cannon based on dust image analysis as described in claim 5, characterized in that, The determination of the dust movement speed and direction includes: Based on the center positions of multiple dust regions belonging to the same dust region in the inter-frame correlation results, calculate the position difference between the center positions of two adjacent dust regions and the acquisition time difference between the acquisition time tags corresponding to two adjacent dust regions; based on the position difference and the acquisition time difference, calculate the motion velocity component and motion direction component of the center position of the dust region between two adjacent dust images. The motion velocity components and motion direction components in the inter-frame correlation results are weighted and averaged to obtain the dust motion velocity and dust motion direction.

8. The control method for pointing a fog cannon based on dust image analysis as described in claim 7, characterized in that, The inter-frame association results include: The dust images are acquired by the image-type dust detector at multiple acquisition times, and the dust images are arranged in chronological order of acquisition time to obtain a dust image sequence. Each frame of the dust image sequence is subjected to grayscale processing and edge enhancement processing to obtain a preprocessed dust image; the preprocessed dust image is then subjected to threshold segmentation to extract connected regions with grayscale values ​​greater than a preset grayscale threshold and connected region areas greater than a preset area threshold, which are used as candidate dust regions. For the candidate dust region, the aspect ratio, edge diffusion, and the difference between the mean gray value inside the candidate dust region and the mean gray value of the external background are calculated. Based on the matching results of the aspect ratio, edge diffusion, and the difference with the preset dust characteristic conditions, candidate dust regions that meet the preset dust characteristic conditions are selected from the candidate dust regions and designated as dust regions. The association process is repeated sequentially according to the arrangement order of two adjacent dust images in the dust image sequence to obtain multiple dust regions belonging to the same dust region in the dust image sequence and their corresponding center positions, which are used as the inter-frame association results.

9. The control method for pointing a fog cannon based on dust image analysis as described in claim 8, characterized in that, The association process includes: For two adjacent dust images in the dust image sequence, calculate the center position, area, and grayscale histogram of the dust region in the previous dust image, and the center position, area, and grayscale histogram of each dust region in the current dust image. Based on the center position of the dust region in the previous dust image and the preset search radius, the search window range is determined in the current dust image. Within the search window, dust regions whose center position distance from the dust region in the previous dust image is less than a preset distance threshold, whose area difference is less than a preset area difference threshold, and whose grayscale histogram similarity is greater than a preset similarity threshold are selected as the associated dust regions in the current dust image.

10. The control method for pointing a fog cannon based on dust image analysis as described in claim 1, characterized in that, The angle mapping relationship is used to map the detection horizontal angle and detection pitch angle to the initial target horizontal angle and initial target pitch angle.

Images