CN115631138A - Zirconium alloy plate laser cutting quality monitoring method and device - Google Patents

Abstract

The invention discloses a laser cutting quality monitoring method and device for a zirconium alloy plate, belonging to the field of laser cutting. Set up a linear array CMOS camera under the side of the laser cutting section. After the laser cutting of the zirconium alloy sheet is completed, the image information of the cutting surface and the bottom surface of the cutting sheet is obtained through the linear array CMOS camera, and the computer image correction and image defect recognition and classification programs are used. , the side shot section is not corrected, the bottom image is corrected, merged into a picture, and processed by grayscale and grid division, and quickly identify and classify, mark, warn, judge, and evaluate defects. The present invention uses a linear array CMOS camera to integrate the cutting section and the bottom surface of the plate into one image, and simultaneously monitors and detects three types of defects online. Compared with the existing laser cutting quality inspection of zirconium alloy plates, the present invention uses lower software and hardware cost, and developed an on-line monitoring method and device with high functional integration, which improved the reliability of detection and product quality.

Description

A method and device for quality monitoring and detection of laser cutting of zirconium alloy plates

technical field

The invention belongs to the technical field of laser cutting, and more specifically relates to a quality monitoring method and device for laser cutting of zirconium alloy plates.

Background technique

The working principle of laser cutting is to guide the laser generated by the laser through optical devices, focus on the surface of the material to melt and vaporize the material, and at the same time use the compressed gas coaxial with the laser beam to blow away the melted material, and make the laser beam and the material along A certain trajectory makes relative movement, thereby forming a certain shape of the slit. Laser cutting technology can be used in the processing of metal and non-metal materials, which can greatly reduce processing time, reduce processing costs, and improve workpiece quality. It is widely used in the cutting of metal sheets in production. Zirconium alloy has good plasticity and can be made into pipes, plates, etc. It is mainly used in the petroleum field and nuclear technology field.

At present, the cutting environment of the laser cutting process is complicated, and the selection of laser cutting parameters such as laser power, cutting speed, gas flow, defocusing amount, etc. is not suitable, and many defects are prone to occur during cutting, such as hanging slag, spatter, cut surface If the roughness exceeds the standard, the key quality indicators of laser cutting products will not be up to standard, which will have a serious impact on the performance of the final assembled product, and even cause serious safety accidents. However, the traditional quality inspection of laser cutting products relies on manual inspection, and due to the fatigue of workers, there may be misjudgment and missed judgment, which consumes a lot of manpower, and the reliability of the inspection results needs to be improved. In addition, the traditional laser cutting does not have a detection device for defects such as dross, spatter, and surface roughness exceeding the standard. It cannot evaluate the products produced by laser cutting of zirconium alloy sheets in a short period of time. It can only be manually inspected after cooling. Whether the slag, spatter, and surface roughness are up to standard, it takes a lot of time to evaluate the cutting quality of laser cutting products. Therefore, it is necessary to develop a highly functionally integrated, accurate and efficient online monitoring method and device for laser cutting quality.

Contents of the invention

In view of the defects of the prior art, the purpose of the present invention is to provide a method and device for laser cutting quality monitoring of zirconium alloy plates, aiming at solving the problem of how to save a lot of manpower and improve the reliability of laser cutting product detection.

In order to achieve the above object, in the first aspect, the present invention provides a method for monitoring and detecting the quality of laser cutting of zirconium alloy plates, the method comprising:

S1. Obtain the picture including the cutting surface and the bottom surface after the laser cutting of the zirconium alloy plate, after correcting the bottom surface of the plate, merge it with the uncorrected cutting section picture into one picture, and perform grayscale and grid division on the image;

S2. Calculate the ratio of the distribution standard deviation and the distribution root mean square height value of the gray histogram of each grid, and compare them with the threshold corresponding to the maximum acceptable roughness respectively. If it exceeds, the grid has a surface roughness exceeding the standard defect , otherwise, the grid does not have defects of surface roughness exceeding the standard;

S3. Input the grayscale images of each grid into the convolutional neural network to obtain the probability of slag defects and splash defects in each grid, and calibrate the defects according to the probability;

S4. Based on the quantity and area of the surface roughness exceeding the standard, hanging slag and splashing three types of defects, the cutting quality of the plate is evaluated.

Preferably, in step S1, correcting the bottom surface of the picture includes:

1) Based on the perspective transformation, identify the intersection line between the bottom surface and the cutting surface and extend it to the edge of the picture;

2) Perspective transformation correction is carried out for the picture below the obtained extension line;

3) Merge the corrected lower picture along the extension line with the cut surface to obtain the final corrected picture.

Preferably, the horizontal distance between the shooting device and the cutting surface is obtained, and the acute angle between the vertical plane of the axis of the shooting device and the bottom surface of the plate is obtained according to the relationship of trigonometric functions as the angle for bottom surface correction.

Preferably, before inputting each gray-scaled grid picture to the convolutional neural network, a selective search algorithm is used for preliminary screening of suspicious grid pictures.

Preferably, in step S3, for the defect determination results obtained, if the maximum probability exceeds 95%, it is determined that the defect exists, and the area is directly marked with a red box, and if the maximum probability is 75% to 95%, it is considered suspected to exist The defect is marked with a yellow box in this area, and the area with the maximum probability below 75% is not marked, and finally it is easy to distinguish when it is displayed on the computer interface through a diagram.

Preferably, the method also includes any of the following treatments:

Processing method 1: Through the pop-up window of the visual display interface and the linear array CMOS monitoring system, the three-color light warning and the buzzer warning sound warning are carried out; among them, the red warning light indicates that the defect is serious, and the number or size of the defect exceeds the normal range, and the buzzer sounds The buzzer emits a continuous warning sound; the yellow warning light is on, indicating that the quality of the cutting surface and the bottom surface are within the normal range, but the score is not excellent, and the buzzer emits an intermittent warning sound; the green warning light is on, indicating that the defect is normal and the score is excellent. No warning sound;

Processing method 2: Perform quantity statistics, size statistics, and defect density statistics on the detected defects, and compare with the standard to determine whether each defect meets the standard.

In order to achieve the above object, in the second aspect, the present invention provides a zirconium alloy plate laser cutting quality monitoring device, the device includes: a shooting device, a guide rail, a supporting connection device and an image processing and analysis module;

The guide rail is located under the side of the laser cutting surface and is parallel to the x-axis of the laser cutting, and is used to drive the shooting device to move smoothly without obstacles, and at the same time to ensure that the shooting device can scan the whole picture of the cutting surface and the bottom surface of the laser cutting workpiece at the same time;

The photographing device is fixed on the guide rail through the supporting connection device, and is used for taking pictures including the bottom surface and the cutting surface of the cutting plate after the laser cutting of the zirconium alloy plate is completed according to the control signal, and uploading to the image processing and analysis module;

The image processing and analysis module is used to process and analyze the captured pictures using the method as described in the first aspect, so as to obtain the position and type of defects and the cutting quality of the zirconium alloy plate.

Preferably, the elevation angle of the shooting device is 45°-65°.

Preferably, the device also includes a protective sheet;

The protective sheet is installed on the surface of the lens of the shooting device, and is used to close according to the control signal during the laser cutting process to isolate the lens of the shooting device from the outside world, and to open it during the scanning and collection process of the shooting device to ensure image collection.

Preferably, the device also includes: a micro ventilation device;

The micro-ventilator is installed on the periphery of the lens of the photographing device, and is used for blowing out gas during the scanning and collecting process of the photographing device according to the control signal, so as to remove residual smoke and dust in the viewing angle of the photographing device.

Generally speaking, compared with the prior art, the above technical solution conceived by the present invention has the following beneficial effects:

(1) The present invention proposes a quality monitoring and testing method for laser cutting of zirconium alloy plates. The image information of the bottom part and the cutting surface of the cutting plate is obtained by one shot and analyzed at the same time, which reduces the number of shots and analyzes and reduces the time spent on quality monitoring and testing. . For the acquired picture, the image is corrected through the computer correction program. When correcting, only the part of the bottom surface is corrected, and the cut surface is still kept as a side view. The two parts are fused into a final correction map, and the image is divided into regions, Select, and classify and judge various defects of the image through the convolutional neural network, and display the image defect information on the computer interface, evaluate the cutting quality of the laser-cut zirconium alloy sheet, and quickly know the quality of the laser-cut zirconium alloy sheet The location and type of defects, and the defect information is fed back to the laser cutting system, the laser cutting parameters are adjusted, and the quality of the laser-cut zirconium plate is monitored and tested.

(2) The present invention proposes a laser cutting quality monitoring device for zirconium alloy plates. On the basis of the existing laser cutting quality, a linear array CMOS camera and its motion assistance and protection device are added below the side of the laser cutting seam, and by means of Computer image processing technology realizes the monitoring and detection of three kinds of defects on two surfaces through one photo, and does not make major changes to the overall processing equipment. Therefore, it saves a lot of manpower and improves the reliability of detection. From a long-term perspective Great cost savings.

Description of drawings

Fig. 1 is a flow chart of a quality monitoring method for laser cutting of zirconium alloy plates provided by the present invention.

Fig. 2 is a schematic diagram of camera erection provided by an embodiment of the present invention.

Fig. 3 is a schematic diagram of bottom correction provided by an embodiment of the present invention.

Fig. 4 is a schematic diagram of defect marking results provided by an embodiment of the present invention.

Detailed ways

In order to make the object, technical solution and advantages of the present invention clearer, the present invention will be further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described here are only used to explain the present invention, not to limit the present invention.

Fig. 1 is a flow chart of a quality monitoring method for laser cutting of zirconium alloy plates provided by the present invention. As shown in Figure 1, the method includes:

Step S1: Set up the camera position and angle.

Fig. 2 is a schematic diagram of camera erection provided by an embodiment of the present invention. As shown in Figure 2, a linear array CMOS camera is installed on the bottom side of the zirconium plate. At this time, the camera device is covered with a protective lens, and the preparations for laser cutting the zirconium plate are done, ready for laser cutting. The main viewing direction of the linear array CMOS camera is perpendicular to the intersection line between the cutting surface 1 and the bottom surface 2, and the elevation angle of the linear array CMOS camera is set at 45° to 65°, which highlights the dross and surface of the cutting surface The rough defect outline can also cover the bottom surface information of the larger plate, which is conducive to the accurate identification and classification of defects by the subsequent neural network model.

A guide rail is installed under the side of the laser cutting surface for the movement of the linear array CMOS camera, and the guide rail is installed parallel to the x-axis of the laser cutting. Connect the linear array CMOS camera with the guide rail through the support device to ensure that the motion of the linear array CMOS camera on the guide rail is unobstructed, stable and will not be damaged by the laser.

Step S2: laser cutting the zirconium alloy plate.

Set the parameters of laser cutting, set cutting power, cutting speed, air pressure and other parameters, clamp the zirconium plate, and carry out laser cutting.

Step S3: line array CMOS image acquisition and reconstruction.

After the laser cutting is completed, after the smoke extraction system absorbs most of the smoke and dust generated during the cutting process, open the protective sheet of the camera device, and turn on the ventilation system to dispel the residual smoke and dust in the line-array CMOS camera angle of view, and turn on the line-array CMOS camera. The camera simultaneously collects optical images of the cutting surface and the bottom surface of the cut piece along the pre-installed guide rail.

The protective sheet of the line array CMOS camera can be used to isolate the lens of the camera device from the outside world through the kit and the protective sheet during the laser cutting process. When the line array CMOS camera scans and collects information, the protective sheet is opened by pulling the soft rope through the motor, without disturbing the image of the camera device. For information collection, the protection sheet is a glass sheet or a rubber sheet, and the shape of the protection sheet is a rectangle.

A small ventilation and smoke removal system is installed around the line array CMOS camera. During the image acquisition process of the camera device, the residual smoke and dust in the viewing angle of the camera device are removed to ensure that the image information collected by the camera is not disturbed by the residual smoke and dust.

The linear CMOS camera moves from the starting end of the plate cutting along the direction of the cutting track to the end of the cutting on the guide rail driven by a servo motor, and completes the reconstruction of the image information of the entire cutting surface and the bottom surface of the plate. The reconstructed cutting surface and the bottom surface are not the front view of the cutting surface and the bottom surface, and the presented image is formed by the connection of two trapezoids.

Step S4: Bottom image correction and cross-sectional image fusion.

Since the camera is placed under the side of the cutting seam, the bottom surface of the collected image and the cutting surface are both angled side views, and the image needs to be corrected. The cutting surface is not corrected, and the bottom surface of the plate is corrected. Then the two parts Fusion to get rectified map.

S4.1: Image correction of the bottom surface of the plate: the correction process adopts the improved perspective transformation based on the geometric intersection of the bottom surface and the cutting surface: the perspective transformation is to rotate the original projection surface around the trace line at a certain angle to obtain the corrected front view. This method is based on Perspective transformation, first identify the intersection line between the bottom surface and the cutting surface and extend it to the edge of the picture, and perform perspective transformation correction on the picture below the obtained extension line, because the perspective transformation will extend and zoom the image, the part that is elongated and enlarged The pixels will change from continuous to discontinuous. Here, the bilinear interpolation method is used, and the gray levels of the four adjacent pixels of the pixel to be obtained are linearly interpolated in two directions to obtain the final corrected image. . Since the detection of the cutting surface needs to pay attention to the roughness value, the correction will easily affect the detection of the roughness, and the correction of the cutting surface has little effect on the detection of dross, so you can choose not to correct the cutting surface.

In the specific correction calculation, a point (x, y, 1) is randomly selected on the image before correction, which corresponds to a point (x′, y′, 1) on the image after correction, and the transformation matrix is H, then the transformation relationship is as follows :

Since the width of the bottom surface of the cut zirconium plate is much larger than the thickness of the sheet, the bottom surface occupies most of the detection area, and the camera placement angle needs to be more biased towards the bottom surface for imaging. After the guide rail is installed, the height from the camera to the bottom surface of the workpiece can be determined. The guide rail is parallel to the x-axis of the laser cutting. After the processing program is loaded, the horizontal distance between the camera and the cutting surface can be obtained. Fig. 3 is a schematic diagram of bottom correction provided by an embodiment of the present invention. As shown in Figure 3, according to the relationship of trigonometric functions, the angle between the camera axis and the intersection line between the cutting surface and the bottom surface can be obtained, and this angle is the angle θ of bottom surface correction.

S4.2: Image fusion of the bottom surface of the plate and the cutting surface: put the corrected bottom surface image and the image directly captured by the line array CMOS camera into one image, and distinguish the bottom surface of the plate from the cutting surface by their boundary line, and the corrected image The bottom surface image is rectangular, but the cut surface image without correction is trapezoidal. The top edge of the rectangle coincides with the bottom edge of the trapezoid, and the dotted line is reconstructed on the coincident line for visual distinction.

Step S5: Defect identification and classification, early warning, marking, judgment and evaluation.

In the identification and classification of defects, based on the quality monitoring method and device for laser cutting of zirconium alloy plates, the image input into the computer is converted into a grayscale image through grayscale processing. The image is grayscale processed to obtain a grayscale image with pixel values in the range [0, 255]. The algorithm used is I(x,y)=a*I_R(x,y)+b*I_G(x,y)+c*I_B(x,y), where a+b+c=1.

S5.1: Defect identification classification: For the detection of surface roughness, the surface roughness of the cutting surface is continuous in a large range, and the surface roughness of the workpiece is related to the standard deviation and root mean square of the image gray histogram distribution. The specific surface The roughness increases with the ratio of the distribution standard deviation of the gray histogram of the image to the root mean square height value of the distribution. The ratio corresponding to each metal is also different. It needs to be photographed for the normal laser-cut zirconium alloy surface. Analyze the grayscale histogram, obtain the distribution standard deviation S and the distribution root mean square height value H of the grayscale histogram, determine the threshold by calculating the ratio S/H, select 100 images for testing to determine the threshold, and test the 100 images In the result, in the data whose roughness is in the normal range, the corresponding maximum value of S/H is selected as the threshold. Cut the corrected picture into small areas to analyze the gray histogram of the picture, and judge whether the surface roughness of the area exceeds the standard according to the threshold value corresponding to the maximum acceptable roughness obtained through the test in advance. The judgment standard is that the calculated value is greater than the threshold value The surface roughness exceeds the standard. In this embodiment, the standard for setting the surface roughness value of the laser-cut zirconium alloy sheet is Ra 3.2.

For the detection of slag and splash, first divide the picture into multiple grids with a side length of 84×84 pixels, and screen the areas to select all suspicious areas. Here, a selective search algorithm is selected, and the specific steps are as follows: : First, divide the input image by double-threshold initialization, recommend the target area, calculate the similarity between adjacent sub-areas in the image from the texture features and overlap of adjacent areas, and finally merge adjacent similar areas continuously , to obtain a smaller number of target areas, thereby narrowing the target search range. Compared with the traditional sliding window exhaustive detection method, this method reduces redundant areas and improves detection efficiency.

Crop the obtained suspicious area into multiple 42×42 pixel pictures. The defect taken by the line array CMOS camera is 0.1mm level, and the 42×42 pixel photo can include the defect. Using the convolutional neural network, input the cropped The 42×42 pixel picture, through 4~6 layers of convolution, pooling, activation, etc., extracts the features of the defect, and realizes the convolution through flattening between the convolution layer and the fully connected layer. The transition between the fully connected layer is input to the fully connected layer 3, and finally the Softmax activation function is used to output the probability value for determining the existence of various types of defects in the image.

The training of the neural network is obtained from the cut products. The input data are 1000 42×42 pixel pictures containing slag defects and 1000 42×42 pixel pictures containing splash defects. For pictures, 80% of the obtained data is used for the training of the CNN convolutional neural network, and the other 20% is used for prediction. When the accuracy of the predicted value reaches more than 99%, it can be considered that the trained neural network is reliable and can be used in actual production.

S5.2: Defect early warning: through the pop-up window of the visual display interface and the linear array CMOS monitoring system, the three-color light early warning and the buzzer warning sound early warning. If the red warning light is on, it means that the defect is serious. If the number or size of the defects exceeds the normal range, the buzzer will sound a continuous warning; Intermittent warning sound; the green warning light is on, indicating that the defect is normal, and the score is excellent, and there is no warning sound.

S5.3: Defect marking: display the marked area of the defect area through the computer interface, and use red for the defect area. For the defect judgment results obtained, the area with the maximum probability exceeding 95% is directly marked with a red frame, the area with the maximum probability between 75% and 95% is marked with a yellow frame, and the area with the maximum probability below 75% is only for two types of defects. For detection, the maximum defect value must not be less than 50%. When the output maximum probability value of a certain defect is not satisfied and is much greater than another, it is considered not to be a defect probability and will not be marked. Finally, it is easy to distinguish when the computer interface is displayed graphically. Fig. 4 is a schematic diagram of defect marking results provided by an embodiment of the present invention. As shown in Figure 4, small circles represent dross defects, ovals represent splash defects, and squares represent roughness defects.

S5.4: Defect judgment: compare with the standard to see whether a single defect meets the standard. Perform quantity statistics, size statistics, and defect density statistics on the detected defects. The number of dross cannot exceed 4, the density cannot exceed 2mm ² /dm ² , the number of splashes cannot exceed 10, and the density cannot exceed 3mm ² /dm ² . The area exceeding the standard cannot exceed 12mm ² .

S5.5: Defect evaluation: combined with the actual measured size and the area of the area where the surface roughness exceeds the standard, the size and quantity of dross and splash, comprehensively evaluate the quality of the plate: based on the key factors such as surface roughness, dross, and splash The comprehensive quality score is obtained by weighted average: the above-mentioned slag, splash, and surface roughness quality evaluation standards are set as 60 points, and no defect is set as a full score of 100 points. The proportions in the scoring are equal, the scores of each index are 60 points according to the standard value, and 100 points for no defect, the scores are evenly distributed with the quality data, and the overall quality score is calculated by weighting the above three defect scores according to the ratio of 1:1:1 , and divide the quality into 4 grades in turn: 90 points and above are excellent, 75 points and above are good, 60 points and above are qualified, and below 60 points are unqualified. And for a certain item of data that has exceeded the standard value, it is classified as unqualified.

Furthermore, after the evaluation of laser cutting products is completed, the detected problems can be fed back to the laser cutting system, and the laser cutting parameters can be adjusted in time to improve product quality and achieve the purpose of monitoring product quality.

It is easy for those skilled in the art to understand that the above descriptions are only preferred embodiments of the present invention, and are not intended to limit the present invention. Any modifications, equivalent replacements and improvements made within the spirit and principles of the present invention, All should be included within the protection scope of the present invention.

Claims (10)

1. A zirconium alloy plate laser cutting quality monitoring and detection method is characterized in that the method comprises: S1. Obtain the picture including the cutting surface and the bottom surface after the laser cutting of the zirconium alloy plate, after correcting the bottom surface of the plate, merge it with the uncorrected cutting section picture into one picture, and perform grayscale and grid division on the image; S2. Calculate the ratio of the distribution standard deviation and the distribution root mean square height value of the gray histogram of each grid, and compare them with the threshold corresponding to the maximum acceptable roughness respectively. If it exceeds, the grid has a surface roughness exceeding the standard defect , otherwise, the grid does not have defects of surface roughness exceeding the standard; S3. Input the grayscale images of each grid into the convolutional neural network to obtain the probability of slag defects and splash defects in each grid, and calibrate the defects according to the probability; S4. Based on the quantity and area of the surface roughness exceeding the standard, hanging slag and splashing three types of defects, the cutting quality of the plate is evaluated. 2. The method according to claim 1, characterized in that, in step S1, correcting the bottom surface of the picture includes: 1) Based on the perspective transformation, identify the intersection line between the bottom surface and the cutting surface and extend it to the edge of the picture; 2) Perspective transformation correction is carried out for the picture below the obtained extension line; 3) Merge the corrected lower picture along the extension line with the cut surface to obtain the final corrected picture. 3. The method according to claim 2, wherein the horizontal distance between the camera and the cutting surface is obtained, and according to the relationship of trigonometric functions, the angle between the vertical plane of the axis of the camera and the bottom surface of the sheet is obtained as the acute angle of the bottom surface. angle. 4. The method according to any one of claims 1 to 3, characterized in that, before inputting each grayscale image of each grid into the convolutional neural network, a selective search algorithm is used for preliminary screening of suspicious grid images . 5. The method according to claim 1, characterized in that, in step S3, for the obtained defect determination result, if the maximum probability exceeds 95%, it is determined that the defect exists, and the area is directly marked with a red frame, the maximum If the probability is between 75% and 95%, it is considered that the defect is suspected to exist, and a yellow frame mark is attached to the area, and the area with the maximum probability below 75% is not marked, and finally it is easy to distinguish when it is displayed on the computer interface through graphics. 6. The method according to claim 1, further comprising any of the following processing methods: Processing method 1: Through the pop-up window of the visual display interface and the linear array CMOS monitoring system, the three-color light warning and the buzzer warning sound warning are carried out; among them, the red warning light indicates that the defect is serious, and the number or size of the defect exceeds the normal range, and the buzzer sounds The buzzer emits a continuous warning sound; the yellow warning light is on, indicating that the quality of the cutting surface and the bottom surface are within the normal range, but the score is not excellent, and the buzzer emits an intermittent warning sound; the green warning light is on, indicating that the defect is normal and the score is excellent. No warning sound; Processing method 2: Perform quantity statistics, size statistics, and defect density statistics on the detected defects, and compare with the standard to determine whether each defect meets the standard. 7. A zirconium alloy plate laser cutting quality monitoring device, characterized in that the device comprises: a photographing device, a guide rail, a supporting connection device and an image processing and analysis module; The guide rail is located under the side of the laser cutting surface and is parallel to the x-axis of the laser cutting, and is used to drive the shooting device to move smoothly without obstacles, and at the same time to ensure that the shooting device can scan the whole picture of the cutting surface and the bottom surface of the laser cutting workpiece at the same time; The photographing device is fixed on the guide rail through the supporting connection device, and is used for taking pictures including the bottom surface and the cutting surface of the cutting plate after the laser cutting of the zirconium alloy plate is completed according to the control signal, and uploading to the image processing and analysis module; The image processing and analysis module is used to process and analyze the captured pictures by the method according to any one of claims 1 to 6, so as to obtain the defect position and type and cutting quality of the zirconium alloy plate. 8. The device according to claim 7, wherein the elevation angle of the photographing device is 45°-65°. 9. The device of claim 7, further comprising a protective sheet; The protective sheet is installed on the surface of the lens of the shooting device, and is used to close according to the control signal during the laser cutting process to isolate the lens of the shooting device from the outside world, and to open it during the scanning and collection process of the shooting device to ensure image collection. 10. The device according to claim 7, further comprising: a micro ventilation device; The micro-ventilator is installed on the periphery of the lens of the photographing device, and is used for blowing out gas during the scanning and collecting process of the photographing device according to the control signal, so as to remove residual smoke and dust in the viewing angle of the photographing device.

Images