CN110926737B - Intelligent screen plate fault monitoring method based on depth image - Google Patents
Intelligent screen plate fault monitoring method based on depth image Download PDFInfo
- Publication number
- CN110926737B CN110926737B CN201911186854.2A CN201911186854A CN110926737B CN 110926737 B CN110926737 B CN 110926737B CN 201911186854 A CN201911186854 A CN 201911186854A CN 110926737 B CN110926737 B CN 110926737B
- Authority
- CN
- China
- Prior art keywords
- depth
- data
- camera
- image
- signal intensity
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Images
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01M—TESTING STATIC OR DYNAMIC BALANCE OF MACHINES OR STRUCTURES; TESTING OF STRUCTURES OR APPARATUS, NOT OTHERWISE PROVIDED FOR
- G01M7/00—Vibration-testing of structures; Shock-testing of structures
- G01M7/02—Vibration-testing by means of a shake table
- G01M7/025—Measuring arrangements
Abstract
The invention relates to a screen plate fault intelligent monitoring method based on a depth image, which comprises the following steps: step 1, determining the matrix arrangement of depth camera installation according to the station of the sieve plate and the internal parameters of the camera; step 2, carrying out dimension conversion on the original data; step 3, processing invalid data in the data after the dimension conversion; step 4, performing pseudo-color rendering on the data after the invalid data processing; step 5, dividing sub-images in the depth image according to the station of each screen plate, and analyzing and judging the working condition of each sub-image independently; and 6, acquiring signal intensity data of the image captured by the camera, and comparing the signal intensity data with the signal intensity data in normal operation to judge whether coal exists on the sieve plate or not and further judge whether the raw coal hopper is blocked or not. Compared with the existing method, the method has the advantages of high detection speed and high detection precision, does not need manual intervention, is a non-invasive monitoring method, and does not influence normal production.
Description
Technical Field
The invention relates to the field of intelligent monitoring of working conditions of large mechanical equipment of an industrial automatic production line, in particular to a screen plate fault intelligent monitoring method based on a depth image, which is used for intelligent monitoring of faults of screen plate inclination or shedding of mechanical vibration.
Background
A vibrating screen is a mechanical device with a complex structure, which is accompanied by severe vibration during starting and running, the vibration amplitude being relatively small but the vibration frequency being large; in the process of stopping and cushioning, the motor drives the vibration frequency of the sieve plate to be close to the self vibration frequency of the sieve plate, resonance occurs, the vibration frequency of the sieve plate is reduced in the shutdown and stopping stage, but the vibration amplitude is increased, and therefore the vibration in different degrees is accompanied in the whole process of the vibrating screen. And the sieve plate is only fixed by the sieve plate buckle and the sieve machine gauge seat. The sieve plate buckle and the sieve gauge seat are both made of polyurethane materials with certain elasticity, and although the materials have the characteristics of high elasticity and corrosion resistance, when the sieve plate is subjected to strong mechanical vibration, the sieve plate is inclined and falls off occasionally.
The sieve dropout trouble is one of the trouble that shale shaker frequently appeared in daily production operation, the shale shaker has the wearing and tearing phenomenon of polyurethane sieve and sieve machine gauge seat in daily production operation, and then cause the sieve buckle not hard up, if the sieve machine continues the operation, the sieve dropout trouble probably takes place, if can not in time discover after the sieve takes off, if there is the material on the sifter, then have a large amount of materials to get into under-sieve chute and pipeline in, cause production accidents such as system's pipeline jam, and then cause whole medium circulation system paralysis and unable normal production.
The existing sieve plate falling fault does not have any on-line monitoring protection device, inspection and patrol can only be carried out by field operators in a mode of empirical judgment, and the fault monitoring has serious hysteresis, so that the problems of fault amplification and the like can be caused.
The two-layer sieve plate falling material abnormal state alarm device of the medium removing sieve is used in the prior art, the state of the sieve plate is monitored in real time, and the problem of hysteresis of judging faults by means of human experience is solved. The device has certain detection effect on the condition that a large amount of screen surface materials slide down after the screen plate falls off, but the detection method has obvious loopholes and defects: for example, the detection mode belongs to hysteresis detection; the detection mode belongs to physical intervention detection; this detection method requires a complicated circuit design and the like.
The analysis can find that an effective monitoring method for the working condition of the sieve plate is not available. The special working environment of the site requires that an interventional or interventional monitoring method cannot be adopted, and the normal production cannot be influenced; in a severe field environment, a monitoring mode of the charged property cannot be adopted.
Aiming at the problems of hysteresis, misjudgment, non-real-time property and the like in the existing sieve plate fault monitoring technology, a real-time monitoring system aiming at the operation state of the sieve plate is needed to be designed, so that faults can be found in time, an alarm can be given in time, and the accident hazard can be reduced to the minimum.
Disclosure of Invention
The invention aims to overcome the defects in the prior art, and provides a screen plate fault intelligent monitoring method based on a depth image.
In order to achieve the purpose, the invention has the following conception:
according to the method, the depth camera is arranged at the position to be detected above the screen surface of the vibrating screen, the camera is an active measuring camera, has the characteristic of self-luminescence, and does not need an external light source. And adjusting the direction of a camera lens, acquiring depth data during the operation of the screen surface, and carrying out online real-time analysis processing on the acquired depth data. The key of the invention is a rapid method for diagnosing the screen plate falling fault, and an intelligent detection algorithm for processing the screen surface depth data comprises the steps of preprocessing original data, positioning the region of each screen plate in an image, extracting the characteristic parameters of the fallen screen plate, calculating and comparing the characteristic parameters with normal characteristic data and the like. And the screen plate working condition on-line monitoring system based on the depth image processes the screen surface depth data, outputs fault information and transmits the fault information to a server interface for dynamic display.
The depth camera is arranged above a vibrating screen system where a screen plate is likely to fall off, and the lens is opposite to the screen surface, so that the maximum signal intensity which is likely to be acquired is ensured. Because the range covered by the lens is limited, the matrix arrangement formed by the cameras is considered when the cameras are installed, so that the lens covers a certain overlapping area, each screen plate is ensured to be in real-time monitoring, and meanwhile, the signal intensity reflected by each screen plate is kept in an ideal range with accurate data. The server interface comprises an industrial computer and intelligent monitoring software for the failure of the sieve plate.
The depth camera based on the light reflection principle is utilized in the invention, and the distance between an object and the camera is taken as a main characteristic basis. Judging the operation condition of the sieve plate at the current moment by analyzing the depth data of the sieve plate; and constructing a model of normal operation of the screen surface and inclination or falling of the screen plate through the collected depth data, and visually judging and positioning the fault.
According to the conception, the invention adopts the following technical scheme:
an intelligent screen plate fault monitoring method based on depth images is used for obtaining the operation working condition of a screen plate and locating a fault point where the screen plate falls off in the operation process of a vibrating screen system. The method comprises the following steps:
step 1, determining the installation number and the installation position of the depth cameras according to stations of the screen plate and internal parameters of the cameras, including ideal signal intensity of the cameras when the cameras run on the screen surface, the coverage range of camera lenses and the like.
And 2, realizing communication between the server and the depth camera, preprocessing the acquired original data, and cutting the original data according to the size of the picture to be displayed, namely converting continuous array data into two-dimensional data.
And 3, processing invalid data in the preprocessed two-dimensional data, including some pixel points with too low signal intensity and overexposure points with too high intensity, and performing corresponding filtering processing on the pixel points.
And 4, performing pseudo-color rendering on the preprocessed two-dimensional data, namely mapping the depth data and the gray value of each pixel point, and mapping the gray image into a pseudo-color image according to a pseudo-color algorithm for clearly and intuitively displaying a field real-time picture on a server.
And 5, dividing areas corresponding to the sieve plates in the depth image according to the spatial distribution of the sieve plates on the working condition site, and judging the working condition of each area of the sieve plate respectively, so that the fault point can be conveniently positioned.
And 6, acquiring signal intensity data reflected by the screen surface, and comparing the signal intensity data with the signal intensity data in normal operation to judge whether coal exists on the screen plate or not and further judge whether the raw coal hopper is blocked or not.
The step 1 comprises the following steps:
step 1.1, selecting a proper camera lens module by combining the coverage range of each screen plate according to the optimal working height range of the camera during operation;
step 1.2, installing a camera, adjusting the focal length of the camera according to the distance between the camera and the screen surface, and finding out the optimal lens built-in parameters of the depth camera shooting screen plate according to external environment conditions;
and step 1.3, determining a camera matrix including the installation position of each camera and the number of cameras required by each screening machine according to the size of the screening surface in each set of equipment and the minimum signal intensity of the camera reflection required by the algorithm.
The step 2 comprises the following steps:
step 2.1, communication between the server and the depth camera is achieved, and a Socket communication mode is adopted between the server and the depth camera; the method comprises the steps that a server firstly sends a connection request instruction to a depth camera, and the connection instruction not only completes communication with the depth camera, but also informs the depth camera to initialize according to the server instruction;
step 2.2, normal data acquisition between the server and the depth camera is realized, namely after the depth camera completes initialization, a depth data acquisition instruction, a signal intensity instruction and a bottom chip temperature instruction are sent to the camera, and camera initialization parameters are readjusted;
step 2.3, after receiving the corresponding instruction, the camera starts to collect corresponding data and sends the collection result to the server in a data stream form; at the moment, the server receives and stores the original data, and analyzes the original data into a two-dimensional array according to the size of the picture.
The step 3 comprises the following steps:
3.1, due to the change of external environment conditions, invalid data possibly exist in the original two-dimensional array, wherein the invalid data comprise data with too low signal intensity and overexposure data with too high signal intensity; at the moment, filtering the original data according to the set minimum signal intensity and maximum signal intensity data; replacing the data smaller than the minimum value according to the minimum value, and replacing the data larger than the maximum value according to the maximum value;
3.2, due to some reasons, noise possibly exists in the pixel field, and the depth data of the noise and the depth data of the surrounding pixel points are obviously different; in industrial fields, the situation that the depth value of adjacent pixel regions is not sudden change is impossible, so noise is filtered.
The step 4 comprises the following steps:
step 4.1, carrying out pseudo color conversion on the two-dimensional depth data; firstly, determining the maximum value and the minimum value in two-dimensional depth data, wherein the maximum value and the minimum value respectively correspond to a gray value of 255 and a gray value of 0 in a gray image; converting the depth value in each pixel point into a gray value according to the following formula:
where g (i, j) represents the depth value of the pixel point at i row and j column after filtering, mindepthIndicates the minimum value of depth, max, in the depth mapdepthRepresenting the maximum depth, g, in the depth imagegray(i, j) represents the gray value at i row and j column after undergoing gray conversion;
step 4.2, in order to make the image presented on the server more intuitive so as to identify more image details, and have stronger distinguishability, the gray image needs to be subjected to pseudo-color processing, that is, image rendering is performed according to certain algorithm according to the gray value in the pixel point, that is, a single-channel image is converted into a three-channel image:
R(i,j)＝fR[ggray(i,j)]
G(i,j)＝fG[ggray(i,j)]
B(i,j)＝fB[ggray(i,j)]
wherein g isgray(i, j) represents the gray value at i row and j column after undergoing gray conversion, fRRepresenting a mapping relationship from the gray-scale image to the red channel; f. ofGRepresenting the mapping from the grayscale image to the green channel, fBAnd a mapping relation from the gray image to a blue channel is expressed, R (i, j) represents a value of a red channel at i rows and j columns after pseudo color rendering, G (i, j) represents a value of a green channel at i rows and j columns after pseudo color rendering, and B (i, j) represents a value of a blue channel at i rows and j columns after pseudo color rendering.
The step 5 comprises the following steps:
step 5.1, normally acquiring a field real-time operation image, clearly and visually judging the position of each sieve plate through a server interface, obtaining the coordinate of the corresponding sieve plate in the image by selecting the position of each sieve plate in the image, and selecting an ROI (region of interest) in the original depth map according to the coordinate;
step 5.2, after obtaining each area needing to be judged, obtaining an original depth image, selecting an area corresponding to the sieve plate in the original image, counting pixels in the area one by one, calculating the average depth value of the pixels in the area, comparing the average depth value with a normal depth threshold value, and judging that the sieve plate has faults at the moment when the average depth is greater than the depth threshold value; and calculating the difference d between the average depth value and the depth threshold value when d is larger than or equal to the thread1Judging that the sieve plate has a falling fault at the moment; when thread2≥d≥thread1Judging that the screen plate has a tilt fault at the moment and judging the threshold value of the screen plate1And thread2Setting by combining data and experience collected by a camera;
and 5.3, judging the working condition of each screen plate through the steps, identifying the normal screen plate on a screen plate simulation interface on the server through a gray color block, identifying the fault screen plate on the screen plate simulation interface through a red color block, and printing a corresponding log on the server interface.
The step 6 comprises the following steps:
step 6.1, the system server sends an instruction, and after the instruction for acquiring the depth data is sent for 5 times, the instruction for acquiring the signal intensity is sent once;
6.2, after the signal intensity data are obtained, analyzing the original data according to the method in the step 3 to obtain the signal intensity data corresponding to each pixel point;
6.3, selecting a central area of the sieve plate, comparing the signal intensity in the central area with the signal intensity in normal operation, namely under the condition that coal exists on the sieve surface, and if the signal intensity fluctuates up and down in a normal range, considering that the system operates normally at the moment, wherein the fluctuation of the signal intensity may be caused by the thickness change of the coal bed on the surface of the sieve surface; if the signal intensity is obviously increased, the situation that no coal exists on the screen surface is shown;
6.4, judging whether the screen surface has coal or not through the steps, wherein the condition that the system runs normally when the screen surface has the coal indicates that the system runs normally is judged, and the server interface is marked through a gray color block; if no coal is on the screen surface, the system idles at the moment, or the screen surface has no coal due to the blockage of a raw coal hopper at the upstream of the screen plate, and the screen surface is marked by a red color block at the moment so as to prompt a user.
Compared with the prior art, the invention has the following advantages:
1. the method is easy to realize, does not need manual intervention, and automatically detects the fault in real time.
2. And the fault is detected in real time, the detection speed is high, and the precision is high.
3. And displaying the fault grade and fault point positioning information in real time.
Drawings
FIG. 1 is a flow chart of the method of the present invention;
fig. 2 is a screen plate real-time monitoring diagram according to an embodiment of the present invention;
fig. 3 shows an intelligent monitoring software interface for screen plate detachment according to an embodiment of the present invention.
Detailed Description
The technical solution in the embodiments of the present invention is clearly and completely described below with reference to the accompanying drawings.
As shown in figure 1, the screen plate fault intelligent monitoring method based on the depth image, disclosed by the invention, is characterized in that a depth camera is used for capturing a field picture in real time, the real-time working condition of the operation of the screen plate is obtained through data preprocessing and data analysis, whether the screen plate is in fault or not is judged, a fault point is positioned, whether the screen plate idles or not is judged, and the like. The method specifically comprises the following steps:
step 1, determining the matrix arrangement of the installation of the depth cameras according to the stations of the sieve plate and the internal parameters of the cameras.
And 1.1, selecting a proper camera lens module by combining the coverage range of each screen plate according to the optimal working height range of the camera during operation. For example, when the ideal working height of the camera is 2.5m, the width of the screening surface is just not more than 5m, and is close to 5m, the lens with the coverage range of 900 is selected to be most suitable. If the ideal working height of the camera is too small relative to the width of the screen deck, a small lens may be used to increase the camera coverage width by increasing the number of cameras.
And step 1.2, installing a camera, adjusting the focal length of the camera according to the distance between the camera and the screen surface, and finding the optimal lens built-in parameters of the depth camera shooting screen plate according to external environment conditions.
And step 1.3, determining a camera matrix including the installation position of each camera and the number of cameras required by each screening machine according to the size of the screening surface in each set of equipment and the minimum signal intensity of the camera reflection required by the algorithm.
And 2, realizing communication between the server and the depth camera, and carrying out dimension conversion on the acquired original data.
And 2.1, realizing communication between the server and the depth camera, wherein a Socket communication mode is adopted between the server and the depth camera. The server first sends an instruction to the depth camera requesting a connection, which not only completes the communication with the depth camera, but also informs the depth camera to initialize according to the server instruction. The specific instructions comprise: an overexposure point enabling instruction, wherein the instruction can inform the depth camera to process the overexposure point; camera integration time is enabled, the instruction can adjust the maximum detection distance of the camera, and meanwhile, the increase of the integration time can also lead to the same distance, the increase of signal intensity and even overexposure of pixel points; and setting a minimum signal strength command to complete the camera minimum signal strength enable.
And 2.2, acquiring normal data between the server and the depth camera, namely after the depth camera completes initialization, sending a depth data acquisition instruction, a signal intensity instruction and a bottom chip temperature instruction to the camera, and readjusting initialization parameters of the camera. The method specifically comprises the following instructions: sending a 'getDistantSorted' instruction to obtain depth data of the sieve plate; sending a 'getAmplexituported' instruction to acquire signal intensity data captured by a lens; and sending a 'getTemperature' instruction to acquire the temperature data of the bottom chip.
Step 2.3, after receiving the corresponding instruction, the camera starts to collect corresponding data and sends the collection result to the server in a streaming form; at this time, the server receives the data, stores the original data, and analyzes the original data into a two-dimensional array according to the size of the picture.
And 3, processing invalid data in the data after the dimension conversion.
And 3.1, due to the change of the external environment condition, invalid data possibly exists in the original two-dimensional array, wherein the invalid data comprises data with too low signal intensity and overexposure data with too high signal intensity. The raw data may be filtered at this point based on the set minimum and maximum signal strength data. Data smaller than the minimum value is replaced by the minimum value, and data larger than the maximum value is replaced by the maximum value.
And 3.2, due to some reasons, noise possibly exists in the pixel field, and the depth data of the noise and the depth data of the surrounding pixel points are obviously different. In industrial fields, the situation that the depth value of adjacent pixel regions is not sudden change is impossible, so noise is filtered. In the embodiment, a mean value filtering algorithm is adopted, and the depth values of the adjacent pixels are used for compensating for the depth values of the mutation pixel points.
Assume that in a square region (3 x 3) of the picture, the pixel value of the center point is the average value of all the pixels. The mean filtering is to perform the above operation on the whole graph. Namely, it is
Wherein i and j respectively represent the row and the column of the pixel point in the image, f (i, j) represents the depth value of the pixel point at the i row and the j column in the two-dimensional image before filtering, s represents the whole image area, M is the total number of the pixel points in the template area, and g (i, j) represents the depth value of the pixel point at the i row and the j column after filtering.
Under the working condition, the noise in the image collected by the camera is mainly Gaussian noise, and the filtering algorithm has a good effect.
And 4, as shown in FIG. 2, performing pseudo-color rendering on the data after the invalid data processing.
Step 4.1, determining the maximum value and the minimum value in the two-dimensional depth data, wherein the maximum value and the minimum value respectively correspond to a gray value of 255 and a gray value of 0 in the gray image, and converting the depth value in each pixel point into the gray value according to the following formula:
where g (i, j) represents the depth value of the pixel point at i row and j column after filtering, mindepthIndicates the minimum value of depth, max, in the depth mapdepthRepresenting the maximum depth, g, in the depth imagegray(i, j) represents the gray value at i row j column after undergoing gray conversion.
Step 4.2, in order to make the image presented on the server more intuitive so as to identify more image details, the resolution is stronger, the gray image needs to be processed with pseudo-color, that is, the image is rendered according to a certain algorithm according to the gray value in the pixel point, that is, the gray image is converted into a color image, so that the single-channel image is converted into a three-channel image:
R(i,j)＝fR[ggray(i,j)]
G(i,j)＝fG[ggray(i,j)]
B(i,j)＝fB[ggray(i,j)]
wherein g isgray(i, j) represents the gray value at i row and j column after undergoing gray conversion, fRRepresenting a mapping relationship from the gray-scale image to the red channel; f. ofGRepresenting the mapping from the grayscale image to the green channel, fBAnd a mapping relation from the gray image to a blue channel is expressed, R (i, j) represents a value of a red channel at i rows and j columns after pseudo color rendering, G (i, j) represents a value of a green channel at i rows and j columns after pseudo color rendering, and B (i, j) represents a value of a blue channel at i rows and j columns after pseudo color rendering.
The specific rendering algorithm is as follows:
dividing the pixel gray-scale value into 5 parts on average, and setting four threshold dividing points between the gray-scale values 0 and 255, wherein s1 is 51, s2 is 102, s3 is 153, and s4 is 204;
the gray value of each part is converted according to the following mapping relation:
| gray | R | G | B | Color |
| 0 | 0 | 0 | 255 | blue |
| 51 | 0 | 255 | 255 | cyan |
| 102 | 0 | 255 | 0 | green |
| 153 | 255 | 255 | 0 | yellow |
| 204 | 255 | 127 | 0 | orange |
| 255 | 255 | 0 | 0 | red |
and 5, as shown in fig. 3, judging the working condition of the sieve plate and positioning a fault point.
And 5.1, normally acquiring a field real-time running image, clearly and visually judging the position of each screen plate through a server interface, obtaining the coordinate of the corresponding screen plate in the image by selecting the position of each screen plate in the image, and selecting an ROI (region of interest) in the original depth map according to the coordinate.
And 5.2, after obtaining each area needing to be judged, acquiring an original depth image, selecting an area corresponding to the sieve plate in the original image, counting pixels in the area one by one, calculating the average depth value of the pixels in the area, comparing the average depth value with a normal depth threshold value, and judging that the sieve plate has faults at the moment when the average depth is greater than the depth threshold value. And calculating the difference d between the average depth value and the depth threshold value when d is larger than or equal to the thread1Then, the sieve plate can be judged to have a falling fault; when thread2≥d≥thread1In the meantime, it can be determined that the screen plate has a tilt fault at this time, and the threshold value thread1And thread2The settings are made in conjunction with data and experience collected by the camera.
And 5.3, judging the working condition of each screen plate through the steps. As shown in fig. 3, a normal screen plate is identified on a screen plate simulation interface on the server through a gray color block, a failed screen plate is identified on the screen plate simulation interface through a red color block, and a corresponding log is printed on the server interface.
And 6, acquiring the signal intensity and judging whether the sieve plate idles.
And 6.1, the system server sends an instruction, and after the instruction for acquiring the depth data is sent for 5 times, the instruction for acquiring the signal intensity is sent once.
And 6.2, after the signal intensity data are obtained, analyzing the original data according to the method in the step 3 to obtain the signal intensity data corresponding to each pixel point.
6.3, selecting a central area of the sieve plate, comparing the signal intensity in the central area with the signal intensity under the condition of normal operation (coal on the sieve surface), and if the signal intensity fluctuates up and down in a normal range, considering that the system operates normally at the moment, wherein the fluctuation of the signal intensity may be caused by the thickness change of the coal bed on the surface of the sieve surface; if the signal intensity is obviously increased, the situation that no coal exists on the screen surface is shown.
6.4, as shown in FIG. 3, the system operates normally when coal exists on the screen surface, and the system is marked on the server interface through a gray color block; if no coal is on the screen surface, the system idles at the moment, or the screen surface has no coal due to the blockage of a raw coal hopper at the upstream of the screen plate, and the screen surface is marked by a red color block at the moment so as to prompt a user.
So far, the screen plate fault diagnosis, the fault point positioning and the judgment of whether the screen plate idles are finished from the step 1 to the step 6.
While the invention has been described with reference to specific embodiments, the invention is not limited thereto, and various equivalent modifications and substitutions can be easily made by those skilled in the art within the technical scope of the invention.
Claims (7)
1. A screen plate fault intelligent monitoring method based on a depth image is characterized by comprising the following steps:
step 1, determining the matrix arrangement of depth camera installation according to the station of the sieve plate and the internal parameters of the camera;
step 2, realizing communication between the server and the depth camera, and carrying out dimension conversion on the acquired original data;
step 3, processing invalid data in the data after the dimension conversion;
step 4, performing pseudo-color rendering on the data after the invalid data processing;
step 5, dividing areas corresponding to the sieve plates in the depth image according to the spatial distribution of the sieve plates on the working condition site, and respectively judging the working condition of each area of the sieve plate, so as to facilitate fault point positioning;
and 6, acquiring signal intensity data of the image captured by the camera, and comparing the signal intensity data with the signal intensity data in normal operation to judge whether coal exists on the sieve plate or not and further judge whether the raw coal hopper is blocked or not.
2. The intelligent screen plate fault monitoring method based on the depth image as claimed in claim 1, wherein the step 1 comprises the following steps:
step 1.1, selecting a proper camera lens module by combining the coverage range of each screen plate according to the optimal working height range of the camera during operation;
step 1.2, installing a camera, adjusting the focal length of the camera according to the distance between the camera and the screen surface, and finding out the optimal lens built-in parameters of the depth camera shooting screen plate according to external environment conditions;
and step 1.3, determining a camera matrix including the installation position of each camera and the number of cameras required by each screening machine according to the size of the screening surface in each set of equipment and the minimum signal intensity of the camera reflection required by the algorithm.
3. The intelligent screen plate fault monitoring method based on the depth image as claimed in claim 1, wherein the step 2 comprises the following steps:
step 2.1, communication between the server and the depth camera is achieved, and a Socket communication mode is adopted between the server and the depth camera; the method comprises the steps that a server firstly sends a connection request instruction to a depth camera, and the connection instruction not only completes communication with the depth camera, but also informs the depth camera to initialize according to the server instruction;
step 2.2, normal data acquisition between the server and the depth camera is realized, namely after the depth camera completes initialization, a depth data acquisition instruction, a signal intensity instruction and a bottom chip temperature instruction are sent to the camera, and camera initialization parameters are readjusted;
step 2.3, after receiving the corresponding instruction, the camera starts to collect corresponding data and sends the collection result to the server in a data stream form; at the moment, the server receives and stores the original data, and analyzes the original data into a two-dimensional array according to the size of the picture.
4. The intelligent screen plate fault monitoring method based on the depth image as claimed in claim 1, wherein the step 3 comprises the following steps:
3.1, due to the change of external environment conditions, invalid data possibly exist in the original two-dimensional array, wherein the invalid data comprise data with too low signal intensity and overexposure data with too high signal intensity; at the moment, filtering the original data according to the set minimum signal intensity and maximum signal intensity data; replacing the data smaller than the minimum value according to the minimum value, and replacing the data larger than the maximum value according to the maximum value;
3.2, due to some reasons, noise possibly exists in the pixel field, and the depth data of the noise and the depth data of the surrounding pixel points are obviously different; in industrial fields, the situation that the depth value of adjacent pixel regions is not sudden change is impossible, so noise is filtered.
5. The intelligent screen plate fault monitoring method based on the depth image as claimed in claim 1, wherein the step 4 comprises the following steps:
step 4.1, carrying out pseudo color conversion on the two-dimensional depth data; firstly, determining the maximum value and the minimum value in two-dimensional depth data, wherein the maximum value and the minimum value respectively correspond to a gray value of 255 and a gray value of 0 in a gray image; converting the depth value in each pixel point into a gray value according to the following formula:
where g (i, j) represents the depth value of the pixel point at i row and j column after filtering, mindepthIndicates the minimum value of depth, max, in the depth mapdepthRepresenting the maximum depth, g, in the depth imagegray(i, j) represents the gray value at i row and j column after undergoing gray conversion;
step 4.2, in order to make the image presented on the server more intuitive so as to identify more image details, and have stronger distinguishability, the gray image needs to be subjected to pseudo-color processing, that is, image rendering is performed according to certain algorithm according to the gray value in the pixel point, that is, a single-channel image is converted into a three-channel image:
R(i,j)＝fR[ggray(i,j)]
G(i,j)＝fG[ggray(i,j)]
B(i,j)＝fB[ggray(i,j)]
wherein g isgray(i, j) represents the gray value at i row and j column after undergoing gray conversion, fRRepresenting a mapping relationship from the gray-scale image to the red channel; f. ofGRepresenting the mapping from the grayscale image to the green channel, fBAnd a mapping relation from the gray image to a blue channel is expressed, R (i, j) represents a value of a red channel at i rows and j columns after pseudo color rendering, G (i, j) represents a value of a green channel at i rows and j columns after pseudo color rendering, and B (i, j) represents a value of a blue channel at i rows and j columns after pseudo color rendering.
6. The intelligent screen plate fault monitoring method based on the depth image as claimed in claim 1, wherein the step 5 comprises the following steps:
step 5.1, normally acquiring a field real-time operation image, clearly and visually judging the position of each sieve plate through a server interface, obtaining the coordinate of the corresponding sieve plate in the image by selecting the position of each sieve plate in the image, and selecting an ROI (region of interest) in the original depth map according to the coordinate;
step 5.2, after obtaining each area needing to be judged, obtaining an original depth image, selecting an area corresponding to the sieve plate in the original image, counting pixels in the area one by one, calculating the average depth value of the pixels in the area, comparing the average depth value with a normal depth threshold value, and judging that the sieve plate has faults at the moment when the average depth is greater than the depth threshold value; and calculating the difference d between the average depth value and the depth threshold value when d is larger than or equal to the thread1Judging that the sieve plate has a falling fault at the moment; when thread2≥d≥thread1Judging that the screen plate has a tilt fault at the moment and judging the threshold value of the screen plate1And thread2Setting by combining data and experience collected by a camera;
and 5.3, judging the working condition of each screen plate through the steps, identifying the normal screen plate on a screen plate simulation interface on the server through a gray color block, identifying the fault screen plate on the screen plate simulation interface through a red color block, and printing a corresponding log on the server interface.
7. The intelligent screen plate fault monitoring method based on the depth image as claimed in claim 1, wherein the step 6 comprises the following steps:
step 6.1, the system server sends an instruction, and after the instruction for acquiring the depth data is sent for 5 times, the instruction for acquiring the signal intensity is sent once;
6.2, after the signal intensity data are obtained, analyzing the original data according to the method in the step 3 to obtain the signal intensity data corresponding to each pixel point;
6.3, selecting a central area of the sieve plate, comparing the signal intensity in the central area with the signal intensity in normal operation, namely under the condition that coal exists on the sieve surface, and if the signal intensity fluctuates up and down in a normal range, considering that the system operates normally at the moment, wherein the fluctuation of the signal intensity may be caused by the thickness change of the coal bed on the surface of the sieve surface; if the signal intensity is obviously increased, the situation that no coal exists on the screen surface is shown;
6.4, judging whether the screen surface has coal or not through the steps, wherein the condition that the system runs normally when the screen surface has the coal indicates that the system runs normally is judged, and the server interface is marked through a gray color block; if no coal is on the screen surface, the system idles at the moment, or the screen surface has no coal due to the blockage of a raw coal hopper at the upstream of the screen plate, and the screen surface is marked by a red color block at the moment so as to prompt a user.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911186854.2A CN110926737B (en) | 2019-11-28 | 2019-11-28 | Intelligent screen plate fault monitoring method based on depth image |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911186854.2A CN110926737B (en) | 2019-11-28 | 2019-11-28 | Intelligent screen plate fault monitoring method based on depth image |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110926737A CN110926737A (en) | 2020-03-27 |
| CN110926737B true CN110926737B (en) | 2021-06-04 |
Family
ID=69847235
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201911186854.2A Active CN110926737B (en) | 2019-11-28 | 2019-11-28 | Intelligent screen plate fault monitoring method based on depth image |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110926737B (en) |
Family Cites Families (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| KR20060084085A (en) * | 2005-01-17 | 2006-07-24 | 주식회사 청석기업 | A device for screen which using sort out circulating aggregate |
| DE102016013406B4 (en) * | 2016-11-11 | 2022-02-03 | Schenck Process Europe Gmbh | Method of operating a vibrating machine condition monitoring system and condition monitoring system |
| DE102017009373B3 (en) * | 2017-10-10 | 2019-05-16 | Schenck Process Europe Gmbh | Mobile device for detecting the state and operating parameters of vibrating machines, equipped vibrating machine and method for detecting the operating and state parameters of vibrating machines |
| CN109905691A (en) * | 2017-12-08 | 2019-06-18 | 浙江舜宇智能光学技术有限公司 | Depth image acquisition device and depth image acquisition system and its image processing method |
| CN108805868B (en) * | 2018-05-25 | 2020-07-31 | 哈尔滨市科佳通用机电股份有限公司 | Image processing method and fault detection method for fault detection of running gear equipment under electric vehicle carrying vehicle |
| CN110503646B (en) * | 2019-08-29 | 2022-03-25 | 联想(北京)有限公司 | Image processing method and device |
-
2019
- 2019-11-28 CN CN201911186854.2A patent/CN110926737B/en active Active
Also Published As
| Publication number | Publication date |
|---|---|
| CN110926737A (en) | 2020-03-27 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110261436B (en) | Rail fault detection method and system based on infrared thermal imaging and computer vision | |
| CN104441526B (en) | Mould in-service monitoring protection system and method based on appearance profile coupling | |
| US7408570B2 (en) | Web inspection system | |
| CN1882078A (en) | Image error detection device for monitoring camera | |
| US20080317356A1 (en) | Image monitoring system | |
| CN110458157B (en) | Intelligent monitoring system for power cable production process | |
| JP5168215B2 (en) | Appearance inspection device | |
| CN112082781B (en) | Vehicle and fault detection method and fault detection device thereof | |
| CN106056597A (en) | Object visual detection method and device | |
| CN112784821A (en) | Building site behavior safety detection and identification method and system based on YOLOv5 | |
| CN106454330A (en) | Fuzziness anomaly detection method for video signals | |
| CN110926737B (en) | Intelligent screen plate fault monitoring method based on depth image | |
| KR20230036557A (en) | A system for monitoring solar panel failures on a large scale using drones | |
| CN112422818B (en) | Intelligent screen dropping remote detection method based on multivariate image fusion | |
| US20220281177A1 (en) | Ai-powered autonomous 3d printer | |
| CN116385948B (en) | System and method for early warning railway side slope abnormality | |
| CN101609580A (en) | Method for intelligently protecting bank self-service equipment | |
| CN116840240A (en) | Visual detection system of power supply distributor | |
| WO2023048017A1 (en) | Water quality monitoring system | |
| CN116704440A (en) | Intelligent comprehensive acquisition and analysis system based on big data | |
| CN115684853A (en) | Unmanned aerial vehicle power transmission line fault detection method and system with ultraviolet imager | |
| CN115205193A (en) | Steel plate surface defect detection method based on microdefect YOLO network | |
| KR20070066829A (en) | Diagnostic system for gel dispensing condition and method thereof | |
| CN108956617A (en) | A kind of detection method of the electronic device solder joint thermal cycle failure based on micro- deformation Intelligence Classifier | |
| CN111222428B (en) | Detection method for device fastening operation based on monitoring video |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |