A method for capturing key frames at the tail of a partition-triggered sintering machine

Abstract

The present invention discloses a method for capturing key frames at the tail of a partition-triggered sintering machine, belonging to the technical field of image capture. The capture method of the present invention includes the following steps: Step 1, collect the video of the cross-section at the tail of the sintering machine; Step 2, manually divide the collected video window into two functional areas; Step 3, monitor the average temperature T r1 in the R1 area to determine whether the sintered ore on the previous trolley is about to fall; Step 4, if the sintered ore on the previous trolley is about to fall, send an instruction to trigger the temperature monitoring of the R2 area; Step 5, monitor the average temperature T r2 in the R2 area to determine whether the dust raised by the falling of the sintered ore is about to block the line of sight; Step 6, if the dust raised by the falling of the sintered ore is about to block the line of sight, then send an instruction to trigger the capture of key frames. The present invention only needs to judge whether the average temperature in the designated area is within the set threshold, with a simpler logic, a smaller amount of data processing, and a fast processing speed.

Description

Technical Field

[0001] The present invention belongs to the technical field of image capture, and more specifically, relates to a method for capturing key frames of the tail of a sintering machine with partition triggering. Background Art

[0002] The cross-section of the tail of a sintering machine contains rich details, such as the brightness, area, uniformity, and pore distribution of the red-hot layer. These details directly and real-time reflect the control situation of the sintering process and the quality of sintered ore. Establishing the connection between the quality of sintered ore and the cross-section of the tail can achieve real-time prediction of the quality of sintered ore, providing an important basis for the precise control of the sintering process. However, in the actual sintering process, the sintering machine is constantly in motion, and the working environment has a lot of dust interference. Capturing a key frame image with clear picture and complete information from the dynamic sintering video is one of the key problems faced by the development of the sintered ore quality prediction technology based on the cross-section of the tail.

[0003] In order to be able to observe the cross-sectional image of the tail of the sintering machine completely, the observer and the cross-section of the tail of the sintering machine should be on the same horizontal plane. The black ore layer in the previous trolley that is rotating and descending blocks the sintering cross-section in the subsequent trolley that is on the same plane as the observer. As the previous trolley slowly descends, the sintered ore cross-sectional image on the subsequent trolley gradually becomes visible. When the sintered ore on the rotating previous trolley completely falls due to the action of gravity, the cross-sectional image of the tail of the sintering machine in the subsequent trolley can be observed completely. After that, the fallen sintered ore impacts the ground, generating a large amount of sintering tail gas and dust. Under the action of the induced draft fan, these tail gas and dust rise rapidly, making the sintering cross-sectional image blurred. Therefore, the best timing for obtaining the cross-section is the short period when the sintered ore on the previous trolley just falls and the tail gas and dust have not been completely lifted.

[0004] The traditional best cross-section acquisition methods include the fixed time interval method and the external contact method. The fixed time interval method captures the best sintering machine tail cross-section in the video by determining the time interval between the sintered ore drops of two adjacent trolleys. However, in reality, at the moment when the sintered ore on the sintering trolley drops, accompanied by complex physical and chemical changes, the sintered ore on the sintering trolley does not completely drop all at once. Due to different sintering conditions, the dropping amplitude and speed of the sintered ore are different, and other sintering parameters such as trolley speed and material layer thickness are adjusted in real time with different sintering conditions. Therefore, the time interval changes. These factors lead to a large error in the fixed time interval method. The external contact method installs sensors on the track where the trolley runs. When the trolley runs to the specified position, the sensor will be triggered. The sensor emits a pulse signal and transmits it to the computer via a serial bus, and the computer captures the cross-section image of the machine tail at that moment. External triggering can capture the sintering machine tail cross-section more accurately. However, the dropping of the sintered ore at the machine tail is a process affected by various factors. The viscosity of the sintered ore at different times is different, resulting in different frictional forces between it and the grate of the trolley. Even if the captured position is the same, the dropping situation of the cross-section is not exactly the same, resulting in poor quality of the captured cross-section image. In addition, in this method, the sensors work in a high-temperature and dusty environment for a long time, with a high component loss rate and increased maintenance costs.

[0005] After retrieval, the application case with the Chinese patent application number 2021110757770 discloses a method and system for obtaining the best cross-section image of the sintering machine tail. It determines the preset area where the sintering machine tail cross-section is located, and uses an image acquisition device and a temperature measurement device to continuously collect real-time image information of the preset area. The image information includes the average temperature and acquisition time of the preset area. Using the historical trend curve of the average temperature of the preset area over the acquisition time, the best acquisition time point is determined, and then the corresponding best cross-section image is extracted according to the best acquisition time point. However, this application case needs to draw the historical change curve of the average temperature in the preset area over time to analyze and determine the best cross-section acquisition time, with a relatively large amount of data processing. And it is necessary to perform image acquisition and calculation at a high frequency to avoid missing key frames.

[0006] For another example, the application case with the Chinese patent application number 2020113151560 discloses a method and device for capturing pictures of the cross-section at the tail of a sintering machine. The method includes: determining a preset area where the cross-section at the tail of the sintering machine is located; continuously obtaining N cross-section pictures of the cross-section at the tail of the sintering machine through an infrared thermal imaging device; analyzing the temperature of each pixel of each cross-section picture, and selecting P0 cross-section pictures that meet the preset conditions from the N cross-section pictures. The preset conditions are that the sintering trolley is not in the tipping state, the cross-section at the tail of the sintering machine is not blocked by high-temperature dust generated by the sintered ore falling to the ground, the lower edge of the cross-section at the tail of the sintering machine is detected, it is not blocked by the previous unshed sintered ore layer, and the lower part is not blocked by the broken blocks falling from the fault. Analyze the temperature of each pixel of each of the P0 cross-section pictures, obtain the target picture therefrom, and obtain the information on the sintering quality according to the clear and complete picture, so as to provide guidance for the operation of the sintering process. However, this application case requires refined calculation and analysis of each picture to obtain multiple calculated values such as the first average temperature, the second average temperature, the third average temperature, the fourth average temperature, the first parameter, the second parameter... of each picture, and analyze the relationship between these calculated values, resulting in a large amount of calculation, low efficiency, and slow processing speed. Summary of the Invention

[0007] 1. Problems to be Solved

[0008] The purpose of the present invention is to provide a method for capturing key frames at the tail of a partition-triggered sintering machine, so as to overcome the deficiencies of the existing capturing methods, such as a large amount of data processing, low capturing efficiency, and low capturing accuracy.

[0009] 2. Technical Solutions

[0010] To solve the above problems, the technical solutions adopted by the present invention are as follows:

[0011] A method for capturing key frames at the tail of a partition-triggered sintering machine according to the present invention includes the following steps:

[0012] Step 1: Video acquisition

[0013] Collect the video of the cross-section at the tail of the sintering machine;

[0014] Step 2: Region division

[0015] Manually divide the video window collected in Step 1 into two functional regions, where the task of the R1 region is to capture the state where the next trolley starts to dump materials, and the task of the R2 region is to capture the complete and effective cross-section at the tail;

[0016] Step 3: Monitor the average temperature T of the R1 region r1 to determine whether the sintered ore of the previous trolley is about to fall;

[0017] Step 4: If the sintered ore on the previous trolley is about to fall, send an instruction to trigger the temperature monitoring of Area R2;

[0018] Step 5: Monitor the average temperature T of Area R2 r2 to determine whether the dust raised by the falling sintered ore is about to block the line of sight;

[0019] Step 6: If the dust raised by the falling sintered ore is about to block the line of sight, send an instruction to trigger the capture of key frames.

[0020] The present invention is mainly based on the change law of the temperature at the tail section of the sintering machine during the falling process of the sintered ore. The video of the tail section of the sintering machine is divided into two functional areas, and the temperature changes in the two functional areas are monitored in real time. According to the monitored temperature values, the timing when the dust raised by the falling sintered ore is about to block the line of sight is judged, that is, the best timing for capturing key frames.

[0021] Furthermore, when the average temperature T of Area R1 r1 reaches the peak value, it is judged that the sintered ore on the previous trolley is about to fall, and the sintered ore on the next trolley is about to enter the wireframe Area R2. Specifically, when 337°C ≤ T r1 ≤ 343°C, the average temperature T of Area R1 r1 reaches the peak value, and at this time, an instruction is sent to trigger the temperature monitoring of Area R2.

[0022] Furthermore, when the average temperature T of Area R2 r2 reaches the peak value, it is judged that the dust raised by the falling sintered ore is about to block the line of sight. Specifically, when 549°C ≤ T r2 ≤ 558°C, the average temperature T of Area R2 r2 reaches the peak value, and at this time, an instruction is sent to trigger the capture of key frames.

[0023] As Figure 2 and Figure 3 shown, the inventors of the present application found during the research process that as the sintered ore continuously falls, the temperatures of Areas R1 and R2 show periodic changes. When the sintered ore on the previous trolley starts to fall, the average temperature of Area R1 just reaches the peak value. And when the smoke and dust raised by the falling sintered ore blocks the lens a moment before, the average temperature of Area R2 just reaches the peak value, and this is the best timing for capturing key frames. Therefore, based on this law, the present invention captures key frames by identifying the occurrence time of the average temperature peak values of Areas R1 and R2.

[0024] To accurately identify and capture the temperature peak values, the inventors experimentally selected 20,000 frames of the R1 and R2 wireframe areas with complete exposed cross-sections, and statistically analyzed the peak values of the average temperatures of a total of 47 cycles. The statistical results show that the average temperature T of the R1 wireframe arear1 The peak value fluctuates slightly between 337°C and 343°C, and the average temperature T of the R2 wireframe area r2 The peak value fluctuates slightly between 549°C and 553°C. Therefore, the present invention utilizes these two temperature thresholds to capture the temperature peak value. Since the present invention only needs to determine whether the average temperature within the defined area is within the set threshold, the logic is simpler, the data processing volume is smaller, and the processing speed is fast, reducing the memory occupancy.

[0025] Meanwhile, the video of the sintering machine tail section is divided into two functional areas. The R1 area is responsible for determining the timing when the sintered ore of the previous trolley breaks and falls. Only after the average temperature of the R1 area reaches the set threshold, the R2 area starts the capture calculation of key frames, with higher capture accuracy. Moreover, when collecting in the R1 area, the collection frequency can be greatly reduced, significantly saving the calculation amount, and the system processing is more efficient and fast.

[0026] Furthermore, in step one, an infrared video of the sintering machine tail section is obtained through an infrared imager, and temperature information is automatically generated.

[0027] Furthermore, in step two, the collected video window is divided into two upper and lower wireframe areas. The size and position of the wireframe R1 are consistent with the non-tilted sintered ore section, and the wireframe R2 is below and adjacent to the wireframe R1, with a similar size.

[0028] Furthermore, when the capture instruction of the key frame is triggered, a picture is automatically intercepted in the video window as the key frame for saving, for subsequent feature analysis.

[0029] 3. Beneficial effects

[0030] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0031] (1) For the method for capturing key frames at the sintering machine tail with partition triggering of the present invention, by dividing the tail section image into two specific functional areas, R1 and R2, and performing partition temperature monitoring, based on the temperature changes in the two specific functional areas of R1 and R2, the timing when the dust raised by the falling of the sintered ore is about to block the line of sight is judged, so as to determine the capture timing of the key frame, greatly improving the capture accuracy of the key frame. Moreover, the overall operation of this method is simple, the calculation amount is small, and the capture efficiency is relatively high.

[0032] (2) For the method for capturing key frames at the sintering machine tail with partition triggering of the present invention, when the average temperature T of the R1 area r1 reaches the peak value, that is, when 337°C ≤ T r1 ≤ 343°C, an instruction is sent to trigger the temperature monitoring of the R2 area; and when the average temperature T of the R2 area r2 reaches the peak value, that is, when 549°C ≤ T r2When the temperature is ≤558°C, send an instruction to trigger the capture of key frames. It is easy to determine the capture timing of key frames, that is, it can be done in one step. After triggering the capture condition, directly capture the key frame image. The logic is relatively simple, direct and fast, with higher efficiency, strong industrial operability, and more accurate control of the capture timing. Description of the Drawings

[0033] Figure 1 It is a schematic diagram of the infrared video partition at the tail of the sintering machine;

[0034] Figure 2 It is a graph of the average temperature change in area R1;

[0035] Figure 3 It is a graph of the average temperature change in area R2;

[0036] Figure 4 It is a flowchart of the algorithm for the partition trigger method of the present invention;

[0037] Figure 5 It is an image of the cross-section of the machine tail captured by using the partition trigger method of the present invention. Detailed Embodiments

[0038] The present invention will be further described below in conjunction with specific embodiments.

[0039] Embodiment

[0040] According to the temperature change law of the cross-section of the machine tail during the dropping process of sinter, a partition trigger capture method is proposed, and its algorithm logic is as Figure 4 shown. This method divides the video area of the cross-section of the machine tail into two functional areas, R1 and R2. First, monitor the temperature through area R1. When T r1 is in the range of 337°C to 343°C, it indicates that the sinter is about to drop. At this time, trigger the temperature monitoring of area R2. When T r2 is between 549°C and 553°C, it indicates that the dust raised by the dropping of the sinter is about to block the line of sight. At this time, trigger the capture of key frames. The specific steps are as follows:

[0041] (1) Obtain the infrared video of the cross-section of the machine tail through an infrared imager installed at the tail of the sintering machine, and automatically generate temperature information;

[0042] (2) Manually divide the collected video window into two wireframes for triggering the key frame capture function (as Figure 1 shown). The size and position of wireframe R1 are the same as those of the non-tilted sinter cross-section. Wireframe R2 is adjacent to it below wireframe R1, and its size also corresponds to the size of wireframe R1;

[0043] (3) Monitor the average temperature T r1 of area R1 of the wireframe;

[0044] (4) Judge T r1 Temperature range. When 337°C ≤ T r1 ≤ 343°C, judge that the sinter on the previous trolley is about to fall, and the sinter on the next trolley will enter the wireframe R2. At this time, send a command to trigger the temperature monitoring of the R2 area;

[0045] (5) Monitor the average temperature T of the wireframe R2 area r2 ;

[0046] (6) Judge T r2 Temperature range. When 549°C ≤ T r2 ≤ 558°C, judge that the dust raised by the falling sinter is about to block the line of sight. At this time, the information of the tail section of the machine is maximally exposed and the picture is clear. Send a command to trigger the capture of key frames;

[0047] (7) Automatically capture pictures in the video window and save them as key frames for subsequent feature analysis.

[0048] This partition trigger capture method has been successfully tested on a sintering machine in a steel plant and achieved good results. The captured key frame images are clear and have comprehensive details (such as Figure 5 ), laying a good foundation for the subsequent quality prediction of sinter.

Claims

1. A method for capturing key frames at the tail of a partition-triggered sintering machine, characterized in that, It includes the following steps: Step 1, video acquisition Acquire the video of the cross-section at the tail of the sintering machine; Step 2, area division Manually divide the video window acquired in Step 1 into two functional areas. The task of area R1 is to capture the state when the next trolley starts to dump materials, and the task of area R2 is to capture the complete and effective cross-section of the machine tail; Step 3. Monitor the average temperature T of area R1 r1 to determine whether the sinter on the previous trolley is about to fall; Step 4, if the sintered ore of the previous trolley is about to fall, send an instruction to trigger the temperature monitoring of area R2; Step 5: Monitor the average temperature T in area R2 r2 to determine whether the dust raised by the falling sinter will soon block the line of sight; Step 6, if the dust raised by the falling of the sintered ore is about to block the line of sight, send an instruction to trigger the capture of key frames.

2. The capture method of the key frame at the tail of a partition-triggered sintering machine according to claim 1, characterized in that: When the average temperature T of the R1 area r1 reaches the peak value, it is determined that the sinter on the previous trolley is about to fall, and the sinter on the next trolley is about to enter the wire frame R2 area.

3. A method for capturing key frames at the tail of a partition-triggered sintering machine according to claim 2, characterized in that: When 337°C ≤ T r1 ≤ 343°C, then send an instruction to trigger the temperature monitoring of area R2.

4. The capture method of key frames at the tail of a partition-triggered sintering machine according to claim 1, characterized in that: When the average temperature T in the R2 area r2 reaches the peak value, it is determined that the dust raised by the falling of the sinter will soon block the line of sight.

5. A method for capturing key frames at the tail of a partition-triggered sintering machine according to claim 4, characterized in that: When 549°C ≤ T r2 ≤ 558°C, a command is sent at this time to trigger the capture of key frames.

6. A method for capturing key frames at the tail of a partition-triggered sintering machine according to any one of claims 1-5, characterized in that: In Step 1, an infrared video of the cross-section at the tail of the sintering machine is obtained through an infrared imager, and temperature information is automatically generated.

7. A method for capturing key frames at the tail of a partition-triggered sintering machine according to any one of claims 1-5, characterized in that: In Step 2, the acquired video window is divided into two wireframe areas, the upper and the lower. The size and position of wireframe R1 are consistent with the cross-section of the non-tilted sintered ore, and wireframe R2 is below and adjacent to wireframe R1.

8. A method for capturing key frames at the tail of a partition-triggered sintering machine according to any one of claims 1-5, characterized in that: When the capture instruction of the key frame is triggered, a picture is automatically intercepted in the video window and saved as a key frame for subsequent feature analysis.

Images