Diesel burner combustion performance detection method based on machine vision

Abstract

The embodiment of the invention discloses a diesel burner combustion performance detection method based on machine vision. The method comprises the steps that a combustion flame video image of a to-be-detected diesel burner is acquired; judging the flame type according to the brightness change characteristics of each frame of flame combustion image, and determining each frame of target grayscale image used for segmenting the image processing area; determining a processing area of each frame of image according to brightness differences between a flame combustion area and a background area and between the flame combustion area and a combustor area in the target gray level image; for the image processing area, enhancing the target flame area by enhancing the image contrast, and realizing accurate binaryzation of the target flame area in each frame of image; and color feature and morphological feature parameter detection is carried out on each frame of target flame area so as to determine the combustion sufficiency and combustion stability of the diesel combustor. According to the embodiment, the combustion sufficiency and the combustion stability of the diesel combustor can be automatically detected and judged, and the detection accuracy is high.

Description

Technical Field

[0001] The embodiments of the present invention relate to the technical field of machine vision, and in particular, to a method for detecting the combustion performance of a diesel burner based on machine vision. Background Art

[0002] Diesel burners are widely used in various energy-consuming industries such as industrial boilers, drying equipment, ironing equipment, field cooking equipment, and kitchen heating equipment. Testing their combustion performance is very important. Testing and analyzing the performance characteristics of existing diesel burners such as ignition characteristics, combustion sufficiency, and combustion stability can provide experimental basis and data support for the subsequent design, research, and application of diesel burners.

[0003] However, traditional testing methods usually rely on manual visual inspection of the flame state to judge, or use corresponding instrument equipment for simple test analysis or indirect experimental testing. There are problems such as being greatly affected by environmental or subjective factors, low test accuracy, and complex test processes, and there are also certain test safety hazards. Summary of the Invention

[0004] The embodiments of the present invention provide a method for detecting the combustion performance of a diesel burner based on machine vision to solve the above technical problems.

[0005] In a first aspect, the embodiments of the present invention provide a method for detecting the combustion performance of a diesel burner based on machine vision, including:

[0006] Obtaining a combustion flame video image of the diesel burner to be detected;

[0007] Judging the flame type according to the brightness change characteristics of each frame of the flame combustion image, and determining the target grayscale image for segmenting the image processing area for each frame;

[0008] Determining the processing area of each frame of image according to the brightness difference between the flame combustion area and the background area and the burner area in the target image;

[0009] For the image processing area, enhancing the target flame area by enhancing the image contrast to achieve accurate binarization of the target flame area in each frame of image;

[0010] Detecting color characteristics and morphological characteristic parameters for each frame of the target flame area to determine the combustion sufficiency and combustion stability of the diesel burner.

[0011] In a second aspect, the embodiments of the present invention provide an electronic device, including:

[0012] One or more processors;

[0013] A memory for storing one or more programs,

[0014] When the one or more programs are executed by the one or more processors, the one or more processors implement the above-mentioned method for detecting the combustion performance of a passive diesel burner based on machine vision.

[0015] In summary, this embodiment provides a method for detecting the combustion performance of a diesel burner based on machine vision. First, a combustion flame video image of the diesel burner to be detected is obtained; the flame type is judged according to the brightness change characteristics of each frame of the flame combustion image, and the target grayscale image for segmenting the image processing area is determined for each frame; according to the brightness differences between the flame combustion area, the background area, and the burner area in the target grayscale image, the processing area of each frame of the image is determined; for the image processing area, the target flame area is enhanced by enhancing the image contrast to achieve precise binarization of the target flame area in each frame of the image; color feature and morphological feature parameters of each frame of the target flame area are detected to determine the combustion sufficiency and combustion stability of the diesel burner. The whole method is automatically realized by machine vision technology, avoiding the subjectivity of manual visual inspection of the flame state, and also eliminating the need to use instrument equipment for testing. It has high detection accuracy, a simple process, and no safety hazards. Description of the Drawings

[0016] In order to more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the following will briefly introduce the drawings required for use in the description of the specific embodiments or the prior art. Obviously, the following drawings are some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0017] Figure 1 is a flowchart of a method for detecting the combustion performance of a diesel burner based on machine vision provided by an embodiment of the present invention;

[0018] Figure 2 is a flowchart of another method for detecting the combustion performance of a diesel burner based on machine vision provided by an embodiment of the present invention;

[0019] Figure 3 is a schematic diagram of the RGB pixel distribution on the cross-section of different color feature regions on a flame combustion image provided by an embodiment of the present invention;

[0020] Figure 4 is a schematic diagram of the binarization of a flame combustion image provided by an embodiment of the present invention, where Figure 4 (a) is the original color image of the flame combustion, Figure 4 (b) is the weighted average grayscale flame image, Figure 4 (c) is the Otsu binarization image;

[0021] Figure 5 It is a schematic diagram of binarization after strengthening the target flame area provided by an embodiment of the present invention. Among them, Figure 5 (a) is the flame grayscale image after strengthening the target flame area, Figure 5 (b) is Figure 5 the Otsu binarization image corresponding to (a);

[0022] Figure 6 It is an RGB pixel distribution diagram on the cross-section of different color feature areas on another flame combustion image provided by an embodiment of the present invention;

[0023] Figure 7 It is a schematic diagram of binarization of another flame combustion image provided by an embodiment of the present invention. Among them, Figure 7 (a) is the original flame combustion color image, Figure 7 (b) is the weighted average grayscale flame image, Figure 7 (c) is the Otsu binarization image;

[0024] Figure 8 It is a schematic diagram of binarization after strengthening the target flame area provided by another embodiment of the present invention. Among them, Figure 8 (a) is the flame grayscale image after strengthening the target flame area; Figure 8 (b) is Figure 8 the Otsu binarization image corresponding to (a);

[0025] Figure 9 It is a schematic diagram of the pixel area curve of the flame area in continuous image frames, the flame height curve in continuous image frames, the area difference curve of the flame area in continuous image frames, and their corresponding average values and standard deviations provided by an embodiment of the present invention. Among them, Figure 9 (a) is the corresponding curve and its average value and standard deviation when preheating is completed, Figure 9 (b) is the corresponding curve and its average value and standard deviation immediately after firing, Figure 9 (c) is the corresponding curve and its average value and standard deviation when combustion is stable, Figure 9 (d) is the corresponding curve and its average value and standard deviation when the firepower overshoot is unstable;

[0026] Figure 10 It is a schematic diagram of the structure of an electronic device provided by an embodiment of the present invention. Detailed implementation manners

[0027] To make the objectives, technical solutions, and advantages of the present invention clearer, the technical solutions of the present invention will be described clearly and completely below. Apparently, the described embodiments are only a part rather than all of the embodiments of the present invention. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the scope protected by the present invention.

[0028] In the description of the present invention, it should be noted that the orientation or positional relationship indicated by the terms "center", "upper", "lower", "left", "right", "vertical", "horizontal", "inner", "outer", etc. is based on the orientation or positional relationship shown in the drawings, and is only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore should not be construed as a limitation of the present invention. In addition, the terms "first", "second", and "third" are only used for descriptive purposes and should not be construed as indicating or implying relative importance.

[0029] In the description of the present invention, it should also be noted that unless otherwise clearly specified and limited, the terms "installed", "connected", and "connected" should be understood in a broad sense. For example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be directly connected or indirectly connected through an intermediate medium, and it can be the communication inside two elements. For those of ordinary skill in the art, the specific meanings of the above terms in the present invention can be understood according to specific situations.

[0030] Figure 1 is a flowchart of a method for detecting the combustion performance of a diesel burner based on machine vision provided by an embodiment of the present invention. This method is executed by an electronic device, such as Figure 1 shown, and this method specifically includes:

[0031] S110. Obtain a combustion flame video image of the diesel burner to be detected.

[0032] In general, the combustion flame of a diesel burner will present different colors and forms at different stages. For example, ① the flame is not fully burned in the early stage, and the flame is a yellow-black state mixed with yellow flames and black smoke. The flame is soft and long, the fluctuation amplitude is large, and the propagation speed is relatively slow; ② With the increase of combustion temperature, especially after entering the stable combustion stage, from the perspective of visible light, the flame appears bright white, blue or even colorless, the flame is short, the fluctuation amplitude is tightened, and the propagation speed is accelerated; ③ When it turns from the stable combustion stage to the state of continuous incomplete combustion with overshoot of firepower, the flame appears thick yellow and black mixed color, and when it turns from the stable combustion stage to the intermittent deflagration state (that is, most of the time it is in a stable combustion state, and intermittent momentary flame deflagration and jumping state), the flame appears dazzling bright yellow. In the above two unstable combustion states, the flames show dynamic characteristics such as large fluctuation area and fast propagation speed.

[0033] Based on the above flame characteristics, this embodiment captures visible light video images of the flame combustion of the diesel burner at various key time periods, such as when preheating is completed, immediately after ignition, when combustion is stable, when firepower is over-regulated or deflagration is unstable, as a basis for subsequent judgment of the combustion performance of the diesel burner.

[0034] S120, judging the flame type according to the brightness change characteristics of each frame of the flame combustion image, and determining the target grayscale image of each frame for segmenting the image processing area.

[0035] The image processing area here refers to a part of the area that can cover the flame combustion area segmented from the flame combustion image. By segmenting the image processing area, the scope of subsequent image processing can be narrowed, and the accuracy and efficiency of image processing can be improved. In this step, according to different flame types, different methods are used to determine the grayscale image used to segment the image processing area. For the convenience of distinction and description, this embodiment refers to the target grayscale image.

[0036] In a specific implementation, each frame of color image acquired in S110 is first grayed. Specifically, considering that when the burning flame is small, the brightness of the background area and the flame burning area are both high, while the brightness of the burner area is low. When the flame area is binary segmented, the background area is easily mis-segmented as the flame area. Therefore, for the small flame burning image, the background difference method can be used to gray the color image to eliminate the influence of the background area. When the burning flame is large, the brightness of the flame burning area is high, while the brightness of the background area and the burner area is relatively low. At this time, the weighted average method is directly used to gray the flame burning image. The specific steps are as follows:

[0037] Read the color background image and the current frame flame combustion image. Among them, the background image refers to the image before the diesel burner is ignited. Optionally, to improve the image processing efficiency, the pixel size of each image is reduced by 0.3 times. Then, convert the background image and the current frame flame combustion image into three grayscale images of R, G, and B respectively, denoted as R b 、G b 、B b and R f 、G f 、B f .

[0038] By using the background difference method, obtain the R-frame background difference map of the current frame flame combustion image. The calculation formula is shown in Equation (1); for the R-frame background difference map, calculate the gray average value standard deviation D of its image pixels in sequence. The calculation formulas are shown in Equation (2) and Equation (3) respectively:

[0039] f(i,j) = R f (i,j) - R b (i,j) (1)

[0040]

[0041]

[0042] In the formula, R f (i,j), R b (i,j) - The R components of the pixel points at (i,j) of the original color flame combustion image and the background image respectively;

[0043] f(i,j) - The gray value of the pixel point at (i,j) of the R-frame background difference map;

[0044] x size - The pixel width of the image;

[0045] y size - The pixel height of the image;

[0046] - The gray average value of all pixels in the R-frame background difference map;

[0047] D - The gray standard deviation of all pixels in the R-frame background difference map.

[0048] Then, according to the pixel standard deviation of the R-frame background difference map, judge the flame type. If D < Dth2 (Dth2 is the no-flame judgment threshold) is satisfied, it is judged that the current frame flame combustion image is a no-flame image. For the no-flame image, the steps of S130 - S150 will no longer be executed, but another method will be used to judge the combustion state, such as Figure 2As shown. If D < Dth1 (the small flame judgment threshold), it is determined that the current frame of the flame combustion image is a small flame image, and the grayscale flame image is obtained based on the R-frame background difference method and used as the target image for determining the image processing area. If the above two conditions are not met, it belongs to a non-small flame image. At this time, the background image and the R-frame grayscale image of the current frame of the flame combustion image are used as the target images for determining the image processing area. Among them, Dth1 > Dth2. For the values of the thresholds Dth1 and Dth2, under the test conditions of this experiment, they can be taken as 35 and 10 respectively.

[0049] S130. Determine the processing area of each frame of the image according to the brightness difference between the flame combustion area, the background area, and the burner area in the target grayscale image.

[0050] For the above-mentioned target image, according to the brightness difference between the flame combustion area, the background area, and the burner area, by finding the brightness jump dividing lines in the horizontal and vertical directions, the image processing area is determined, further narrowing the image processing range and improving the processing efficiency.

[0051] In a specific implementation manner, the four-side dividing lines can be determined respectively in the following way:

[0052] ① Determine the lower dividing line of the image processing area

[0053] If the target image is the R-frame background difference map of the aforementioned small flame image, set the lower boundary of the target image as the lower boundary line of the image processing area (i.e., y = y size -1).

[0054] If the target image is the aforementioned R-frame grayscale image, determine the lower boundary line of the image processing area by finding the jump point of the horizontal cumulative pixel value distribution curve of the background R-frame grayscale image. Optionally, the specific determination method is as follows:

[0055] (a) For the R-frame grayscale image of the background image, calculate and obtain its horizontal cumulative pixel value distribution curve (denoted as L h ), and the calculation formula is shown in Equation (4):

[0056]

[0057] (b) After smoothing the curve L h , calculate and obtain the cumulative value difference curve between consecutive points on the curve (denoted as L1), and the calculation formula is shown in Equation (5), where let L1(0) = L1(1):

[0058] L1(j) = L h (j) - L h (j - 1), j ∈ [1, y size -1] (5)

[0059] (c) Perform Savitzky-Golay filtering on curve L1 and draw a horizontal line L on L1 av1 , and its value is the sum of the average value of the vertical coordinates of L1 (denoted as L av1 ) and twice the standard deviation (denoted as L std1 ), that is, its value is L av1 + 2L std1 . The calculation formula is shown in Equations (6) and (7):

[0060]

[0061]

[0062] Find the left jump point on curve L1 that crosses the horizontal line L av1 (denote its abscissa value as P d , that is, its pixel row coordinate in the image is P d ).

[0063] (d) Determine the lower boundary line of the image processing area as y = P d .

[0064] ② Determine the upper boundary line of the image processing area

[0065] For the target image (R-frame background difference map or flame R-frame grayscale map), calculate and obtain its vertical cumulative pixel value distribution curve (denoted as L v ). The calculation formula is shown in Equation (8):

[0066]

[0067] After smoothing it, calculate the pixel average value of the column where the second largest peak point on curve L v is located (denoted as reference value P x ). Find the uppermost pixel point in the target image whose pixel value is equal to P x (denote its pixel row coordinate in the image as P u ), and accordingly determine the upper boundary line of the image processing area as y = P u .

[0068] ③ Determine the left and right boundary lines of the image processing area

[0069] For the smoothed curve L v above, obtain the cumulative value difference curve between consecutive points (denoted as L2, and its calculation method is the same as Equation (5)). Similarly, perform Savitzky-Golay filtering on it, and draw a horizontal line of its ordinate average value on L2 (denote this average value as L av2, the calculation formula is the same as formula (6)). Find the jump points on both sides of the horizontal line L that cross the curve L2 (record their abscissa values as P av2 and P l , that is, their corresponding image pixel column coordinates are P r and P l ). Based on this, determine that the left and right dividing lines of the image processing area are: x = P r x = P l and x = P r .

[0070] Determine the image processing area according to the upper, lower, left, and right dividing lines, and all subsequent image processing steps are carried out within this area.

[0071] S140. For the image processing area, enhance the target flame area by enhancing the image contrast to achieve accurate binarization of the target flame area in each frame of the image.

[0072] For small flame images, considering that both the background area and the flame combustion area have high brightness, while the burner area has low brightness, when binarizing and segmenting the flame area, the background area is easily mis-segmented as the flame area. Therefore, the background difference method is adopted to grayscale the color image to eliminate the influence of the background area, and then based on the Otsu method for binarization, the target flame area is segmented and extracted.

[0073] For non-small flame images, the brightness and color of the flame combustion directly affect the brightness and color of the background area, thus affecting the distinguishability between the two, and further affecting the binarization segmentation of the flame area. Analyze the brightness and color conditions of the flame and the background area during the combustion process, which can be mainly divided into the following three situations. The binarization algorithms for them are given respectively below:

[0074] ① When the flame is bright (mainly showing bright white characteristics) and the background is bright yellow or bright red, the R, G, and B component values of the background area around the flame are all large, and the distinguishability from the flame area is relatively small, especially the R component value has the smallest distinguishability (as shown in Figure 3 ). At this time, the color flame image is grayscaled based on the weighted average method to obtain a weighted average grayscale flame image. The specific calculation formula is shown in formula (9):

[0075] f(i, j) = 0.299R f (i, j) + 0.587G f (i, j) + 0.114B f (i, j) (9)

[0076] In the formula, R f (i, j), G f (i, j), B f(i,j) —— The R, G, and B components of the pixel at (i,j) in the color flame combustion image.

[0077] At this time, observing its weighted average gray-scale flame image ( Figure 4 (b)), it can be seen that although there are differences in the pixel values between some background areas near the flame and the flame area, compared with the background areas in the distance, the two are relatively close. When directly using the Otsu method for binary segmentation, this part of the background area is likely to be segmented as the flame area, resulting in an "under-segmentation" situation in the target flame area and affecting the accuracy of its binary segmentation.

[0078] At this time, based on the Otsu method, binary segmentation can be performed on the weighted average gray-scale flame image after contrast enhancement for the background area and the flame area to improve the accuracy of the segmentation of the target flame area. The specific binary segmentation algorithm is as follows:

[0079] (a) For the weighted average gray-scale flame image, perform binary segmentation based on the Otsu method to obtain the target flame area (denoted as area A1). At this time, there is often an "under-segmentation" situation in the segmented and extracted target flame area, and some background areas are segmented into it.

[0080] (b) Traverse and obtain the pixel values of each pixel point within the range of area A1 in the weighted average gray-scale flame image in turn, find the minimum pixel value among them (denoted as Pix_Min1), and then traverse each pixel point outside area A1 in the weighted average gray-scale flame image in turn, and assign its pixel value to Pix_Min1. Thus, the gray-scale values of the bright and dark areas of the background are equalized, and further through the contrast stretching method, the contrast between the background area and the flame area is increased, so as to obtain the weighted average gray-scale flame image after contrast enhancement (denoted as F1), as Figure 5 (a) shown.

[0081] (c) For the weighted average gray-scale flame image F1 after contrast enhancement, perform binary segmentation based on the Otsu method to obtain the final target flame area, as Figure 5 (b) shown. At this time, the segmented and extracted target flame area is relatively accurate, and the aforementioned "under-segmentation" situation is well improved.

[0082] ② When the flame is yellow and black mixed, and the background is dark yellow or dark red, the R, G, and B component values of the yellow and black mixed flame area are all small, and the distinguishability from its nearby background area is small, especially the distinguishability of its B component value is the smallest (as Figure 6 shown), at this time, observing its weighted average gray-scale flame image, it can be seen that, as Figure 7As shown, the pixel values of the yellow-and-black mixed flame area are relatively close to those of the background area. When directly using Otsu's method for binary segmentation, the yellow-and-black mixed flame area is extremely likely to be segmented into the background area, resulting in "over-segmentation" of the target flame area and affecting the accuracy of its binary segmentation.

[0083] Observe the R, G, and B component value characteristics of the flame combustion image (mainly including bright white, bright yellow, and yellow-and-black mixed flame areas, as well as the background area), and adopt the following binary segmentation algorithm:

[0084] (a) It can be observed that the R, G, and B component values of the bright white flame area are the largest, and all three component values are close to 255. The R, G, and B component values of the bright white flame area, bright yellow flame area, yellow-and-black mixed flame area, and background area show a decreasing characteristic in turn.

[0085] (b) The R component values of the bright white and bright yellow flame areas are relatively large, and there is an obvious difference from the R component value of the background area. However, the R component value of the yellow-and-black mixed flame area is relatively small, and it is closer to the R component value of the background area compared with the bright white and bright yellow flame areas. At this time, for the R-frame flame grayscale image, binary segmentation can be performed based on Otsu's method to obtain the bright white and bright yellow flame areas, and the extracted binary target area is denoted as A2.

[0086] (c) Compared with the difference between the R and B component values of the bright white flame area and the background area, the difference between the R and B component values of the bright yellow and yellow-and-black mixed flame areas is usually relatively obvious. Based on the 2(R - B) method for grayscale conversion of the color flame image, the bright yellow and yellow-and-black mixed flame areas can be enhanced, and then binary segmentation can be performed based on Otsu's method. Usually, the bright yellow and yellow-and-black mixed flame areas can be extracted, and the extracted target area is denoted as A3. Among them, the specific operation of the 2(R - B) method is: subtract the B-frame grayscale image from the R-frame grayscale image, and then multiply the grayscale difference of each pixel point by 2. Specifically, when the background area shows dark red, the difference between the R and B component values of the background area will become obvious, and it is closer to the difference between the R and B component values of the bright yellow and yellow-and-black mixed flame areas compared with the difference between the R and B component values of the bright white flame area. At this time, in the above binary target area A3, the dark red background area will also be segmented into it, resulting in "over-segmentation". Therefore, traverse the pixel values of each pixel point in the 2(R - B) frame flame grayscale image in turn to find the maximum pixel value (denoted as Pix_Max1), and then traverse each pixel point within the range of area A2 in the 2(R - B) frame flame grayscale image in turn, and assign its pixel value to Pix_Max1, thereby obtaining the enhanced flame grayscale image of the bright white, bright yellow, and yellow-and-black mixed flame areas (denoted as F2).

[0087] (d) For the flame grayscale image F2, perform binary segmentation based on Otsu's method to obtain bright white, bright yellow, and yellow-black mixed flame regions. Denote the extracted binary target region as A4. By observing the flame grayscale image F2 and the extracted target region A4, it can be seen that there are still a small number of yellow-black mixed flame regions whose pixel grayscale values are darker than those of other flame regions, resulting in incomplete extraction. Therefore, traverse the pixel values of each pixel point outside the target region A4 in image F2 in sequence to find the maximum pixel value (denoted as Pix_Max2). Then, traverse each pixel point within the region A4 in image F2 in sequence and assign its pixel value to Pix_Max2. Thus, the equalization of the grayscale values of the bright and dark regions of the flame is achieved, and the contrast between the background region and the flame region is enhanced through contrast stretching to obtain a new flame grayscale image (denoted as F3). At this time, the bright white, bright yellow, and yellow-black mixed flame regions are overall better enhanced, as Figure 8 (a) shown.

[0088] (e) For the flame grayscale image F3 with enhanced target flame regions, perform binary segmentation based on Otsu's method to obtain the target flame region, as Figure 8 (b) shown. At this time, the segmented and extracted target flame region is relatively accurate, which better improves the aforementioned "over-segmentation" situation.

[0089] ③ When the flame is bright or bright yellow and the background is dark yellow or dark red, in its weighted average grayscale flame image, the distinction between the flame region and the background region is relatively large. By directly performing binary segmentation on the weighted average grayscale flame image based on Otsu's method, the target flame region can be effectively obtained.

[0090] Finally, perform morphological operation processing such as hole filling, erosion and dilation, and denoising on the binary images of the target flame regions obtained in the above two major categories and four cases of flame combustion brightness to obtain the final binary image of the target flame region.

[0091] Among them, the above four cases of flame combustion brightness can be automatically judged based on the following algorithm: For the R-frame background difference image, after binary segmentation based on Otsu's method, calculate and obtain the pixel area of its target flame region (denoted as S sub ), set the judgment threshold S th (small flame area judgment threshold). When S sub ≤S th is satisfied, it is judged as a small flame image; for the weighted average grayscale flame image, calculate and obtain the average grayscale pixels of the two regions, denoted as Avg_Fpix and Avg_Bpix, respectively, based on the flame region and the background region determined after its binary segmentation according to Otsu's method. Set the judgment thresholds P th1 (flame region brightness judgment threshold), P th2 (background region bright yellow / bright red judgment threshold), Pth3 (Brightness judgment threshold of the background area), P th4 (Judgment threshold for the yellow and black mixture in the flame area), P th5 (Dark yellow / dark red judgment threshold of the background area), when Avg_Fpix > P th1 and Avg_Bpix > P th2 , or Avg_Bpix > P th3 it is judged that the flame is bright, the background is bright yellow / bright red, and the binary segmentation of the target flame area is prone to the situation of "under" segmentation. When Avg_Fpix < P th4 and Avg_Bpix < P th5 it is judged that the flame is yellow and black mixed, the background is dark yellow / dark red, and the binary segmentation of the target flame area is prone to the situation of "over" segmentation. Otherwise, it is judged that the flame is bright or bright yellow, and the background is dark yellow or dark red, and the discrimination between the two is relatively large, and the target flame area is easy to be binary segmented. For the thresholds S th , P th1 , P th2 , P th3 , P th4 , P th5 , P

[0092] S150. Detect the color features and morphological feature parameters for each frame of the target flame area to determine the combustion sufficiency and combustion stability of the diesel burner.

[0093] The combustion sufficiency and stability of the diesel burner are closely related to the characteristics of the combustion flame color, morphology, etc. When the flame burns insufficiently, the flame presents a bright yellow, yellow and black mixed state, the flame is soft and long, the fluctuation amplitude is large, and the propagation speed is relatively slow. When the flame burns stably, the flame presents a bright white, light blue or even colorless state, the flame is short, the fluctuation amplitude is tightened, and the propagation speed is accelerated.

[0094] To this end, this embodiment analyzes the current combustion sufficiency of the burner based on the color of the burning flame, and judges the current combustion stability of the burner based on the fluctuation amplitude of the morphological characteristics such as the height and area of ​​the flame area in the continuous frame flame combustion image (such as its average value, standard deviation, etc.). When the burning flame is light blue or even almost colorless, the combustion sufficiency is the best. At this time, the combustion state can be comprehensively judged based on the "no flame" state under the visible light perspective, combined with other relevant parameters such as the numerical range of the infrared thermal image test temperature and its stability; when the burning flame is bright white, the combustion sufficiency is good, and the R, G, and B component values ​​of the flame area are relatively close (all close to 255); when the burning flame is bright yellow or a mixed color of yellow and black, the combustion is not very complete, and the flame area presents a B component value of <G分量值<R分量值状态,尤其是其B分量值明显低于R分量值。对于燃烧稳定性,在开火后的不同燃烧阶段,其中稳定燃烧阶段的形态特征值波动最小,且其红外热像测试温度数据值较大、平稳性也最好,据此可判断燃烧器的燃烧稳定性。由此,对比分析各型燃烧器在不同燃烧阶段的燃烧充分性和燃烧稳定性。具体检测判断算法如下:

[0095] ①Judging combustion sufficiency based on color feature detection

[0096] For the current frame 2 (RB) grayscale flame image, the pixel values ​​of each pixel in the target flame area are traversed and obtained in turn, and the minimum pixel value (denoted as Pix_Min2) is found. Then, each pixel outside the target flame area is traversed in turn, and its pixel value is assigned to Pix_Min2. In this way, the contrast between the incomplete combustion flame area and the background area and the fully burned flame area in the 2 (RB) grayscale flame image is enhanced. Based on the Otsu method, the binarization can be performed to effectively segment and extract the incomplete combustion flame area. The area of ​​the target flame area of ​​the current frame (denoted as S0) and the area of ​​the incomplete combustion flame area (denoted as S1) are calculated. The sufficiency of the current combustion is judged according to the ratio of S1 to S0 (called the proportion of incomplete combustion area). The closer the ratio is to 0, the more complete the combustion is. Set the combustion sufficiency judgment threshold R th (This value can be optimized and adjusted according to actual use requirements, and the value here is 0.25). When S1 / S0 is satisfied <R th When the combustion is sufficient, it is judged as complete combustion; otherwise, it is judged as incomplete combustion.

[0097] ② Judging combustion stability based on morphological feature detection

[0098] Based on the binary image of the target flame area, calculate the height of the target area (i.e., the height of the flame pixels) and the area of the target area (i.e., the pixel area of the flame area) in the current frame. Process the flame combustion video images frame by frame according to the above steps to obtain the flame height values and the flame area of each image frame, and draw the flame combustion height curve (denoted as L H ), the flame combustion area curve (denoted as L A ). Calculate the average value corresponding to the height curve L H and the average value and standard deviation corresponding to the area curve L A respectively, and denote them as V H-avg , V A-avg , V A-std . The calculation method refers to Equations (2) and (3).

[0099] Based on the above morphological feature parameters of V H-avg , V A-avg , V A-std , combined with the average value of the highest temperature in the continuous image frames detected by the infrared thermal imager (denoted as T av ), the combustion stability of the burner can be judged. When T av > T th (the temperature threshold during stable combustion), and at the same time V H-avg < V H-th1 (the flame height threshold during stable combustion), V A-avg < V A-th1 (the flame area threshold during stable combustion) and V A-std < V A-th2 (the fluctuation amplitude threshold of the flame area during stable combustion), it can be judged that the burner is in a stable combustion state. Among them, when T av > T th , when V H-avg , V A-avg , V A-std are close to 0, the combustion flame presents a light blue or even almost colorless state at this time, and the burner is in a stable and full combustion state. For the values of the thresholds T th , V H-th1 , V A-th1 , V A-th2 , they can be adjusted by testing the infrared thermal imaging temperature range, the fluctuation amplitude of the flame height, the fluctuation amplitude of the flame area, etc. when the burner is in a stable combustion state in the actual use environment. Usually, different types of burners (differences in preheating tube structure, combustion power, combustion sufficiency, etc.) and different use environment conditions (such as differences in use environment temperature, altitude, etc.) will have different corresponding thresholds.

[0100] Further, to verify the effectiveness of the method in this embodiment, taking the BRS-7 type as an example, the processing results and analysis of this method are shown below. Specifically, in any selected frame of different combustion stages, the image processing area is determined, the target flame area is extracted, and the proportion of the insufficient combustion area is calculated. Finally, it is concluded that: ① The image processing area determination algorithm in this embodiment can accurately lock the flame combustion area, effectively narrow the image processing range, and greatly improve the processing efficiency; ② The image binarization algorithm proposed for the study of two major categories and four cases of flame combustion brightness can effectively enhance the image contrast, strengthen the target flame area, improve the "under-segmentation" and "over-segmentation" situations, and accurately extract the target flame area; ③ The calculated proportion of the insufficient combustion area in different combustion stages is consistent with the intuitive judgment result of the human eye, and can effectively complete the judgment of the flame combustion sufficiency; ④ The combustion sufficiency in the stable combustion stage is the best, indicating that the current combustion state of the burner can be determined by combining the infrared thermal image temperature data and the proportion value of the insufficient combustion area.

[0101] Taking the BRS-7 type as an example, the pixel area curve of the flame area, the flame height curve of the continuous image frames, the area difference curve of the flame area of the continuous image frames and their corresponding average values and standard deviations in different combustion stages (i.e., four combustion stages: when preheating is completed, after firing, during stable combustion, and when the firepower overshoots and is unstable, and 2s duration is arbitrarily selected for each stage) are processed and calculated in sequence, as Figure 9 shown. Finally, it is concluded that: By comparing and analyzing the average value and standard deviation of the flame area of the continuous image frames, and the average value of the flame height in the four combustion stages of when preheating is completed, after firing, during stable combustion, and when the firepower overshoots and is unstable, it can be seen that the corresponding values during stable combustion are the smallest, the corresponding values when the firepower overshoots and is unstable are the largest, and the corresponding values when preheating is completed and after firing are in the middle, which is consistent with the intuitive judgment result of the human eye, indicating that the judgment of the combustion stable state can be effectively completed based on the above three indicators.

[0102] Meanwhile, the continuous image frame flame area pixel area curve, continuous image frame flame height curve, continuous image frame flame area difference curve and their corresponding average values and standard deviations of other types of burners at two typical combustion stages (any 2s duration is selected for each stage) when the preheating is completed and the combustion is stable are processed and calculated in sequence. Combining with the infrared thermal imaging temperature data of the corresponding stages, the combustion stability of each type of burner is compared and analyzed, and the comparison and analysis results are as follows: ① For each type of burner, the average value and standard deviation of the continuous image frame flame area and the average value of the continuous image frame flame height in the stable combustion stage are the smallest. The combustion stability state of each type of burner can be effectively judged based on the above three indicators; ② The infrared thermal imaging temperature of high-power burners (such as Brother Jiedeng BRS-7, Aiwanke D61, Rhinoceros 203, and Kaku MA-01) can reach 700-900°C during stable combustion. Among them, the Aiwanke D61 and Rhinoceros 203 burners show a "flameless" combustion state during the stable combustion stage, and the values of the continuous image frame flame area pixel area and the continuous image frame flame height are both 0, with the best combustion sufficiency and stability. The Brother Jiedeng BRS-7 and Kaku MA-01 burners are second, indicating that the combustion stability and sufficiency of the four-vertical-tube flat-head structure burners are better than those of the single / double-span tube structure; ③ The infrared thermal imaging temperature of low-power burners (such as Brother Jiedeng BRS-29B, Aiwanke DK58, and Yuqing brand YQ-001) can reach 600-700°C during stable combustion. Comparing the average value and standard deviation of the continuous image frame flame area and the average value of the continuous image frame flame height, YQ-001 < DK58 < BRS-29B, indicating that the YQ-001 burner has the best combustion stability, that is, the combustion stability of the single-tube across-structure burner is better than that of the single-tube side-span burner.

[0103] In summary, this embodiment provides a method for detecting the combustion performance of a diesel burner based on machine vision. First, the combustion flame video image of the diesel burner to be detected is obtained; the flame type is judged according to the brightness change characteristics of each frame of the flame combustion image, and the target grayscale image for dividing the image processing area is determined; according to the brightness difference between the flame combustion area and the background area and the burner area in the target grayscale image, the processing area of each frame of the image is determined; for the image processing area, the target flame area is enhanced by enhancing the image contrast to achieve accurate binarization of the target flame area in each frame of the image; for each frame of the target flame area, color feature and morphological feature parameter detection are performed to determine the combustion sufficiency and combustion stability of the diesel burner. The whole method is automatically realized through machine vision technology, avoiding the subjectivity of manual visual inspection of the flame state, and there is no need to use instrument equipment for testing. The detection accuracy is high, the process is simple, and there is no safety hazard.

[0104] Figure 10The following is a schematic structural diagram of an electronic device provided by an embodiment of the present invention. As Figure 10 shown, the device includes a processor 60, a memory 61, an input device 62, and an output device 63. The number of processors 60 in the device may be one or more. Figure 10 Here, one processor 60 is taken as an example. The processor 60, the memory 61, the input device 62, and the output device 63 in the device may be connected through a bus or other means. Figure 10 Here, connection through a bus is taken as an example.

[0105] The memory 61, as a computer-readable storage medium, can be used to store software programs, computer-executable programs, and modules, such as program instructions / modules corresponding to the method for dynamically updating node routing in a wireless short-range communication network in an embodiment of the present invention. The processor 60 executes various functional applications and data processing of the device by running the software programs, instructions, and modules stored in the memory 61, that is, implements the above-mentioned method for dynamically updating node routing in a wireless short-range communication network.

[0106] The memory 61 may mainly include a program storage area and a data storage area. Among them, the program storage area can store an operating system and application programs required for at least one function. The data storage area can store data created according to the use of the terminal, etc. In addition, the memory 61 may include a high-speed random access memory, and may also include a non-volatile memory, such as at least one magnetic disk storage device, a flash memory device, or other non-volatile solid-state storage devices. In some instances, the memory 61 may further include a memory remotely provided with respect to the processor 60, and these remote memories can be connected to the device through a network. Examples of the above network include but are not limited to the Internet, an enterprise intranet, a local area network, a mobile communication network, and combinations thereof.

[0107] The input device 62 can be used to receive input digital or character information, and generate key signal inputs related to the user settings and function controls of the device. The output device 63 may include a display device such as a display screen.

[0108] An embodiment of the present invention also provides a computer-readable storage medium, on which a computer program is stored, and when the program is executed by a processor, it implements the method for dynamically updating node routing in a wireless short-range communication network in any embodiment.

[0109] The computer storage medium of the embodiments of the present invention may adopt any combination of one or more computer-readable media. The computer-readable media may be computer-readable signal media or computer-readable storage media. The computer-readable storage media may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination of the above. More specific examples (a non-exhaustive list) of the computer-readable storage media include: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In this document, the computer-readable storage media may be any tangible medium that contains or stores a program, which can be used by or in conjunction with an instruction execution system, apparatus, or device.

[0110] The computer-readable signal media may include a data signal propagated in a baseband or as part of a carrier wave, which carries the computer-readable program code. Such a propagated data signal may take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the above. The computer-readable signal media may also be any computer-readable media other than the computer-readable storage media, which can send, propagate, or transmit a program for use by or in conjunction with an instruction execution system, apparatus, or device.

[0111] The program code contained on the computer-readable media may be transmitted by any appropriate medium, including but not limited to wireless, wire, optical cable, RF, etc., or any suitable combination of the above.

[0112] The computer program code for performing the operations of the present invention may be written in one or more programming languages or combinations thereof. The programming languages include object-oriented programming languages such as Java, Smalltalk, C++, and also include conventional procedural programming languages such as the C language or similar programming languages. The program code may be executed entirely on the user's computer, partially on the user's computer, executed as an independent software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In the case of a remote computer, the remote computer may be connected to the user's computer through any type of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computer (for example, by using an Internet service provider to connect through the Internet).

[0113] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions described in the foregoing embodiments, or perform equivalent replacements on some or all of the technical features; and these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the technical solutions of the embodiments of the present invention.

Claims

1. A method for detecting the combustion performance of a diesel burner based on machine vision, characterized in that, Including: Obtain the combustion flame video image of the diesel burner to be detected; Judge the flame type according to the brightness change characteristics of each frame of the flame combustion image, and determine the target grayscale image for segmenting the image processing area for each frame; Determine the processing area of each frame of image according to the brightness difference between the flame combustion area and the background area and the burner area in the target grayscale image; For the image processing area, enhance the target flame area by enhancing the image contrast to achieve accurate binarization of the target flame area in each frame of the graph; Detect the color features and morphological feature parameters for each frame of the target flame area to determine the combustion sufficiency and combustion stability of the diesel burner.

2. The method according to claim 1, wherein The step of judging the flame type according to the brightness change characteristics of each frame of the flame combustion image and determining the target grayscale image for segmenting the image processing area for each frame includes: Regard the image before the diesel burner is ignited as the background image, and set the area outside the diesel burner in the background image to white; According to the background difference method, calculate the R-frame grayscale difference map between the current frame of the flame combustion image and the background image; Determine the flame type of the current frame of the flame combustion image according to the pixel fluctuation amplitude of the difference map; If the flame type is a small flame, use the R-frame grayscale difference map as the target grayscale image for determining the image processing area; If the flame type is a non-small flame, jointly use the R-frame grayscale image of the background image and the R-frame grayscale image of the flame combustion image as the target grayscale image for determining the image processing area.

3. The method according to claim 1, characterized in that, The step of determining the processing area of each frame of image according to the brightness difference between the flame combustion area and the background area and the burner area in the target grayscale image includes: For the current frame of the target grayscale image, determine the current frame of the image processing area by finding the brightness jump dividing lines in the horizontal and vertical directions according to the brightness difference between the flame combustion area and the background area and the burner area.

4. The method according to claim 3, wherein When the current frame of the target grayscale image is the R-frame grayscale difference map, the step of determining the current frame of the image processing area by finding the brightness jump dividing lines in the horizontal and vertical directions according to the brightness difference between the flame combustion area and the background area and the burner area includes: For the R-frame grayscale difference map, directly set the lower boundary line of the current frame of the flame combustion image as the lower dividing line of the current frame of the image processing area; For the R-frame grayscale difference map, construct the first distribution curve of the vertical cumulative pixel values; determine the upper dividing line of the current frame of the image processing area according to the peak point of the first distribution curve; According to the first distribution curve, construct the difference curve between consecutive points of the vertical cumulative pixel values; determine the left and right dividing lines of the current frame of the image processing area according to the jump points on both sides of the difference curve that cross the average horizontal line; Use each dividing line to segment the current frame of the flame combustion image to obtain the current frame of the image processing area.

5. The method according to claim 3, characterized in that, When the target grayscale image of the current frame includes the R-frame grayscale image of the background image and the R-frame grayscale image of the flame combustion image of the current frame, determining the image processing area of the current frame according to the brightness difference between the flame combustion area, the background area, and the burner area, including: Construct a second distribution curve of the horizontal cumulative pixel values for the R-frame grayscale image of the background image; determine the lower boundary line of the image processing area of the current frame according to the jump points of the second distribution curve; Construct a third distribution curve of the vertical cumulative pixel values for the R-frame grayscale image of the flame combustion image of the current frame; determine the upper boundary line of the image processing area of the current frame according to the peak points of the third distribution curve; Construct a difference curve between consecutive points of the vertical cumulative pixel values according to the third distribution curve; determine the left and right boundary lines of the image processing area of the current frame according to the jump points on both sides of the difference curve that cross the average value horizontal line; Segment the flame combustion image of the current frame using each boundary line to obtain the image processing area of the current frame.

6. The method according to claim 1, wherein For the image processing area, enhancing the target flame area by enhancing the image contrast to achieve accurate binarization of the target flame area in each frame of the image, including: Judging the flame brightness situation of the image processing area of the current frame according to the flame area area feature and the grayscale value features of the flame and background areas; If the flame brightness situation is a small flame, grayscale the color image of the image processing area by background difference method, and perform binarization on the grayscale image based on Otsu's method to obtain the target flame area; If the flame brightness situation is that the flame is bright and the background is bright yellow / bright red, grayscale the color image of the image processing area by weighted average method, enhance the contrast between the bright white area and the bright yellow / bright red area in the grayscale image, and perform binarization on the enhanced grayscale image based on Otsu's method to obtain the target flame area; If the flame brightness situation is that the flame is yellow and black mixed and the background is dark yellow / dark red, perform binarization on the R-frame grayscale image of the image processing area of the current frame based on Otsu's method to obtain the bright white and bright yellow flame areas; grayscale the color image of the image processing area of the current frame based on the 2(R - B) method, enhance the bright yellow and yellow and black mixed color flame areas, and further enhance the bright white and bright yellow flame areas in the grayscale image by the maximum pixel grayscale assignment method for the target area; for the grayscale image of the bright white, bright yellow, and yellow and black mixed color flame areas after enhancement, achieve the equalization of the grayscale values of the bright white, bright yellow, and yellow and black mixed flame areas in the grayscale image by the minimum pixel grayscale assignment method for the target area, reduce the discrimination degree during binarization of the bright white, bright yellow, and yellow and black mixed flame areas, and enhance the discrimination degree during binarization with the background area; perform binarization on the image after re-enhancement based on Otsu's method to obtain the target flame area; If the flame brightness situation is that the flame is bright / bright yellow and the background is dark yellow / dark red, grayscale the color image of the image processing area by the weighted average method, and perform binary segmentation on the grayscale image based on Otsu's method to obtain the target flame area.

7. The method according to claim 6, wherein Enhancing the contrast between the bright white area and the bright yellow / bright red area in the grayscale image includes: Performing binarization on the grayscale image based on Otsu's method to obtain an under-segmented flame area, where the under-segmented flame area includes a bright white flame area and a bright yellow / bright red background area; Modifying the pixel values outside the under-segmented flame area to the minimum pixel value in the under-segmented flame area; Performing contrast stretching on the modified grayscale image to enhance the contrast between the bright white flame area and the bright yellow / bright red background area.

8. The method according to claim 6, wherein Further enhancing the bright white and bright yellow flame areas in the grayscale image by the maximum pixel gray value assignment method for the target area includes: modifying the pixel values of the bright white and bright yellow flame areas in the grayscale image to the maximum pixel value in the grayscale image to enhance the bright white, bright yellow, and yellow-black mixed flame areas; Correspondingly, achieving the equalization of the gray values of the bright white, bright yellow, and yellow-black mixed flame areas in the grayscale image by the minimum pixel gray value assignment method for the target area, reducing the distinguishability during the binarization of the bright white, bright yellow, and yellow-black mixed flame areas, and enhancing the distinguishability during the binarization with the background area includes: performing binarization on the enhanced image based on Otsu's method to obtain the bright white, bright yellow, and yellow-black mixed flame areas; modifying the pixel values within the bright white, bright yellow, and yellow-black mixed flame areas to the maximum pixel value outside the bright white, bright yellow, and yellow-black mixed flame areas in the enhanced image; performing contrast stretching on the modified image to enhance the contrast between the bright white, bright yellow, and yellow-black mixed flame areas and the background area in the image.

9. The method according to claim 1, wherein Detecting the color features and morphological feature parameters for each frame of the target flame area to determine the combustion sufficiency of the diesel burner includes: Enhancing the contrast between the non-fully combusted flame area and the background area and the fully combusted flame area for the current frame 2(R - B) grayscale flame image; Performing binarization on the enhanced image to segment and extract the non-fully combusted flame area therein; Judging the combustion sufficiency of the diesel burner according to the ratio of the area of the target flame area of the current frame to the area of the non-fully combusted flame area.

10. The method according to claim 1, wherein Detecting the color features and morphological feature parameters for each frame of the target flame area to determine the combustion stability of the diesel burner includes: Processing the target flame area frame by frame to obtain the flame height value and the flame area of each frame, and drawing the flame height curve and the flame area curve; Judging the combustion stability of the diesel burner according to the fluctuation amplitude characteristics of the two curves.

Images