CN113222929A - Smoke concentration detection method and device based on total variation - Google Patents
Smoke concentration detection method and device based on total variation Download PDFInfo
- Publication number
- CN113222929A CN113222929A CN202110495053.5A CN202110495053A CN113222929A CN 113222929 A CN113222929 A CN 113222929A CN 202110495053 A CN202110495053 A CN 202110495053A CN 113222929 A CN113222929 A CN 113222929A
- Authority
- CN
- China
- Prior art keywords
- image
- total variation
- smoke concentration
- variation
- smoke
- 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.)
- Withdrawn
Links
- 239000000779 smoke Substances 0.000 title claims abstract description 186
- 238000001514 detection method Methods 0.000 title claims abstract description 41
- 230000008033 biological extinction Effects 0.000 claims abstract description 53
- 238000004364 calculation method Methods 0.000 claims abstract description 19
- 238000004458 analytical method Methods 0.000 claims abstract description 15
- 238000000034 method Methods 0.000 claims description 46
- 238000012544 monitoring process Methods 0.000 claims description 17
- 238000004422 calculation algorithm Methods 0.000 claims description 16
- 238000012545 processing Methods 0.000 claims description 11
- 238000003860 storage Methods 0.000 claims description 9
- 230000010354 integration Effects 0.000 claims description 8
- 238000000605 extraction Methods 0.000 claims description 6
- 238000005070 sampling Methods 0.000 claims description 5
- 238000005457 optimization Methods 0.000 claims description 3
- 230000009467 reduction Effects 0.000 claims description 3
- 230000009191 jumping Effects 0.000 abstract description 6
- 230000007423 decrease Effects 0.000 abstract description 5
- 230000006870 function Effects 0.000 description 25
- 238000010586 diagram Methods 0.000 description 8
- 238000004590 computer program Methods 0.000 description 7
- 230000008569 process Effects 0.000 description 7
- 238000005259 measurement Methods 0.000 description 4
- 238000011156 evaluation Methods 0.000 description 3
- 230000008447 perception Effects 0.000 description 3
- 230000003068 static effect Effects 0.000 description 3
- 238000010521 absorption reaction Methods 0.000 description 2
- 230000015556 catabolic process Effects 0.000 description 2
- 238000006731 degradation reaction Methods 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 230000007613 environmental effect Effects 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 241001028048 Nicola Species 0.000 description 1
- 230000002238 attenuated effect Effects 0.000 description 1
- 235000013405 beer Nutrition 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000033228 biological regulation Effects 0.000 description 1
- 230000000739 chaotic effect Effects 0.000 description 1
- 238000006243 chemical reaction Methods 0.000 description 1
- 230000003247 decreasing effect Effects 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 230000023077 detection of light stimulus Effects 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 239000000284 extract Substances 0.000 description 1
- 230000004907 flux Effects 0.000 description 1
- 230000010365 information processing Effects 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 230000008520 organization Effects 0.000 description 1
- 230000001902 propagating effect Effects 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 230000011218 segmentation Effects 0.000 description 1
- 238000012795 verification Methods 0.000 description 1
- 230000000007 visual effect Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/0002—Inspection of images, e.g. flaw detection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/40—Extraction of image or video features
- G06V10/60—Extraction of image or video features relating to illumination properties, e.g. using a reflectance or lighting model
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/40—Scenes; Scene-specific elements in video content
- G06V20/41—Higher-level, semantic clustering, classification or understanding of video scenes, e.g. detection, labelling or Markovian modelling of sport events or news items
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/40—Scenes; Scene-specific elements in video content
- G06V20/46—Extracting features or characteristics from the video content, e.g. video fingerprints, representative shots or key frames
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10016—Video; Image sequence
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Multimedia (AREA)
- Software Systems (AREA)
- Computational Linguistics (AREA)
- Quality & Reliability (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Investigating Or Analysing Materials By Optical Means (AREA)
- Fire-Detection Mechanisms (AREA)
Abstract
The invention provides a smoke concentration detection method and device based on total variation. The jumping part and the smooth fluctuation part form a digital image, and correspondingly, the difference of the texture boundary of the video image containing smoke is reduced, and the jumping part is relatively smooth and can be treated as a blurred image. Based on the characteristic properties of Total Bound Variation (TBV), the Total Variation can effectively characterize the difference between image boundaries. The image blur corresponding to the increase in smoke density increases and the video image variation reflecting this smoke information decreases. On the basis of a total variation theory, a relation model between total variation and image ambiguity is constructed to accurately measure the relation between the smoke concentration and the total variation, a functional is constructed on the basis of a Koschmieder theory, variation is solved, and an extreme value is solved for an objective function by the idea of piecewise smooth analysis, so that an extinction coefficient is solved, and the current smoke concentration is obtained. The invention avoids errors caused by complex calculation, has accurate and stable output and wide application prospect.
Description
Technical Field
The invention belongs to the technical field of information processing, and particularly relates to a smoke concentration detection method and device based on total variation.
Background
The smoke concentration is a measure of the amount of smoke and the degree of damage to the smoke, and is also a weather indicator for knowing the stability of the atmosphere and the vertical structure. The smoke concentration quantitative measurement has important significance for the fields of traffic transportation safety, fire early-stage smoke detection and military affairs, and is an extremely important factor. The traditional smog degree instrument detection method is limited by a sampling space, is difficult to reflect the real condition of the environment, and is expensive, poor in flexibility and not suitable for popularization.
In 2013, a method for detecting smoke concentration based on image gray scale characteristics is provided by a Yuanfei pavilion, in fire detection, the accuracy and the anti-interference capability of fire detection are improved, but the method can only realize the detection of light white smoke concentration, has certain dependence on the gray scale characteristics of a background image during measurement, and is not ideal in black smoke detection effect. In 2014, Miao Ligang improved smoke detection, realized video smoke detection through the dark channel method, roughly approximated the smoke concentration to the intensity of the dark channel. In 2017, martian corp provides a method for calculating smoke concentration through image structure similarity, the conversion coefficient is replaced by a negative value of image complexity, the method realizes quantitative measurement of smoke concentration under different background image conditions based on the image structure similarity, and the smoke concentration value can be measured accurately while the existence of smoke is detected qualitatively. In 2018, Cheng Xiagang proposes a variation framework to process the time-varying property of the extinction coefficient, and extracts the extinction coefficient by performing piecewise function fitting on an observed brightness curve, thereby realizing real-time atmospheric visibility estimation in haze weather.
However, smoke has a certain similarity to noise, the background image is also blurred by the presence of smoke in the image, resulting in a decrease in image quality, and the greater the smoke density, the greater the degree of image quality degradation. Some of the smoke detection methods are realized by detecting the dynamic characteristics or the static characteristics of smoke, and have certain dependence on the gray level characteristics of a background image, which can result in weaker capability when distinguishing objects similar to the smoke characteristics and certain limitation when distinguishing the objects similar to the smoke characteristics; some methods for realizing visibility detection of the smoke image do not accurately and quantitatively calculate the smoke concentration. The core of the image-based smoke detection method is how to represent the self characteristics of smoke, and a stable and efficient characteristic extraction algorithm is constructed, so that the method becomes the key for reducing smoke concentration errors.
Disclosure of Invention
The invention aims to overcome the defects in the prior art and provide a smoke concentration detection method and device based on total variation, which can efficiently and stably detect smoke concentration and reduce smoke concentration detection errors.
In order to achieve the purpose, the invention is realized by adopting the following technical scheme:
in a first aspect, the invention provides a smoke concentration detection method based on total variation, which comprises the following steps:
acquiring a video image and a clear image;
extracting the video image to obtain an image containing smoke concentration information;
performing image brightness analysis on the image containing the smoke concentration information, performing total variation calculation to obtain a total variation (TBV) value of the image,
and calculating and obtaining an extinction coefficient, namely the smoke concentration according to the total variation (TBV) value of the image.
Further, the method for calculating the total variation (TBV) value of the image by performing the total variation calculation comprises the following steps:
respectively calculating the total variation TBV of the clear image ffAnd the total variation TBV of the blurred image ggRespectively solved by the following formula:
wherein f represents a sharp image, the function f (x, y) represents the (x, y) th pixel in the image, and the pixel of the blurred image is the function g (x, y); wherein Ω is represented by the following formula:
Ω＝{(x,y):0≤x≤m,0≤y≤n}
and m and n are the sizes of the original clear images.
Further, the method for calculating and acquiring the extinction coefficient according to the total variation (TBV) value of the image comprises the following steps:
and acquiring a brightness curve of the image according to the total variation (TBV) value of the image, and acquiring an extinction coefficient k according to the brightness curve of the image.
Further, the method for obtaining the brightness curve of the image according to the total variation (TBV) value of the image and obtaining the extinction coefficient k according to the brightness curve of the image includes:
based on a smoke concentration estimation algorithm of total variation, regarding the image containing smoke concentration information as a fuzzy image, combining image characteristics, and then continuously performing approximate comparison on the brightness estimation value of the target object and the observed actual brightness value of the target object to obtain a brightness curve;
and constructing a functional, solving the variation of the functional, solving an extreme value of the objective function, performing convolution and integration processing on the brightness curve based on the total variation principle, and then solving the adjusted extinction coefficient by using a quasi-Newton iteration method. Further, the method for constructing the functional, solving the variation of the functional, solving the extreme value of the objective function, performing convolution and integration processing on the brightness curve based on the total variation principle, and then solving the adjusted extinction coefficient by using a quasi-Newton iteration method comprises the following steps of:
establishing an objective function to obtain a formula:
the above formula n is the sampling times, Q is the objective function, and the variables of the above formula are as follows:
deriving a cost function of
Further simplified to obtain
In the formula, L: the brightness of the target observed by the observation point; l is0: the target has own brightness; l isf: background sky brightness; k: extinction coefficient; d: the distance between the observation point and the target object;
according to the optimization algorithm, the free brightness L of the target object is obtained0And extinction coefficient kiNamely the formula:
on the basis, multiple iterations are carried out through a quasi-Newton iteration method, an extinction coefficient can be obtained, and the smoke concentration k is obtained.
Further, the method for extracting the video image and acquiring the image containing the smoke concentration information comprises the following steps:
denoising the video image to obtain a denoised video;
extracting monitoring frames with equal time intervals according to the de-noised video;
and performing ROI (region of interest) region extraction on the monitoring frames with equal time intervals to obtain an image containing smoke concentration information.
In a second aspect, the present invention provides a smoke concentration detection device based on total variation, the device comprising:
a video acquisition module: for acquiring video images;
an image extraction module: the image acquisition module is used for extracting the video image and acquiring an image containing smoke concentration information;
a calculation analysis module: the system is used for carrying out image brightness analysis on the image containing the smoke concentration information, carrying out total variation calculation and solving a total variation (TBV) value of the image;
a concentration acquisition module: and calculating and acquiring an extinction coefficient, namely the smoke concentration according to the total variation (TBV) value of the image.
In a third aspect, the invention provides a smoke concentration detection device based on total variation, which comprises a processor and a storage medium;
the storage medium is used for storing instructions;
the processor is configured to operate in accordance with the instructions to perform the steps of the method according to any of the above.
Compared with the prior art, the invention has the following beneficial effects:
1. based on the characteristic properties of the total variation, the total variation can effectively characterize the difference between the image boundaries. The image fuzziness corresponding to the increase of the smoke concentration is increased, and the variation of the video image reflecting the smoke information is reduced; on the basis of a total variation theory, a total variation value of an extreme value state is expressed, a relation model between the total variation and the image ambiguity is constructed, the relation between the smoke concentration and the total variation is accurately measured, and the criterion is used for evaluating the smoke concentration; the invention constructs an advanced smoke concentration monitoring system based on a video sensing system, improves the transparency of environmental smoke detection, and has the advantages of easy operation and the like based on a video analysis detection method;
2. the smoke concentration image detection method based on total variation comprises the steps of extracting monitoring frames at equal time intervals, analyzing an image containing smoke concentration information, solving an extreme value of a target function by the aid of the idea of piecewise stationary analysis to obtain a total bounded value, taking an extinction coefficient from a constant variable as a function of time through the total bounded variation principle on the basis, constructing a functional, further performing multiple iteration approximation through a quasi-Newton iteration method, solving the extinction coefficient, and estimating a real-time smoke concentration value. The algorithm shields the complex calculation of the smoke characteristics, reduces the calculation amount, avoids errors caused by complex calculation, has accurate and stable output of the algorithm result, and has better reliability and effectiveness;
3. based on a video image total variation technology, a bridge between image and information perception is established at a high-level semantic level, a smoke concentration perception system is established, and the perception and quantitative expression of smoke concentration information are realized;
4. the method fully utilizes the images acquired by reconnaissance monitoring, shields complex calculation of static characteristics of smoke, avoids errors caused by complex calculation, has accurate and stable result output, has the characteristics of high operability, good flexibility and the like, and has wide application prospect.
Drawings
FIG. 1 is a flow chart of the system of the present invention.
FIG. 2 is a graph showing the relationship between total variation and smoke concentration according to the present invention.
FIG. 3 shows the luminance curve and extinction coefficient of the present invention.
Fig. 4 is a relationship between the TBV of the haze image of the present invention and the smoke density thereof.
FIG. 5 is a comparison of the curve fit of the brightness of the target according to the present invention.
Detailed Description
The invention is further described below with reference to the accompanying drawings. The following examples are only for illustrating the technical solutions of the present invention more clearly, and the protection scope of the present invention is not limited thereby.
The first embodiment is as follows:
the embodiment provides a smoke concentration detection method based on total variation, which comprises the following steps:
first, the total variation is solved. The monitoring image of normal weather is clear, the main body is a high-frequency signal, and the boundary difference is large. The image blur degree corresponding to the increase of the smoke density is increased, and the TBV (total variation) value of the video image reflecting the smoke information is reduced. After haze exists, the image becomes fuzzy, namely a large amount of noise appears on a clear image, the thicker the haze is, the more the noise is, and the higher the smoke concentration is. When the smoke concentration caused by haze is higher and higher, the TBV (total variation) value of the monitoring image is smaller and smaller. At lower smoke concentrations, the value of TBV (total variation) is also lower. This is because the entire environment covers haze, which corresponds to noise superimposed on a clear video image, and therefore, the characteristic difference between the image boundaries is reduced, resulting in a smaller TBV (total variation). When the dense haze becomes light, the process of denoising the video image is actually performed, and when the haze becomes light or disappears, the characteristic difference between the dense haze and the corresponding image boundary is displayed.
And secondly, fitting approximation. On the basis of a total variation theory, a relation model between total variation and image ambiguity is constructed for accurately measuring the relation between smoke concentration and total variation and for the criterion of smoke concentration evaluation. Based on the smoke concentration estimation algorithm of total variation, the image containing smoke concentration information is regarded as a fuzzy image, the image characteristics are combined, and then the brightness estimation value of the target object and the observed actual brightness value of the target object are continuously subjected to approximation comparison to obtain a brightness curve.
And thirdly, determining the smoke concentration. On the basis of the property that the total variation value of an image is correspondingly reduced when the image becomes fuzzy, when the smoke concentration is detected, a functional is constructed on the basis of the Koschmieder theory, the variation is solved, an extreme value is solved for an objective function by the idea of piecewise smooth analysis, a brightness curve is subjected to convolution and integration processing on the basis of the total variation principle, and then the extinction coefficient is solved by a quasi-Newton iteration method.
The smoke concentration is a measure of the amount of smoke and the degree of damage to the smoke, and is also a weather indicator for knowing the stability of the atmosphere and the vertical structure. The smoke concentration quantitative measurement has important significance for the fields of traffic transportation safety, fire early-stage smoke detection and military affairs, and is an extremely important factor. The traditional smog degree instrument detection method is limited by a sampling space, is difficult to reflect the real condition of the environment, and is expensive, poor in flexibility and not suitable for popularization. Currently, a video sensing platform is widely applied, an advanced smoke concentration monitoring system is constructed based on a video sensing system, and the transparency of environmental smoke detection is improved. The video analysis-based detection method has the advantages of easiness in operation and the like, and becomes a research hotspot.
The smoke density detection method based on the total variation can be regarded as a blurred image in which the difference of the texture boundary of a video image containing smoke is small. Based on the characteristic properties of the total variation, the total variation can effectively characterize the difference between the image boundaries. The image blur corresponding to the increase in smoke density increases and the video image variation reflecting this smoke information decreases. On the basis of a total variation theory, a relation model between total variation and image fuzziness is constructed to accurately measure the relation between the smoke concentration and the total variation, a brightness curve is approximated through a total variation principle and then based on a segmentation stability idea, an extinction coefficient is obtained, and a real-time smoke concentration value is estimated. The algorithm fully utilizes the images acquired by reconnaissance monitoring, shields complex calculation of static characteristics of smoke, avoids errors caused by complex calculation, is accurate and stable in result output, has the characteristics of high operability, good flexibility and the like, and has a wide application prospect.
The implementation principle is as follows: the difference of the texture boundary of the video image containing smoke is reduced, the full variation value of the video image is correspondingly reduced, the image is more blurred when the smoke is larger, and the image can be treated as a blurred image when the full variation value of the video image is smaller. First, the full variation value is calculated. Based on the characteristic properties of the total variation, the total variation can effectively represent the difference between the image boundaries, and the total variation value is calculated by the definition of the total variation; then, fitting approximation is performed. And determining the relation between the total variation of the haze image and the smoke concentration of the haze image, and fitting the calculated TBV value of the image and the corresponding smoke concentration of the image. Based on a smoke concentration estimation algorithm of total variation, an image containing smoke concentration information is regarded as a fuzzy image, a brightness curve is given by combining image characteristics, and then a brightness estimation value of a target object and an observed actual brightness value of the target object are continuously subjected to approximation comparison; finally, the smoke concentration is determined. Based on the total variation principle, the luminance curve is processed by convolution and integration, and then the extinction coefficient is obtained by a quasi-Newton iteration method. The smoke concentration detection algorithm flow based on the total variation is shown in figure 1.
The invention provides a smoke concentration detection method and device based on total variation. The jumping part and the smooth fluctuation part form a digital image, and correspondingly, the difference of the texture boundary of the video image containing smoke is reduced, and the jumping part is relatively smooth and can be treated as a blurred image. Based on the characteristic properties of Total Bound Variation (TBV), the Total Variation can effectively characterize the difference between image boundaries. The image blur corresponding to the increase in smoke density increases and the video image variation reflecting this smoke information decreases. On the basis of a total variation theory, a relation model between total variation and image ambiguity is constructed to accurately measure the relation between the smoke concentration and the total variation, a functional is constructed on the basis of a Koschmieder theory, variation is solved, and an extreme value is solved for an objective function by the idea of piecewise smooth analysis, so that an extinction coefficient is solved, and the current smoke concentration is obtained. The invention avoids errors caused by complex calculation, has accurate and stable output and wide application prospect.
1 definition of Smoke concentration
Because the amount of smoke generated under different conditions is different, the result is influenced to a certain extent according to different measuring methods. Therefore, currently there is no uniform definition of smoke concentration. The mainstream method of defining the smoke concentration by selecting the extinction coefficient aiming at smoke such as haze is disclosed, and the smoke concentration mentioned in the later article refers to the extinction coefficient.
The extinction coefficient is an important parameter for detecting the smoke concentration, and is an absorption value of a detected medium to light and is divided into an absorption coefficient and a scattering coefficient. Beer lambert's law states that light energy will be attenuated continuously when propagating in a chaotic or absorbing medium. For plane waves, in the linear range, the energy consumed is proportional to the distance traveled by the light. As shown in formula (1)
dIr(f)＝-Ir(f)k(f,z)dz (1)
Where f denotes the clear image light wave frequency, dz is the differential distance, Ir(f) K (f, z), which represents the flux density incident on the face where dz is located, is the extinction coefficient, also known as the molar extinction coefficient, which is a function of f and z, which is closely related to smoke concentration.
Based on the above principle in 1924, Koschmieder summarizes the burger-lambert (Bougner-Lamber) law, deduces a relational formula between an extinction coefficient and the brightness of a target object, and lays the theoretical foundation for smoke concentration detection. That is, as shown in the formula (2)
L＝L0e-kd+Lf(1-e-kd) (2)
Wherein L is the brightness of the target observed at the observation point, L0Indicating the brightness of the subject, LfRepresenting background sky brightness, k is an extinction coefficient, d is the observation point-to-target distance, and for the environment, represents the distance from a monitoring camera to the target.
Duntley derived atmospheric attenuation law based on Koschmieder's law in 1949, as shown in formula (3)
C＝C0e-kd (3)
Wherein C represents the contrast of the brightness of the object, C0Is the target inherent brightness contrast. Let ε be (4)
Epsilon is called a visual contrast threshold, and the International Civil Aviation Organization (ICAO) recommends that epsilon be 0.05, so that the atmospheric smoke concentration shown in formula (5) can be obtained
(5) The formula is a basic formula for detecting the smoke concentration, and Con (k) represents the smoke concentration.
2 Smoke concentration detection method
The monitoring image of normal weather is clear, the main body is a high-frequency signal, and the boundary difference is large. After haze exists, the image becomes fuzzy, namely a large amount of noise appears on a clear image, the thicker the haze is, the more the noise is, and the higher the smoke concentration is. Verification data shows that when the smoke concentration caused by haze is higher and higher, the total variation value of the monitoring image is smaller and smaller. On the basis of the property that the total variation value of an image is correspondingly reduced when the image becomes fuzzy, a functional is constructed based on a Koschmieder theory and variation is solved, and an extreme value is solved for an objective function by the idea of piecewise smooth analysis when the smoke concentration is detected, so that an extinction coefficient is solved, the current smoke concentration is obtained, and the smoke concentration detection method based on the total variation is provided.
2.1 Total variational calculation
The jumping part and the smooth fluctuation part form a digital image, and correspondingly, the difference of the texture boundary of the video image containing smoke is reduced, and the jumping part is relatively smooth and can be treated as a blurred image. Based on the characteristic properties of the total variation, the total variation can effectively characterize the difference between the image boundaries. The image blur corresponding to the increase in smoke density increases and the video image variation reflecting this smoke information decreases. On the basis of a total variation theory, a relation model between total variation and image ambiguity is constructed to accurately measure the relation between the smoke concentration and the total variation, so that the total variation value of an extreme value state is expressed to be used as a criterion for smoke concentration evaluation.
According to the full variation definition, let f denote a sharp image, and the function f (x, y) denotes the (x, y) th pixel in the image. When the image is blurred due to the sharp smoke, the pixels of the blurred image are set as a function g (x, y), and when the noise influence is neglected, the function g (x, y) is h (x, y) f (x, y), and h (x, y) is a degradation function. The fully variant TBV discrete form of any MxN digital image z is
On the basis, the clearness can be respectively calculatedTotal variation TBV of image ffAnd the total variation TBV of the blurred image ggThe equations (7) and (8) are used to solve.
Omega in the formulas (7) and (8) is expressed as shown in the formula (9)
Ω＝{(x,y):0≤x≤m,0≤y≤n} (9)
And m and n are the sizes of the original clear images.
The relationship between smoke concentration and total confinement is shown in fig. 2. In the figure, the dotted line is a full variation value, the curve of the full variation value is the variation trend along with the reduction of the smoke concentration, the solid line is the smoke concentration and represents the fuzzy degree, and the curve of the full variation value is the variation trend along with the reduction of the smoke concentration. It can be seen from the opposite trend that when a clear image becomes blurred due to the increase of smoke, the full variation value of the clear image is correspondingly smaller, and the image becomes blurred more and less as the smoke is larger.
2.2 Brightness Curve fitting approximation
1. Fitting
Nicolas gave a luminance curve versus extinction coefficient in 2006, with distance on the horizontal axis and luminance on the vertical axis, as shown in fig. 3. In the figure, L1～L5The luminance curves are 5, and the corresponding extinction coefficients are respectively k 1-k 5. In the figure, the extinction coefficients have the following relationship
k1＜k2＜k3＜k4＜k5 (10)
The corresponding smoke concentration (Con) has the following relationship
Con(L1)＞Con(L2)＞Con(L3)＞Con(L4)＞Con(L5) (11)
It can be seen that the higher the smoke concentration, the more the curve of the luminance curve is curvedThe higher the degree. Conversely, the lower the smoke concentration, the lower the curvature. Experimental data show that under the condition that the smoke concentration is within 0.015dB/m, the brightness curve of a video image and L in the graph4And L5This feature provides the possibility for the algorithm implementation herein to be close, i.e. the curve is relatively flat.
Based on the characteristic properties of the total variation, the difference between the boundaries can be effectively characterized. When the smoke concentration is increased, the variation is reduced, and the corresponding extinction coefficient is increased, namely L in the graph5The corresponding total bounded value is less than L1The corresponding variation value. According to the existing theory, L is a monotone decreasing curve, so L can be considered0And k is an undetermined coefficient, and an L curve is fitted on the basis of variation. When the fitting value is close to L, the extinction coefficient k can be obtained.
Based on the monitoring video, a scene is collected, the scene belongs to the process of changing the dense haze into the light haze, the TBV value of the image is calculated in the process, and the TBV value is fitted with the corresponding smoke concentration.
The results are shown in FIG. 4. In the figure, the vertical axis represents the TBV value, and the horizontal axis represents the smoke concentration value. It can be seen from the figure that the value of TBV is also lower when the smoke concentration is lower. This is because the entire environment covers haze, which is equivalent to noise superimposed on a clear video image, and therefore, the characteristic difference between the image boundaries is reduced, resulting in a smaller TBV. When the dense haze becomes light, the process of denoising the video image is actually performed, and when the haze becomes light or disappears, the characteristic difference between the dense haze and the corresponding image boundary is displayed. As can be seen from the figure, the TBV values at this time are all increased, which is also consistent with that shown in fig. 4.
2. Approximation
And (3) based on a smoke concentration estimation algorithm of total variation, regarding the image containing smoke concentration information as a blurred image, and combining the image characteristics to give a brightness curve. Firstly, based on the total variation principle, the convolution and integration processing is carried out on the brightness curve, and then the extinction coefficient k is obtained in an iterative mode. If the extinction coefficient is found to be accurate, the smoke concentration estimation will also be accurate. Then, in order to check whether the solution of the extinction coefficient is accurate, further check is needed, and the method is to continuously perform approximate comparison between the estimated value of the brightness of the object and the observed actual value of the brightness of the object. And when the brightness estimated value of the target object is close to the actual value or the approximation error is small, the obtained extinction coefficient k is considered to be more accurate.
Several target brightness curves were compared by fitting, as shown in fig. 5. As can be seen from fig. 5(a) and 5(b), the approximation effect is better. The estimated value is substantially close to the actual value. The extinction coefficients of the two figures are referenced 0.0257 and 0.0218, respectively, and the experimental values are 0.0263 and 0.0224, respectively. The approximation of fig. 5(c) is less effective than the first two. The mean absolute relative errors of the three plots were 3.72%, 4.36%, and 6.14%, respectively.
As can be seen from fig. 5, when the smoke concentration is low, especially the article algorithm can better approximate the brightness curve of the target object. As the smoke density increases, the approximation error increases due to the increase in the curvature of the luminance curve, and thus the estimation error also increases. The data show a minimum error of 3.42% and a maximum error of 9.29% within 10% of national regulations.
2.3 Smoke Density determination
According to equation (2), for different positions, i.e. k varies constantly, there is a corresponding L, as shown in equation
In the formula, the relevant parameters have the following meanings:
l: the brightness of the target observed by the observation point;
L0: the target has own brightness;
Lf: background sky brightness;
k: extinction coefficient;
d: the observation point-to-target object distance is the distance from the camera to the target object. On the basis of the formula (12), an objective function is established to obtain a formula
The above equation n is the sampling number. The pair of formula (13) has
Where p represents a norm, and in this context, let p equal to 1, the formula is obtained
Deriving from (13) and (15) a cost function of
Further simplified to obtain
According to the optimization algorithm, the free brightness L of the target object is obtained0And extinction coefficient kiI.e. the formula
On the basis, multiple iterations are carried out through a quasi-Newton iteration method, an extinction coefficient can be obtained, and the smoke concentration k is obtained.
Example two:
the embodiment provides a smog concentration detection device based on total variation, the device includes:
a video acquisition module: for acquiring video images;
an image extraction module: the image acquisition module is used for extracting the video image and acquiring an image containing smoke concentration information;
a calculation analysis module: the system is used for carrying out image brightness analysis on the image containing the smoke concentration information, carrying out total variation calculation and solving a total variation (TBV) value of the image;
a concentration acquisition module: and calculating and acquiring an extinction coefficient, namely the smoke concentration according to the total variation (TBV) value of the image.
Example three:
the embodiment of the invention also provides a smoke concentration detection device based on total variation, which comprises a processor and a storage medium;
the storage medium is used for storing instructions;
the processor is configured to operate in accordance with the instructions to perform the steps of the method of:
first, the total variation is solved. The monitoring image of normal weather is clear, the main body is a high-frequency signal, and the boundary difference is large. The image blur degree corresponding to the increase of the smoke density is increased, and the TBV (total variation) value of the video image reflecting the smoke information is reduced. After haze exists, the image becomes fuzzy, namely a large amount of noise appears on a clear image, the thicker the haze is, the more the noise is, and the higher the smoke concentration is. When the smoke concentration caused by haze is higher and higher, the TBV (total variation) value of the monitoring image is smaller and smaller. At lower smoke concentrations, the value of TBV (total variation) is also lower. This is because the entire environment covers haze, which corresponds to noise superimposed on a clear video image, and therefore, the characteristic difference between the image boundaries is reduced, resulting in a smaller TBV (total variation). When the dense haze becomes light, the process of denoising the video image is actually performed, and when the haze becomes light or disappears, the characteristic difference between the dense haze and the corresponding image boundary is displayed.
And secondly, fitting approximation. On the basis of a total variation theory, a relation model between total variation and image ambiguity is constructed for accurately measuring the relation between smoke concentration and total variation and for the criterion of smoke concentration evaluation. Based on the smoke concentration estimation algorithm of total variation, the image containing smoke concentration information is regarded as a fuzzy image, the image characteristics are combined, and then the brightness estimation value of the target object and the observed actual brightness value of the target object are continuously subjected to approximation comparison to obtain a brightness curve.
And thirdly, determining the smoke concentration. On the basis of the property that the total variation value of an image is correspondingly reduced when the image becomes fuzzy, when the smoke concentration is detected, a functional is constructed on the basis of the Koschmieder theory, the variation is solved, an extreme value is solved for an objective function by the idea of piecewise smooth analysis, a brightness curve is subjected to convolution and integration processing on the basis of the total variation principle, and then the extinction coefficient is solved by a quasi-Newton iteration method.
As will be appreciated by one skilled in the art, embodiments of the present application may be provided as a method, system, or computer program product. Accordingly, the present application may take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, the present application may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, and the like) having computer-usable program code embodied therein.
The present application is described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the application. It will be understood that each flow and/or block of the flow diagrams and/or block diagrams, and combinations of flows and/or blocks in the flow diagrams and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instruction means which implement the function specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
The above description is only a preferred embodiment of the present invention, and it should be noted that, for those skilled in the art, several modifications and variations can be made without departing from the technical principle of the present invention, and these modifications and variations should also be regarded as the protection scope of the present invention.
Claims (8)
1. A smoke concentration detection method based on total variation is characterized by comprising the following steps:
acquiring a video image;
extracting the video image to obtain an image containing smoke concentration information;
analyzing the brightness of the image containing the smoke concentration information, calculating the total variation to obtain the total variation (TBV) value of the image,
and calculating and obtaining an extinction coefficient, namely the smoke concentration according to the total variation (TBV) value of the image.
2. The smoke density detection method based on total variation according to claim 1, wherein the method of performing total variation calculation to find the total variation (TBV) value of the image comprises the steps of:
respectively calculating the total variation TBV of the clear image ffAnd the total variation TBV of the blurred image ggRespectively solved by the following formula:
wherein f represents a sharp image, the function f (x, y) represents the (x, y) th pixel in the image, and the pixel of the blurred image is the function g (x, y); where Ω represents a region of the image, and is expressed as follows:
Ω＝{(x,y):0≤x≤m,0≤y≤n}
in the formula, m and n are the sizes of the original clear image.
3. The method according to claim 1, wherein the method for calculating and acquiring the extinction coefficient according to the total variation (TBV) value of the image comprises:
and acquiring a brightness curve of the image according to the total variation (TBV) value of the image, and acquiring an extinction coefficient according to the brightness curve of the image.
4. The smoke density detection method based on total variation as claimed in claim 3, wherein the method for obtaining the brightness curve of the image according to the total variation (TBV) value of the image and obtaining the extinction coefficient according to the brightness curve of the image comprises:
based on a smoke concentration estimation algorithm of total variation, regarding the image containing smoke concentration information as a fuzzy image, combining image characteristics, and then continuously performing approximate comparison on the brightness estimation value of the target object and the observed actual brightness value of the target object to obtain a brightness curve;
and constructing a functional, solving the variation of the functional, solving an extreme value of the objective function, performing convolution and integration processing on the brightness curve based on the total variation principle, and then solving the adjusted extinction coefficient by using a quasi-Newton iteration method.
5. The method for detecting smoke concentration based on total variation as claimed in claim 4, wherein the method for constructing functional and calculating variation thereof, extremizing the objective function, performing convolution and integration processing on the brightness curve based on total variation principle, and then obtaining the adjusted extinction coefficient by quasi-Newton iteration method comprises:
establishing an objective function to obtain a formula:
in the above formula, n is the sampling frequency, Q is the objective function, the variation is taken, the cost function is derived, and the reduction is carried out to obtain
In the formula, L: the brightness of the target observed by the observation point; l is0: the target has own brightness; l isf: background sky brightness; k: extinction coefficient; d: the distance between the observation point and the target object;
according to the optimization algorithm, the free brightness L of the target object is obtained0And extinction coefficient kiNamely the formula:
on the basis, multiple iterations are carried out through a quasi-Newton iteration method, an extinction coefficient can be obtained, and the smoke concentration k is obtained.
6. The smoke concentration detection method based on total variation as claimed in claim 1, wherein the method for extracting the video image and acquiring the image containing the smoke concentration information comprises the following steps:
denoising the video image to obtain a denoised video;
extracting monitoring frames with equal time intervals according to the de-noised video;
and performing ROI (region of interest) region extraction on the monitoring frames with equal time intervals to obtain an image containing smoke concentration information.
7. A smoke concentration detection device based on total variation, the device comprising:
a video acquisition module: for acquiring video images;
an image extraction module: the image acquisition module is used for extracting the video image and acquiring an image containing smoke concentration information;
a calculation analysis module: the system is used for carrying out image brightness analysis on the image containing the smoke concentration information, carrying out total variation calculation and solving a total variation (TBV) value of the image;
a concentration acquisition module: and calculating and acquiring an extinction coefficient, namely the smoke concentration according to the total variation (TBV) value of the image.
8. A smoke concentration detection device based on total variation is characterized by comprising a processor and a storage medium;
the storage medium is used for storing instructions;
the processor is configured to operate in accordance with the instructions to perform the steps of the method according to any one of claims 1 to 6.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110495053.5A CN113222929A (en) | 2021-05-07 | 2021-05-07 | Smoke concentration detection method and device based on total variation |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110495053.5A CN113222929A (en) | 2021-05-07 | 2021-05-07 | Smoke concentration detection method and device based on total variation |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN113222929A true CN113222929A (en) | 2021-08-06 |
Family
ID=77091536
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110495053.5A Withdrawn CN113222929A (en) | 2021-05-07 | 2021-05-07 | Smoke concentration detection method and device based on total variation |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113222929A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114387273A (en) * | 2022-03-24 | 2022-04-22 | 莱芜职业技术学院 | Environmental dust concentration detection method and system based on computer image recognition |
-
2021
- 2021-05-07 CN CN202110495053.5A patent/CN113222929A/en not_active Withdrawn
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114387273A (en) * | 2022-03-24 | 2022-04-22 | 莱芜职业技术学院 | Environmental dust concentration detection method and system based on computer image recognition |
| CN114387273B (en) * | 2022-03-24 | 2022-05-31 | 莱芜职业技术学院 | Environmental dust concentration detection method and system based on computer image recognition |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109743879B (en) | Underground pipe gallery leakage detection method based on dynamic infrared thermography processing | |
| CN109633633B (en) | Life signal enhancement method based on segmented classification enhancement processing | |
| EP2846306B1 (en) | Image processing apparatus for removing haze contained in still image and method thereof | |
| CN109190624B (en) | Kitchen oil smoke concentration detection method based on image processing | |
| WO2020232710A1 (en) | Haze image quality evaluation method and system, storage medium, and electronic device | |
| CN115797641B (en) | Electronic equipment gas leakage detection method | |
| CN108564002B (en) | Method and system for detecting time sequence change of remote sensing image | |
| CN110889843B (en) | SAR image ship target detection method based on maximum stable extremal region | |
| CN109993797A (en) | Door and window method for detecting position and device | |
| CN111489352A (en) | Tunnel gap detection and measurement method and device based on digital image processing | |
| CN115841434A (en) | Infrared image enhancement method for gas concentration analysis | |
| CN114119436A (en) | Infrared image and visible light image fusion method and device, electronic equipment and medium | |
| Fan et al. | Dim small target detection based on high-order cumulant of motion estimation | |
| EP3936854A1 (en) | Automated detection and quantification of gas emissions | |
| CN114596329A (en) | Gas image enhancement and gas leakage detection method and system | |
| CN115564710A (en) | Fire smoke detection method and device based on LK optical flow method and storage medium | |
| CN115343685A (en) | Multi-dimensional ground penetrating radar detection method, device and equipment applied to disease identification | |
| CN117368153A (en) | Road state detection method based on infrared laser and thermal imaging | |
| CN111985314A (en) | ViBe and improved LBP-based smoke detection method | |
| CN113222929A (en) | Smoke concentration detection method and device based on total variation | |
| CN113705380A (en) | Target detection method and device in foggy days, electronic equipment and storage medium | |
| CN117710266A (en) | Night road visibility monitoring method based on video image | |
| CN113705381A (en) | Target detection method and device in foggy days, electronic equipment and storage medium | |
| CN105719252A (en) | Real-time high dynamic range drawing method and device of railway patrol image | |
| CN116137067A (en) | Hyperspectral image anomaly detection method based on coefficient analysis collaborative representation |
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 | ||
| WW01 | Invention patent application withdrawn after publication | ||
| WW01 | Invention patent application withdrawn after publication |
Application publication date: 20210806 |