A smoke detection method based on image processing

Abstract

The present invention discloses a smoke detection method based on image processing, which belongs to the field of image processing technology and includes the following steps: S1, collecting real-time scene images of the detection area, generating a detection threshold, and using the detection threshold to screen a detection particle set from the real-time scene image; S2, determining a smoke detection coefficient according to the detection particle set; S3, determining whether there is smoke in the detection area through the smoke detection coefficient. Compared with the traditional physical sensor detection method, the present invention is based on image processing technology, does not require additional installation of sensor equipment, and saves manpower and material costs.

Description

Technical Field

[0001] The invention belongs to the technical field of image processing, and in particular relates to a smoke detection method based on image processing. Background Art

[0002] With the increasing number of high-rise and underground buildings and large-scale comprehensive buildings in cities, the potential fire hazards have also greatly increased. The number of fires and the losses they cause are increasing year by year. Fires not only cause property losses, but also produce smoke containing harmful gases. Excessive smoke concentration will directly threaten personal safety, and the research on smoke and its concentration detection technology has become a necessary issue. Traditional smoke detection methods mostly rely on physical sensors, such as smoke particle sensors, but they have problems such as limited sensitivity and susceptibility to environmental interference. With the continuous development of image processing technology, image-based smoke detection methods have gradually become a new research direction. Summary of the invention

[0003] In order to solve the above problems, the present invention proposes a smoke detection method based on image processing.

[0004] The technical solution of the present invention is: a smoke detection method based on image processing comprises the following steps:

[0005] S1, collecting real-time scene images of the detection area, generating a detection threshold, and using the detection threshold to screen a detection particle set from the real-time scene image;

[0006] S2. Determine the smoke detection coefficient based on the detected particle set;

[0007] S3. Determine whether there is smoke in the detection area through the smoke detection coefficient.

[0008] Furthermore, S1 includes the following sub-steps:

[0009] S11, collecting a real-time scene image of the detection area, and calculating a color attenuation value of the real-time scene image;

[0010] S12, generating edge sharpness of a real-time scene image;

[0011] S13, taking the pixel value variance of all pixels in the real-time scene image as the overall deviation;

[0012] S14, calculating a detection threshold according to the edge sharpness and overall deviation of the real-time scene image;

[0013] S15, determining whether the number of pixel points whose pixel values ​​in the real-time scene image are greater than the detection threshold is greater than or equal to the attenuation value, if so, taking the pixel points whose pixel values ​​are greater than the detection threshold as the detection particle set, otherwise proceeding to S16;

[0014] S16. Pixel points whose pixel values ​​are greater than twice the overall deviation are taken as the detection particle set.

[0015] The beneficial effect of the above further scheme is that in the present invention, the edge sharpness and the overall deviation value are generated by using the Brenner function and the pixel value variance, and the detection threshold for selecting the detection particle set is generated according to these two values. The pixels included in the screened detection particle set may have pixel blur, which is used to determine the smoke detection coefficient.

[0016] Furthermore, in S11, the calculation formula of the color attenuation value S of the real-time scene image is:

[0017] ;

[0018] ; In the formula, ρ n represents the attenuation weight of the nth pixel in the real-time scene image, N represents the number of all pixels in the real-time scene image, k n Represents the pixel value of the nth pixel in the real-time scene image, Indicates a floor operation.

[0019] Furthermore, in S12, the edge sharpness of the real-time scene image is calculated using the Brenner function.

[0020] The beneficial effect of the above further scheme is that in the present invention, the Brenner function is mainly used to calculate the clarity or sharpness of an image, which is calculated based on the square of the difference in grayscale values ​​of adjacent pixels in the image (especially at a certain distance, such as 2 pixels).

[0021] Since the edge of smoke gradually merges with the background and the edge information is not obvious, the existing Brenner function is used to calculate the edge sharpness.

[0022] Furthermore, in S14, the calculation formula of the detection threshold y is:

[0023] ; Where B represents edge sharpness, P represents overall deviation, max(·) represents maximum value calculation, and exp (·) represents exponential calculation.

[0024] Furthermore, S2 includes the following sub-steps:

[0025] S21, taking the number of detected particle sets as the kernel size of the Gaussian function;

[0026] S22. Based on the kernel size of the Gaussian function, calculate the smoke detection coefficient corresponding to the detection particle set.

[0027] The beneficial effect of the above further solution is: in the present invention, Gaussian blur can reduce high-frequency noise in the image and can evenly blur the details in the image, so the calculated smoke detection coefficient is more accurate.

[0028] Furthermore, in S22, the calculation formula of the smoke detection coefficient J is:

[0029] ; In the formula, σ represents the standard deviation of all particles in the detection particle set, c represents the kernel size of the Gaussian function, P1 represents the value of the first particle in the detection particle set, and P i represents the value of the i-th particle in the detection particle set, P I It represents the value of the Ith particle in the detected particle set, exp(·) represents exponential calculation, and min(·) represents minimum value calculation.

[0030] All particles have corresponding pixel values ​​that can be used to calculate the standard deviation.

[0031] Further, in S3, it is determined whether there is a pixel point in the real-time scene image whose pixel value is greater than or equal to the smoke detection coefficient. If so, smoke exists in the detection area, otherwise, smoke does not exist in the detection area.

[0032] The beneficial effects of the present invention are as follows: the present invention analyzes the edge sharpness and overall deviation of the real-time scene image of the detection area to obtain a detection threshold, and uses the detection threshold to generate a detection particle set; and then uses the scale of the detection particle set to adjust the characteristic parameters of the Gaussian function to improve the detection accuracy. Compared with the traditional physical sensor detection method, the present invention is based on image processing technology, does not require additional installation of sensor equipment, and saves manpower and material costs. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] Figure 1 Flowchart of smoke detection method based on image processing. DETAILED DESCRIPTION

[0034] The embodiments of the present invention will be further described below in conjunction with the accompanying drawings.

[0035] like Figure 1 As shown, the present invention provides a smoke detection method based on image processing, comprising the following steps:

[0036] S1, collecting real-time scene images of the detection area, generating a detection threshold, and using the detection threshold to screen a detection particle set from the real-time scene image;

[0037] S2. Determine the smoke detection coefficient based on the detected particle set;

[0038] S3. Determine whether there is smoke in the detection area through the smoke detection coefficient.

[0039] In this embodiment of the present invention, S1 includes the following sub-steps:

[0040] S11, collecting a real-time scene image of the detection area, and calculating a color attenuation value of the real-time scene image;

[0041] S12, generating edge sharpness of a real-time scene image;

[0042] S13, taking the pixel value variance of all pixels in the real-time scene image as the overall deviation;

[0043] S14, calculating a detection threshold according to the edge sharpness and overall deviation of the real-time scene image;

[0044] S15, determining whether the number of pixel points whose pixel values ​​in the real-time scene image are greater than the detection threshold is greater than or equal to the attenuation value, if so, taking the pixel points whose pixel values ​​are greater than the detection threshold as the detection particle set, otherwise proceeding to S16;

[0045] S16. Pixel points whose pixel values ​​are greater than twice the overall deviation are taken as the detection particle set.

[0046] In the present invention, the edge sharpness and the overall deviation value are generated by using the Brenner function and the pixel value variance, and the detection threshold for selecting the detection particle set is generated according to these two values. The pixels contained in the screened detection particle set may have pixel blur, which is used to determine the smoke detection coefficient.

[0047] In the embodiment of the present invention, in S11, the calculation formula of the color attenuation value S of the real-time scene image is:

[0048] ;

[0049] ; In the formula, ρ n represents the attenuation weight of the nth pixel in the real-time scene image, N represents the number of all pixels in the real-time scene image, k n Represents the pixel value of the nth pixel in the real-time scene image, Indicates a floor operation.

[0050] In the embodiment of the present invention, in S12, the edge sharpness of the real-time scene image is calculated using the Brenner function.

[0051] In the present invention, the Brenner function is mainly used to calculate the clarity or sharpness of an image, and is calculated based on the square of the difference in grayscale values ​​of adjacent pixels in the image (especially those separated by a certain distance, such as 2 pixels).

[0052] Since the edge of smoke gradually merges with the background and the edge information is not obvious, the existing Brenner function is used to calculate the edge sharpness.

[0053] In the embodiment of the present invention, in S14, the calculation formula of the detection threshold y is:

[0054] ; Where B represents edge sharpness, P represents overall deviation, max(·) represents maximum value calculation, and exp (·) represents exponential calculation.

[0055] In this embodiment of the present invention, S2 includes the following sub-steps:

[0056] S21, taking the number of detected particle sets as the kernel size of the Gaussian function;

[0057] S22. Based on the kernel size of the Gaussian function, calculate the smoke detection coefficient corresponding to the detection particle set.

[0058] In the present invention, Gaussian blurring can reduce high-frequency noise in an image and can evenly blur details in an image, so the calculated smoke detection coefficient is more accurate.

[0059] In the embodiment of the present invention, in S22, the calculation formula of the smoke detection coefficient J is:

[0060] ; In the formula, σ represents the standard deviation of all particles in the detection particle set, c represents the kernel size of the Gaussian function, P1 represents the value of the first particle in the detection particle set, and P i represents the value of the i-th particle in the detection particle set, P I It represents the value of the Ith particle in the detected particle set, exp(·) represents exponential calculation, and min(·) represents minimum value calculation.

[0061] All particles have corresponding pixel values ​​that can be used to calculate the standard deviation.

[0062] In the embodiment of the present invention, in S3, it is determined whether there are pixels in the real-time scene image whose pixel values ​​are greater than or equal to the smoke detection coefficient. If so, smoke exists in the detection area; otherwise, smoke does not exist in the detection area.

[0063] Those skilled in the art will appreciate that the embodiments described herein are intended to help readers understand the principles of the present invention, and should be understood that the protection scope of the present invention is not limited to such specific statements and embodiments. Those skilled in the art can make various other specific variations and combinations that do not deviate from the essence of the present invention based on the technical revelations disclosed by the present invention, and these variations and combinations are still within the protection scope of the present invention.

Claims

1. A smoke detection method based on image processing, characterized in that: The following steps are involved: S1, collecting real-time scene images of the detection area, generating a detection threshold, and using the detection threshold to screen a detection particle set from the real-time scene image; S2. Determine the smoke detection coefficient based on the detected particle set; S3, determining whether there is smoke in the detection area through the smoke detection coefficient; The S1 comprises the following sub-steps: S11, collecting a real-time scene image of the detection area, and calculating a color attenuation value of the real-time scene image; S12, generating edge sharpness of a real-time scene image; S13, taking the pixel value variance of all pixels in the real-time scene image as the overall deviation; S14, calculating a detection threshold according to the edge sharpness and overall deviation of the real-time scene image; S15, determining whether the number of pixel points whose pixel values ​​in the real-time scene image are greater than the detection threshold is greater than or equal to the attenuation value, if so, taking the pixel points whose pixel values ​​are greater than the detection threshold as the detection particle set, otherwise proceeding to S16; S16, taking the pixel points whose pixel values ​​are greater than twice the overall deviation as the detection particle set; In S11, the calculation formula of the color attenuation value S of the real-time scene image is: ; ; In the formula, ρ n represents the attenuation weight of the nth pixel in the real-time scene image, N represents the number of all pixels in the real-time scene image, k n Represents the pixel value of the nth pixel in the real-time scene image, Indicates a round-down operation; In S12, the edge sharpness of the real-time scene image is calculated using the Brenner function; In S14, the calculation formula of the detection threshold y is: ; Wherein, B represents edge sharpness, P represents overall deviation, max(·) represents maximum value calculation, and exp (·) represents exponential calculation; The S2 comprises the following sub-steps: S21, taking the number of detected particle sets as the kernel size of the Gaussian function; S22, calculating the smoke detection coefficient corresponding to the detection particle set based on the kernel size of the Gaussian function; In S22, the calculation formula of the smoke detection coefficient J is: ; In the formula, σ represents the standard deviation of all particles in the detection particle set, c represents the kernel size of the Gaussian function, P1 represents the value of the first particle in the detection particle set, and P i represents the value of the i-th particle in the detection particle set, P I It represents the value of the Ith particle in the detected particle set, exp(·) represents exponential calculation, and min(·) represents minimum value calculation.

2. The smoke detection method based on image processing according to claim 1, characterized in that: In S3, it is determined whether there is a pixel point whose pixel value is greater than or equal to the smoke detection coefficient in the real-time scene image. If so, smoke exists in the detection area; otherwise, smoke does not exist in the detection area.

Images