No longer afraid of PM2.5 evaluation of real-time video surveillance fogging technology

Tag: Video fogging Security monitoring

This article refers to the address: http://

The effect of fog on video surveillance

Recently, PM2.5, a professional vocabulary in the meteorological field, has become a popular topic of social concern. Droplets in the air and small solid particles not only endanger human health, cause smog and lead to frequent traffic accidents, but also significantly reduce the quality of outdoor surveillance video. In the smoggy weather, the color of the image is dim, the contrast is low, and the details of some important targets are submerged in the fog, and the practicality of the video surveillance system is greatly affected.

Therefore, removing the fog in the video and improving the image quality has become a key technology to enhance the application value of the outdoor video surveillance system.

First, the impact of fog on video surveillance

It is generally believed that the atmospheric medium consists mainly of air molecules, water vapor and aerosols. Aerosols are dispersions of small particles suspended in a gas. Some particles have high hygroscopicity and act as a center of condensation, and their size is related to the relative humidity of the environment, the supply of water vapor, and the degree of agglomeration by collision. Due to the size, type, and degree of aggregation of various particles in the atmosphere, various weather conditions such as Clearness, Haze, Fog, Cloud, and Rain are generated.

Under clear weather conditions, the light reflected from the surface of the object is not affected by various components in the atmosphere, such as scattering, absorption, reflection, etc., and can directly reach the imaging device to obtain a clear and fog-free image.

In foggy weather, light reflected from the surface of the object is affected by suspended particles in the air as it reaches the imaging device. Aerosol particles are the main factor in the formation of haze and the root cause of image quality degradation. The main effects are as follows:

(1) The aerosol particles have a scattering effect on the light, and the scattering loss attenuates the intensity of the "transmitted light", resulting in a decrease in the contrast of the image.

(2) Due to the non-uniformity of the aerosol particles, the spherical wave is distorted into an aspherical wave, resulting in blurred images, reduced edges and details.

(3) The particle size of the aerosol particles is large, and the self-imaging of the particles cannot be ignored, and can be approximated as "noise".

(4) The scattering part of the aerosol by the aerosol particles will be superimposed with the original forward scattering part due to the multiple scattering effect, resulting in a certain blur.

Second, the comparative analysis of real-time video fogging technology and other fog-passing techniques

Currently known fog-transparent algorithms can be roughly divided into two categories: one is a non-model image enhancement method, which enhances the contrast of the image to meet the requirements of subjective vision to achieve clarity; the other is model-based image The restoration method, which examines the cause of image degradation, models the degradation process, and uses reverse processing to finally solve the image restoration problem.

Video monitoring system through fog technology formula

Typical methods for performing transmissive processing in an enhanced manner include: histogram equalization, filter transform methods, and fuzzy logic based methods. The histogram equalization method, in which the globalization method has a small amount of computation but insufficient enhancement of details; the local equalization method works well, but may introduce a block effect, a large amount of calculation, a noise is amplified, and the algorithm effect is difficult to control. The filter-transformed fog-transparent algorithm can obtain relatively good processing results through local processing, but they are computationally intensive, resource-intensive, and unsuitable for devices with high real-time requirements. The effect of fogging based on fuzzy logic is not ideal.

The enhancement-based method can improve the image contrast to some extent and enhance the recognizability by enhancing the region of interest. However, this method fails to compensate for the cause of the image degradation process, so it can only improve the visual effect without obtaining a good fog effect.

At present, the methods based on image restoration mainly include the following methods: filtering method, maximum entropy method and image degradation function estimation method. The filtering method, such as the Kalman filtering method, has a large amount of calculation as a whole. The maximum entropy method can obtain higher resolution but its nonlinearity, large amount of calculation, and difficulty in numerical solution.

Image degradation function estimation method is mostly designed according to certain physical models (such as atmospheric scattering model and polarization model of fog permeability). It is necessary to collect multiple images as reference images at different time points in order to determine multiple parameters in the physical model. And finally solved to obtain the resulting image in the fog-free state. This limits the application of such methods in real-time monitoring.

Security products are now used in a variety of complex scenarios, bad weather, all-weather real-time monitoring of the product portability and power consumption, processing effects, processing adaptability and other aspects have put forward more stringent requirements. Good video fogging technology should combine the advantages of image enhancement and image restoration based on the atmospheric transmission model, so that it can obtain ideal image effects and be actually engineered.

In image processing, the following model is generally used to express the foggy image seen:

I(X)=J(X)t(x)+A(1-t(x))

I represents the intensity of the image seen, J is the intensity of the scene, A is the component of the atmospheric light, and t is the part that describes the light that is not scattered through the medium. The goal of fogging is to recover J, A, t, and J from I, which corresponds to the resulting image after fogging. Among them, J(X)t(x) is called the direct attenuation term, which represents the part of the scene where the light is attenuated in the medium. A(1-t(x) is the atmospheric light component, which is caused by the forward scattering.

After fully analyzing the advantages and disadvantages of the fogging theory and conducting in-depth research and exploration, Hikvision (002415, shares it) combined with the special requirements of video image fogging in the field of security monitoring, developed a real-time video fogging technology. . The technology is based on the principle of atmospheric optics, which distinguishes the depth of field and the fog concentration of different regions of the image for filtering, and obtains accurate and natural fog-transparent images.

The real-time video fog-transmission technology can automatically adjust according to the change of the fog condition to adapt to various scene applications, avoiding the situation that the near-field fog is excessively black and the foreground is blurred; at the same time, the efficiency and complexity of the realization are ensured, and the entire fog is ensured. Real-time and engineerable.

Traditional dehazing method and real-time fog contrast

At the same time, this real-time video fogging technology can not only effectively remove the effects of fog, but also avoid the color error caused by the transition processing of some scenes, and the unreality caused by excessive mist removal. As shown in Figure 2-1, where a is the image containing the mist, b is the effect of the traditional fogging method, and c is the effect of the real-time fogging technique. It can be seen that since the fog itself is not thick, due to the lack of self-adaptive ability in the conventional fogging method b, blackening and color distortion appear in the red frame, and the image effect in the red frame in c is natural and true.

Comparison of effects between traditional dehazing methods and real-time fog-through technology (left to right a to c)

Comparison of traditional defogging methods with real-time fog-through technology (left to right a to c)

The advantages of real-time video fogging technology compared with other fog-through methods are mainly reflected in the following aspects:

1) Strong fogging ability. Real-time video fogging technology can accurately remove the corresponding degree of fog according to different depth of field. The traditional image enhancement method may have a better near-field fogging effect, while the distant view still has a lot of fogging problems; the fog-removing effect based on the image restoration method is related to the accuracy of the selected component, and the component selection is accurate and the fog-transparent effect is obtained. Preferably, if the component selection is inaccurate, poor results may occur, so the performance of the method of the fog-through algorithm is not stable enough.

2) Good permeability. The image processed by real-time video fogging technology is transparent and has good contrast. Other methods leave fog due to inaccurate estimation of depth of field, making the image after fogging not very transparent.

3) The level of detail retention is high. The real-time video fog-passing technology contains special processing to keep the details, so the fog-transparent image can retain or even partially enhance the details hidden in the fog, which is difficult to achieve by other methods.

4) The color saturation is high and the reduction ability is strong. Real-time video fogging technology does not change the color tone of the image but merely increases the saturation of its color. Other fogging methods may cause color distortion.

5) Does not cause image darkening. Real-time video fogging technology does not appear to be dark as a result of brightness enhancement performance. Other methods may cause problems with reduced contrast.

6) A wide range of applications. Real-time video fogging technology can also be used to process fog-free images, improve the contrast and saturation of the original image, and at the same time improve the transparency of the image and enhance the visual quality of the image.

Third, the application prospect of real-time video fogging technology in the field of security monitoring

Real-time video fogging technology can improve the quality of video surveillance from multiple angles. First of all, it is a fog-through technology that can be used for aerosol treatment of various weather conditions caused by aerosols; it is also an enhancement algorithm that can significantly improve the contrast of images, make images transparent and clear; Enhance the details of the image, so that the original hidden image details are fully displayed; it can enhance the saturation of the image, make the image colorful and lively, vivid, and the image after fogging maintains accurate color tone and natural appearance, thus obtaining Good image quality and visual experience.

Real-time video fogging technology has strong engineering capabilities. The fog-through technology can be applied to various resolutions including megapixel-level high-definition images; real-time accurate fog-through processing can be guaranteed at various resolutions; current fog concentration can be evaluated in real time according to the target scene and adaptively adjusted. The fog intensity changes without fear of the scene; since the fogging technique can obtain accurate depth of field information, the non-uniform mist in the same scene can be accurately removed without leaving the fog. From the current test results, this technology has a good application prospect in outdoor video surveillance systems.

The real-time video fog-passing function has been initially implemented on some cameras, and the actual fog-transparent processing results are shown in Figures 2-2 and 2-3. It is the two sides of an image, the left side is the image after fogging, and the right side is the original image without fogging. It can be clearly seen that the left image is significantly improved in terms of details, transparency, color, and the like.

Real-time video through fog image contrast with the original image

Real-time video through fog image contrast with the original image

Real-time video through fog image contrast with the original image

Real-time video through fog image contrast with the original image

Real-time fog-through technology can be combined with video compression and intelligent analysis technology to bring greater value to it. Since the current mainstream video compression algorithms are lossy compression, it will cause damage to the lower contrast details in the image, while the foggy video generally has low contrast and less detail, so it is often blurred and unclear after being encoded and compressed. restore. Real-time through-fog technology can effectively enhance image contrast and detail, ensure that valuable information is not lost by encoding compression, and significantly improve information efficiency. For intelligent analysis, the image processed by real-time fog-transparent technology can significantly reduce the error rate of the analysis result, especially the false negative rate, thus greatly improving the practicality of the intelligent analysis system.

With the development of industry and its impact on the climate, smog has become a common weather phenomenon, which has a great impact on the picture quality of outdoor application monitoring systems. At present, in actual engineering, radar, infrared and other means are used to sense and monitor targets in the haze climate. Due to technical and cost constraints, radar, infrared and other technologies have not been ideally suited in the field of video surveillance. The emergence of real-time video fogging technology can bring many aspects to the field of video surveillance.

From the application scenario, real-time video fogging technology can be used in a variety of outdoor applications, such as highway traffic accident-prone area, highway bayonet location; bus auxiliary driving facility monitoring area; public security organs focus on Locations and areas; key monitoring areas for power plants and power transmission equipment; primary and secondary schools, urban commercial centers and city squares. From the perspective of applied industries, including transportation industry, public security industry, education, aviation, digital products, remote sensing image processing, food safety monitoring, and even special military applications. From the perspective of applied products and solutions, real-time video fogging technology can be applied to the front camera, fast ball fog and image quality in the field of security monitoring; it can be applied to DVR to improve image quality; Large-screen display to enhance its color saturation and image quality; it can also be used in embedded client software to improve the quality of preview images.

Circuit Board

Circuit Board Parts,Circuit Board,Printed Circuit Board,Socket Circuit Board

Dongguan baiyou electronic co.,ltd , https://www.dgbaiyou.com