CN115396570A - A low-light high-temperature industrial endoscope - Google Patents

Abstract

The invention discloses a low-light high-temperature industrial endoscope, which includes a shell assembly, a low-light level optical imaging system installed in the shell component, and an image brightness processing module, wherein: the low-light level optical imaging system is used for collecting Image video stream; image brightness processing module, which is connected with the Youguang level optical imaging system, used to receive the furnace image video stream output by the Youguang level optical imaging system, and perform brightness processing on the furnace image video stream to obtain high brightness Clear furnace image video stream, the invention solves the technical problems of insufficient brightness, fuzzy details, and unclear outlines of images in industrial furnaces under extremely weak light conditions, and the designed fidelity function, consistency function and The regularization function utilizes the strong correlation between the optimized image and the original image, achieves the purpose of eliminating ghosting in the enhanced image while ensuring the brightness of the enhanced image, and effectively improves the image quality of the enhanced image.

Description

A low-light high-temperature industrial endoscope

technical field

The invention mainly relates to the technical field of high-temperature industrial endoscopes, in particular to a low-light high-temperature industrial endoscope.

Background technique

The acquisition of internal information of industrial kiln is the key to realize its precise control, ensure its efficient operation, and achieve the goal of reducing consumption and emission. The high temperature industrial endoscope is the core equipment that can directly obtain the internal information of the industrial furnace. However, the harsh environment of high temperature and strong corrosion inside the industrial furnace places extremely strict requirements on the temperature resistance and corrosion resistance of the endoscope. At the same time, the extremely weak light environment inside the industrial kiln (that is, the dim light environment with an illuminance lower than 0.0001 Lux) causes problems such as missing details, blurred outlines, and poor clarity in images captured by the endoscope. This will affect the operator's accurate judgment on the working status of the kiln, and it will be impossible to select appropriate means to control the kiln in time, resulting in fluctuations in furnace conditions, equipment failures, and ultimately huge economic losses. The brand-new low-light high-temperature industrial endoscope can withstand high temperature and strong corrosion environment, overcome the extremely weak light environment inside the industrial kiln (illuminance is lower than 0.0001Lux), realize low-light level imaging, and provide operators with high-quality The imaging in the furnace clearly shows the internal operation of the furnace, and provides an operating basis for the adjustment of the operation status of the industrial furnace to avoid accidents.

The imaging endoscope equipment in the kiln mainly includes: industrial endoscope without light source and industrial endoscope with light source. Light source industrial endoscopes are divided into retractable and non-retractable types. Now explain one by one as follows.

Industrial endoscope without light source: Generally, an optical camera is directly installed at the front end of the endoscope, and the object under test is photographed by using ambient light. The collected image information is transmitted from the tail of the endoscope through the signal line, and is externally connected to the host computer for display. This method has clear imaging and simple structure, and can increase the shooting angle and image width by designing the endoscope tube body with movable wide-angle components and angle adjustment units, and select an angle with a clear image to shoot and image. However, because it does not provide a light source and does not design any image brightness enhancement algorithm, it is difficult to capture clear and bright images when the ambient light is insufficient. At the same time, the camera is installed directly at the front end of the endoscope. Even if a cooling system is designed, the high temperature close to the center of the furnace will still burn the camera at the front end, making it unable to operate normally. Therefore, this equipment is not suitable for industrial kilns with high temperature environments Shot inside the furnace.

Retractable industrial endoscope with light source: through the design of the retractable device, it only extends into the detection device when shooting is required. This kind of equipment, on the one hand, prevents the camera from being in the harsh environment of the furnace for a long time and prolongs the service life of the endoscope; on the other hand, it reduces the pollution of the camera lens by the dust in the kiln to ensure the clarity of imaging. However, this kind of equipment cannot provide continuous and real-time images in the furnace, and it is difficult to make timely adjustments according to the operating status of the industrial furnace, and there are certain potential safety hazards.

Non-retractable industrial endoscope with light source: using multiple LED lamp beads, distributed in a circular array around the body of the endoscope, providing high-brightness illumination for the endoscope, which is convenient for highlighting high-definition shooting of dark or weak light objects . However, because the LED lamp beads are distributed in a circle, the light is relatively scattered, and the light loss is serious during the irradiation process, which cannot provide sufficient brightness for shooting inside the kiln in a dim light environment. At the same time, the LED lamp bead array is located at the front of the endoscope, which is easily damaged in the high temperature environment inside the industrial kiln for a long time. Therefore, this equipment is not suitable for shooting inside industrial kilns under high temperature and low light environment.

The utility model patent whose publication number is CN213338205U is a kind of endoscopic camera which is convenient to observe and image clearly. In order to obtain clear and stable imaging of the endoscope, its equipment is composed of components such as motors, sliders, fixed plates, adjustment rods and adjustment rods. Its main function and realization principle are divided into two parts, one part is to use components such as slider, slide rail, first motor, fixing plate, support rod and adjustment rod to realize the fixation of the endoscope main body inside the detected object; the other part The camera angle is finely adjusted by using components such as the second motor, an angle lever, a rotating tooth, and an adjusting tooth. The specific working process of the equipment is as follows: firstly, the first motor drives the threaded rod to rotate, so that the endoscope slider slides along the slide rail, so that the fixed plate, support rod and adjustment rod that were originally attached to the outside of the main body are against the center of the detection area. On the inner side, the entire body of the endoscope is fixed to obtain a stable and clear image. Then, turn on the second motor to drive the camera to rotate through the angle lever to realize the adjustment of the camera angle. This patent provides a new form of device fixing, and after fixing, it indirectly guarantees that the acquisition of the image will not be affected by the outside, and a clear image can be stably obtained. However, this equipment lacks cooling components and cannot withstand the harsh environment of high temperature and high pressure inside the industrial furnace. At the same time, due to the existence of movable components such as slide rails and sliders, the airtightness of the industrial furnace cannot be guaranteed under long-term work, and it has a certain degree of safety. hidden danger, so the device cannot be applied to industrial kilns.

The utility model patent whose publication number is CN213399057U is a kind of high-brightness industrial endoscope for viewing large spaces. In this patent, the array of LED lamp beads is placed around the lens of the endoscope in a surrounding arrangement to provide illumination for the shooting of the endoscope and improve the imaging brightness. At the same time, the patent also designs a corresponding heat dissipation structure for the LED lamp bead array to prevent the overheating of the LED from affecting the normal operation of the endoscope. This patent provides a new idea of using LED lamp bead array as the illumination source of endoscope to finally improve the imaging effect. The provided lighting has scattered light and severe light loss, which cannot provide bright and effective lighting for clear shooting inside the industrial kiln. Therefore, this device is not suitable for image acquisition inside industrial kilns in low-light environments.

The invention patent with the publication number CN114026495A is a kind of endoscope that can finely adjust the shooting angle. An angle adjustment unit realizes the rotation of the endoscope camera at any angle and maximizes the detection range. At the same time, the device is designed with an air supply pipe to prevent the endoscope camera from being damaged due to high temperature. However, this device only uses the movable wide-angle components and angle adjustment unit to increase the viewing angle of the endoscope and improve the width of the picture. It does not perform additional brightness processing on the image, and still cannot obtain a clear picture of the interior of the industrial kiln under insufficient light. .

Contents of the invention

The low-light high-temperature industrial endoscope provided by the present invention solves the technical problems of insufficient brightness, fuzzy details, and unclear outlines of images acquired by existing industrial endoscopes in high-temperature airtight, low-illumination industrial kilns.

In order to solve the above-mentioned technical problems, the low-light high-temperature industrial endoscope proposed by the present invention includes a housing assembly, a low-light optical imaging system installed in the housing assembly, and an image brightness processing module, wherein:

Youguang level optical imaging system, used to collect images and video streams in the furnace;

The image brightness processing module is connected with the Youguang level optical imaging system, and is used to receive the furnace image video stream output by the Youguang level optical imaging system, and perform brightness processing on the furnace image video stream, so as to obtain a bright and clear furnace image. Image video stream.

Further, the shell assembly includes a front shell assembly and a rear shell assembly, wherein:

The front shell assembly has a double-layer structure, and the outer layer in the double-layer structure is an air-cooled circulation channel, and the outer side of the inner layer is a double-helical winding water-cooled circulation flow channel;

The rear shell assembly is used to install the CCD imaging chip, the imaging drive circuit and the image brightness processing module, and the rear shell assembly is provided with an air-cooled air inlet, a water-cooled water inlet and a water-cooled water outlet.

Furthermore, the Youguang high temperature industrial endoscope also includes a cooling system, in which:

The cooling system adopts an air-cooled-water-cooled dual cooling system, and the cold air in the air-cooled-water-cooled dual cooling system enters the air-cooled circulation channel from the air-cooled air inlet, and the gas circulates in the air-cooled circulation channel to form an air-cooled effect;

The water flow in the air-cooled-water-cooled dual cooling system enters the water-cooled circulation flow channel from the water-cooled water inlet, and is discharged at the water-cooled water outlet. The discharged water can flow in again after being processed, and thus circulates to form a water-cooling effect.

Further, the low-level optical imaging system includes sequentially connected imaging objective lens group, relay lens group, focusing lens group and CCD imaging chip, wherein:

The imaging objective lens group includes a negative focal power lens group and a positive focal power lens group, and the negative focal power lens group is used to scatter the incident light, and uniformly scatter the collected optical information with a large field of view and a large amount of incoming light to the positive focal power lens The group provides initial conditions for the long-distance parallel transmission of optical information, and the positive power lens group is used to re-gather the scattered light to form a beam parallel to the optical axis;

The relay lens group is used to transmit the image in the furnace obtained by the imaging objective lens group to the CCD imaging chip;

The focus lens group is used to adjust the focus distance when the CCD imaging chip is blurred, so as to ensure clear imaging;

The CCD imaging chip is used to convert the clear image transmitted by the focusing lens group into a digital image.

Further, the image brightness processing module includes an input processing unit and an FPGA video stream image processing unit and an output processing unit sequentially connected to the input processing unit, wherein:

The input processing unit is used to decode the video stream of the furnace image output by the Youguang optical imaging system, and store the decoded furnace image video stream in the DDR SDRAM storage module;

The FPGA video stream image processing unit is used to perform image preprocessing and brightness synthesis on the decoded furnace image video stream, specifically including a MicroBlaze processing module and a signal processing task trigger module connected to the MicroBlaze processing module, a preset processing task module and DDR SDRAM memory module, where:

The signal processing task trigger module is used to judge whether the number of image frames in the furnace image video stream meets the algorithm requirements in the preset processing task module according to the furnace image video stream stored in the DDR SDRAM storage module, and send tasks to the MicroBlaze processing module command to realize task triggering, task configuration and task management;

The preset processing task module is used to pre-store the algorithms of image preprocessing and brightness synthesis processing tasks;

The MicroBlaze processing module is used to receive the task trigger, task configuration and task management commands sent by the signal processing task trigger module, and according to the pre-stored image preprocessing and brightness synthesis processing task algorithms in the preset processing task module, the decoded Image preprocessing and brightness synthesis of the image and video stream in the furnace;

The DDR SDRAM storage module is used to store the decoded image video stream of the input processing unit and perform data interaction with the signal processing task trigger module and the MicroBlaze processing module;

The output processing unit is used to output the image video stream in the furnace processed by the FPGA video stream image processing unit.

Further, the preset processing task module includes sequentially connected three-dimensional digital comb filter algorithm submodule, video highlight synthesis algorithm submodule based on temporal and spatial domain adaptation, and white balance algorithm submodule based on multi-color dynamic domain algorithm, wherein:

The three-dimensional digital comb filter algorithm sub-module is used to perform three-dimensional digital comb filter on the decoded furnace video stream;

The video highlight synthesis algorithm sub-module based on time-space domain adaptation is used to perform video highlight synthesis on the furnace video stream after the three-dimensional digital comb filter;

The white balance algorithm sub-module based on the multi-color dynamic domain algorithm is used to perform white balance processing on the video stream in the furnace after video brightness synthesis.

Further, the video highlight synthesis algorithm submodule based on temporal and spatial domain adaptation includes a sequentially connected video frame acquisition submodule, nonlinear transformation submodule, superimposition submodule and composite video frame acquisition submodule, wherein:

The video frame acquisition sub-module is used to collect video frames in the video stream after the three-dimensional comb filtering, and the video frames include the current video frame and the video frames adjacent to the current video frame;

The non-linear transformation sub-module is used for non-linear transformation of the video frame;

The overlay submodule is used to overlay the non-linearly transformed video frames to obtain an enhanced image;

The composite video frame acquisition sub-module is used to optimize the enhanced image using the maximum a posteriori model to obtain a composite video frame, thereby obtaining a highlighted video stream.

Further, the composite video frame acquisition submodule includes a fidelity function calculation submodule, a consistency function calculation submodule, a regularization function calculation submodule, a posterior probability calculation submodule and a composite video frame calculation submodule connected in sequence, wherein :

The fidelity function calculation sub-module is used to calculate the fidelity function, wherein the calculation formula of the fidelity function is:

为O_t和Y_j的联合分布的方差，B和D分别表示模糊矩阵和下采样矩阵，S表示协方差矩阵，

表示t时刻的优化图像O_t到j时刻的增强图像Y_j的运动补偿矩阵；Among them, ψ(O _t , Y _j ) is a fidelity function, which is used to measure the similarity between the optimized image O _t at time t and the enhanced image Y _j at time j,

is the variance of the joint distribution of O _t and Y _j , B and D represent the fuzzy matrix and the downsampling matrix respectively, S represents the covariance matrix,

Represent the motion compensation matrix of the optimized image O _t at moment t to the enhanced image Y _j at moment j;

The consistency function calculation sub-module is used to calculate the consistency function, where the calculation formula of the consistency function is:

表示t-1时刻的输出优化图像，

表示一致性函数，衡量t时刻的优化图像O_t和t-1时刻的输出优化图像

之间的相似性，

是O_t和

的联合分布的方差，

是t时刻优化图像O_t到t-1时刻的输出优化图像

的运动补偿矩阵；in,

Indicates the output optimized image at time t-1,

Represents the consistency function, which measures the optimized image O _t at time t and the output optimized image at time t-1

the similarity between

is O _t and

The variance of the joint distribution of ,

is the optimized image O at time _t and the output optimized image at time t-1

The motion compensation matrix;

The regularization function calculation sub-module is used to calculate the regularization function, wherein the calculation formula of the regularization function is:

Among them, L(O _t ) is the regularization function, P is the size of the moving window, α is used to constrain the smoothness of the optimized image, H _l and V _j indicate that the optimized image O _t at time t is moved in the horizontal and vertical directions respectively Operators for l and j pixels;

The posterior probability calculation submodule is used to obtain the posterior probability of the optimized image O _t according to the fidelity function, consistency function and regularization function, and the calculation formula of the posterior probability is:

Among them, p(O _t ) is the posterior probability of the optimized image O _t , r is the video frame number of the original video used to output the optimized image before time t, and b is the frame number of the original video used to output the optimized image after time t Video image frame number;

The composite video frame calculation sub-module is used to obtain the composite video frame according to the posterior probability of the output optimization image, and the specific calculation formula is:

Further, the white balance algorithm submodule includes a space conversion submodule, an absolute deviation mean calculation submodule, an area to be processed acquisition submodule, an average and deviation calculation submodule, a candidate white point acquisition submodule, and a fusion submodule connected in sequence, wherein :

The space conversion sub-module is used to convert the image frames in the highlighted video stream from the RGB space to the HSV space and the YCbCr space respectively, and obtain the H channel, the S channel, the V channel, the Y channel, the Cb channel and the Cr channel;

The absolute deviation mean calculation sub-module is used to divide the image frame into a preset number of regions, and calculate the absolute deviation mean of the H channel, the S channel, the Cb channel and the Cr channel of each region;

The area to be processed acquisition sub-module is used to obtain the area to be processed that requires white balance processing according to the absolute deviation mean value of the H channel, the S channel, the Cb channel and the Cr channel;

Mean value and deviation calculation sub-module, used to calculate the mean value and deviation of the H channel, S channel, Cb channel and Cr channel of the area to be processed;

The candidate white point acquisition sub-module is used to obtain the candidate white point according to the mean value and deviation of the H channel, S channel, Cb channel and Cr channel of the area to be processed. The calculation formula is specifically:

Among them, H(i,j), S(i,j), Cb(i,j) and Cr(i,j) are i pixel and j pixel of H channel, S channel, Cb channel and Cr channel respectively , AvrH _all , AvrS _all , AvrCb _all and AvrCr _all are the mean values of H channel, S channel, Cb channel and Cr channel respectively, DevH _all , DevS _all , DevCb _all and DevCr _all are H channel, S channel, Cb channel and The variance of the Cr channel;

The fusion sub-module is used to select a reference white point from the candidate white points, and fuse the reference white point with the luminance information of HSV space and YCbCr space. The specific calculation formula is:

Among them, img represents the fused image after the reference white point is fused with the brightness information of HSV space and YCbCr space, imgR, imgG and imgB represent the channel images corresponding to the R channel, G channel, and B channel respectively, and θ represents the multi-color domain fusion degree , V _max and Y _max are the maximum value of lightness and brightness respectively, and R _avgw , G _avgw , B _avgw are the pixel grayscale mean values of R, G, and B channels respectively.

Furthermore, the Youguang high-temperature industrial endoscope also includes a display module that sends the high-brightness and clear image video stream in the furnace to the host computer for display.

Compared with the prior art, the present invention has the advantages of:

The low-light high-temperature industrial endoscope provided by the present invention includes a shell assembly, a low-light optical imaging system installed in the shell assembly, and an image brightness processing module, wherein: the low-light optical imaging system is used to collect images and videos in the furnace stream; the image brightness processing module is connected with the Youguang level optical imaging system, and is used to receive the furnace image video stream output by the Youguang level optical imaging system, and perform brightness processing on the furnace image video stream, so as to obtain a bright and clear image Furnace image video stream, the invention solves the technical problems of insufficient brightness, fuzzy details, and unclear outlines of images in industrial furnaces under extremely weak light conditions, and the designed fidelity function, consistency function and regularization The function utilizes the strong correlation between the optimized image and the original image, achieves the purpose of eliminating ghosting in the enhanced image while ensuring the brightness of the enhanced image, and effectively improves the image quality of the enhanced image.

Key technical points of the present invention:

In terms of optical imaging system design, a new idea is adopted, using a reverse telephoto objective lens group with a large field of view and a large amount of light input, a Hopkins rod relay lens group with no optical loss that can offset transmission distortion and deviation, and manual focus The focusing lens group that makes the image clear can improve the brightness of the video picture from the hardware level, improve the quality of the video picture, and realize high-quality imaging in dim light environments.

In terms of the image processing algorithm processing platform, the algorithm is integrated into the FPGA chip to realize the hardware of the algorithm, get rid of the dependence on the host computer, make the equipment integrate the optical imaging system and image processing algorithm, and ensure the integrity and integrity of the equipment function. Independence, and the fidelity function, consistency function and regularization function are designed, and the strong correlation between the optimized image and the original image is used to achieve the purpose of eliminating ghosting in the enhanced image while ensuring the brightness of the enhanced image , effectively improving the image quality of the enhanced image.

Description of drawings

Fig. 1 is the overall structure diagram of the low-light high-temperature industrial endoscope according to Embodiment 1 of the present invention;

Fig. 2 is the low-light high-temperature industrial endoscope optical imaging system of Embodiment 1 of the present invention;

Fig. 3 is a block diagram of a low-light high-temperature industrial endoscope image brightness processing module according to Embodiment 1 of the present invention;

FIG. 4 is a schematic diagram of highlight video synthesis in Embodiment 1 of the present invention;

FIG. 5 is a flow chart of white balance processing in Embodiment 1 of the present invention;

Fig. 6 is a working flow chart of the low-light high-temperature industrial endoscope image brightness processing module according to Embodiment 1 of the present invention.

Reference signs:

U1, air-cooled circulation channel; U2, water-cooled circulation flow channel; U3, CCD imaging chip; U4, air-cooled air inlet; U5, image brightness processing module; U6, water-cooled water inlet; U7, water-cooled water outlet; M1, take Image objective lens group; M2, relay lens group; M3, Hopkins rod lens; M4, focusing lens group; M5, CCD photosensitive chip; S1, FPGA video stream image processing unit; S2, preset processing task module; S3, signal processing task trigger module; S4, MicroBlaze processing module; S5, DDR SDRAM storage module; S6, CCD; S7, input processing unit; S8, output processing unit; S9, host computer.

Detailed ways

In order to facilitate the understanding of the present invention, the present invention will be described more fully and in detail below in conjunction with the accompanying drawings and preferred embodiments, but the protection scope of the present invention is not limited to the following specific embodiments.

Embodiment one

As shown in Figure 1, the low-light high-temperature industrial endoscope of the embodiment of the present invention is composed of a housing assembly, a cooling system, a low-light level optical imaging system, and an image brightness processing module U5. The specific working process is as follows: First, rely on the dual cooling system that works simultaneously with air cooling and water cooling to ensure the normal operation of the equipment in the high temperature environment in the industrial furnace, and then adjust the focusing lens group M4 of the low-light optical imaging system, Ensure the clarity of image capture, and through the image capture objective lens group and relay lens group M2, the image information captured by the endoscope is transmitted to the CCD photosensitive chip M5 without loss, and then the image is pre-imaged by the image brightness processing module U5 Processing and brightness synthesis to obtain high-brightness and clear internal videos of industrial kilns. The following is a detailed description of each component of the device:

Housing components:

The shell of the embodiment of the present invention is mainly made of high temperature resistant nickel-based superalloy, as shown in FIG. 1 . The shell has a double-layer structure, wherein the outer layer is an air-cooled circulation channel U1, and the outer side of the inner layer is a double-helical water-cooled circulation flow channel U2. There is an air-cooled air inlet U4 and a water-cooled water inlet and outlet at the rear end of the housing. The rear cylindrical shell mainly includes CCD imaging chip U3, imaging drive circuit, image brightness processing module U5, video signal output interface and power supply, etc.

cooling system

The embodiment of the present invention adopts an air-cooled-water-cooled dual cooling system to ensure long-term stable operation of the equipment in a high-temperature environment in the furnace, as shown in FIG. 1 . The cold air enters the air-cooling channel in the inner layer of the shell from the air inlet, and the gas circulates in the air-cooling channel to form an air-cooling effect. The water flow enters the spiral pipe from the water-cooling water inlet U6, and finally is discharged from the water-cooling water outlet U7. The discharged water can flow in again after being processed, so as to form a water-cooling effect through circulation. In the embodiment of the present invention, the inlet and outlet of the dual cooling system are designed at the rear end of the device, which can not only cool the endoscope body in a high temperature environment, but also cool the imaging and image brightness processing module U5 at the rear end of the endoscope, and improve the endoscope. Stability of working in high temperature and harsh environment.

The low light level imaging system of the embodiment of the present invention is shown in Figure 2, including:

The imaging objective lens group M1, the relay lens group M2, the focusing lens group M4 and the CCD imaging chip U3. The details of each part are as follows:

(1) Image taking objective lens group M1

The reverse telephoto structure imaging has the characteristics of a smaller image plane when the field of view is larger. Under the same aperture, it can not only collect a wider image frame, but also increase the amount of light entering the lens, achieving a large field of view for the endoscope. Provide support for low-light level imaging with wide field of view and wide viewing angle. The imaging objective lens group M1 based on the reverse telephoto structure is mainly composed of a negative focal power lens group and a positive focal power lens group. The negative focal power lens group is responsible for scattering the incident light, and evenly scatters the collected optical information with large field of view and large amount of light to the positive focal power lens group, providing initial conditions for the long-distance parallel transmission of optical information. The positive focal power lens group re-gathers the scattered light to form a beam parallel to the optical axis. The gathered parallel beam prevents the divergence loss of part of the light and avoids the decrease in the brightness of the optical image during the collection process.

(2) Relay lens group M2

Since the front end of the endoscope is in a high temperature and highly corrosive environment in the furnace, the imaging chip is easy to burn out in this environment for a long time, and considering that the back end of the endoscope is outside the furnace, it can ensure the long-term and stable operation of the imaging system, so the imaging chip is installed behind the endoscope. Based on the above imaging chip installation scheme, it is necessary to build a set of relay lens group M2 to transmit the image acquired by the imaging objective lens group M1 to the rear end of the endoscope, increase the length of the optical imaging system, and at the same time realize the darkening of the collected image. loss transmission. Considering that the Hopkins rod lens M3 consists of exactly the same two rod lenses, and each rod lens group consists of a thick biconvex lens and two symmetrical concave lenses, the Hopkins rod lens M3 has complete symmetry structure. This structure can not only offset the distortion and deviation generated during the transmission process, but at the same time, because the light passes through the lens with a higher refractive index most of the time, the light loss is extremely small. Therefore, a relay lens group M2 composed of three Hopkins rod lenses M3 is designed to realize the optical loss-free transmission of the optical image from the front end to the back end of the endoscope.

(3) Focusing lens group M4

Since the conditions inside the industrial kiln are dynamically changing, in order to obtain a clear image inside the furnace, the focusing lens group M4 is designed. When the image of the CCD photosensitive chip M5 is blurred, the focus can be refocused by adjusting the focusing lens group M4 to ensure the clarity of the image.

(4) CCD imaging chip U3

The CCD imaging chip U3 converts the clear image transmitted through the focusing lens group M4 into a digital image, which is convenient for subsequent processing of the collected image.

As shown in FIG. 3 , the image brightness processing module U5 includes an input processing unit S7 , an FPGA video stream image processing unit S1 and an output processing unit S8 .

(1) Input processing unit S7

After conversion by CCD S6, the video stream composed of digital image frames needs to be decoded by a video decoding chip (such as TVP5150, etc.), and the decoded data information is stored in the DDR SDRAM storage module S5.

(2) FPGA video stream image processing unit S1

The FPGA video stream image processing unit S1 is the core of the image brightness processing module U5. Its main functions include processing task signal triggering, three-dimensional digital comb filtering, video highlight synthesis algorithm based on time-space domain adaptation, and multi-color dynamic domain algorithm-based white balance. The above functions and algorithm implementations are all embedded in the FPGA-based video stream image processing module. This module is a hardware platform embedded system with MicroBlaze as the core, developed with Verilog HDL hardware description language, and its process block diagram is shown in Figure 3. Its composition includes:

(1) MicroBlaze processing module S4

MicroBlaze is a highly flexible and configurable soft-core embedded processor. Its instruction functions cover read and write operations and calculation operations. The instruction operation is connected to DRR SDRAM through the DRR interface to read relevant data and perform calculations, or Write the operation result into DDR SDRAM and store the data.

(2) Signal processing task trigger module S3

The signal processing task trigger module S3 is mainly responsible for obtaining the image frame information stored in the DDR SDRAM in real time when the FPGA video stream image processing unit S1 is running, and sending task commands to MicroBlaze to realize task triggering, task configuration and task management functions. In terms of task triggering, the image frame information collected by the CCD and stored by the decoder is used to determine whether the number of image frames meets the algorithm requirements of the preset processing task module S2, and thus send relevant task commands to MicroBlaze; in terms of task configuration, pre-set task The storage space required for operation, the calculation mode of the task, etc., and the calculation stage of MicroBlaze are set according to the actual progress of the task; in terms of task management, before the task is loaded, configure relevant operation instructions for MicroBlaze, and obtain the image frame data required by the task through DDR SDRAM , and read the current algorithm task from the preset processing task module S2, and load related functions. During the task loading process, according to the hardware resources required for the current task execution, the storage space size is allocated and adjusted in real time to improve the task execution efficiency and ensure the stable operation of algorithms and functions.

(3) Preset processing task module S2

The preset task processing module S2 is a memory that pre-stores algorithms of tasks to be processed and commands for executing corresponding tasks. In the embodiment of the present invention, the algorithms and task commands to be processed mainly include three-dimensional digital comb filter, video highlight synthesis algorithm based on time-space domain adaptation and white balance based on multi-color dynamic domain algorithm. The detailed steps of each algorithm are as follows:

Step1: Three-dimensional digital comb filter:

Since the received video signal has not undergone any preprocessing, the luminance (L) and chrominance (C) in the original video signal are highly mixed, and the video image has problems such as bright color crossing, color overlapping, and serious clutter interference. Therefore, in order to preserve image details and solve the above problems at the same time, a digital comb filter will be used to filter the original video signal, and the delayer, adder, subtractor and bandpass filter of the digital comb filter will be designed to Finally, a video signal with no cross-color, no dot-like noise, no hanging point, and wider bandwidth of bright color signals can be obtained, which provides a high-quality image basis for subsequent image brightness processing. At the same time, because the three-dimensional comb filter has a very good bright color separation effect on static videos, and considering that the decline of the top material level of the blast furnace is a slow-changing process, the changes in two consecutive frames of images are small and relatively static, so the three-dimensional comb filter can be used The shape filter performs calculation processing on two consecutive frames of images of the furnace top material surface video to obtain a clearer video image.

Step2: video highlight synthesis algorithm based on spatio-temporal domain adaptation

The image processed by the comb filter needs to be combined with a high-brightness video. The schematic diagram of the implementation of the high-brightness video combination in this embodiment is shown in FIG. 4 . Specific steps are as follows:

(1) The difference between light and dark of the original video image frame I _t is relatively large. In order to avoid the overexposure phenomenon during superimposition processing, the video image frame is nonlinearly transformed to obtain I′ _t , and the gray value of the darker pixel in the original image is Increase to suppress the gray value of the brighter part. The following transformations can be used:

P _y and P _x represent the gray value after transformation and the gray value before transformation respectively, n can be used to control the overall improvement of brightness, and is selected according to the gray distribution of the image, and n=2 is taken here.

(2) Overlay the transformed image. Select I′ _t-1 , I′ _t and I′ _t+1 for superposition to obtain an enhanced image Y _t . Since three frames are selected for image brightness enhancement, ghosting may appear in the enhanced image Y _t , so the enhanced image needs to be optimized.

(3) The maximum a posteriori model is used to optimize the superimposed image Y _t . The posterior probability is jointly determined by the fidelity function ψ(·,·), the consistency function φ(·,·) and the regularization function L(·).

(4) The fidelity function ψ(·,·) measures the similarity between the optimized image and the original image, and the Mahalanobis distance is used for calculation. The calculation formula is,

为O_t和Y_j的联合分布的方差，B和D分别表示模糊矩阵和下采样矩阵，S表示协方差矩阵，

表示t时刻的优化图像O_t到j时刻的增强图像Y_j的运动补偿矩阵。Among them, ψ(O _t , Y _j ) is a fidelity function, which is used to measure the similarity between the optimized image O _t at time t and the enhanced image Y _j at time j,

is the variance of the joint distribution of O _t and Y _j , B and D represent the fuzzy matrix and the downsampling matrix respectively, S represents the covariance matrix,

Represents the motion compensation matrix of the optimized image O _t at time t to the enhanced image Y _j at time j.

(5) The consistency function φ(·,·) is used to control the time consistency of the synthesized frame, and its calculation formula is,

表示t-1时刻的输出优化图像，

表示一致性函数，衡量t时刻的优化图像O_t和t-1时刻的输出优化图像

之间的相似性，

是O_t和

的联合分布的方差，

是t时刻优化图像O_t到t-1时刻的输出优化图像

的运动补偿矩阵。in,

Indicates the output optimized image at time t-1,

Represents the consistency function, which measures the optimized image O _t at time t and the output optimized image at time t-1

the similarity between

is O _t and

The variance of the joint distribution of ,

is the optimized image O at time _t and the output optimized image at time t-1

motion compensation matrix.

(6) The regularization function L( ) is used to suppress image noise information, and the calculation formula is,

Among them, L(O _t ) is the regularization function, P is the size of the moving window, α is used to constrain the smoothness of the optimized image, H _l and V _j indicate that the optimized image O _t at time t is moved in the horizontal and vertical directions respectively Operator for l and j pixels.

(7) The posterior probability of the synthesized video frame is:

Among them, p(O _t ) is the posterior probability of the optimized image O _t , r is the video frame number of the original video used to output the optimized image before time t, and b is the frame number of the original video used to output the optimized image after time t The number of video image frames. The fidelity function, consistency function and regularization function designed in the embodiment of the present invention utilize the strong correlation between the optimized image and the original image, and achieve the purpose of eliminating ghosting in the enhanced image while ensuring the brightness of the enhanced image , effectively improving the image quality of the enhanced image.

(8) The output synthetic image frame is:

Combined with the above functions, the output synthetic image frame can be expressed as:

对原始视频流进行处理后即可以得到高亮视频流。(9) According to the formula (7), the composite brightness enhanced image frame can be obtained by using the least square method

After processing the original video stream, the highlighted video stream can be obtained.

Step3: White balance based on multi-color dynamic domain algorithm

There is chromatic aberration in the video image after highlight processing. In order to improve the image quality, a white balance process is performed on the image based on a multi-color dynamic domain algorithm to obtain an image closer to the real color temperature in the kiln. This embodiment realizes the process of white balance processing The figure is shown in Figure 5, and the specific steps of the white balance algorithm are as follows:

(1) Convert the image from RGB space to HSV space and YCbCr space respectively, and record the results as H, S, V, Y, Cb, Cr.

(2) Divide the image into 8 areas to enhance the robustness of the algorithm, and calculate the mean absolute deviation DevH, DevS, DevCb, and DevCr of H, S, Cb, and Cr in each area. The calculation formula is:

Where N is the number of pixels in each area.

(3) Find the |DevH+DevS| and |DevCb+DevCr| of each area. When one of the indicators in an area is too small, it indicates that the color distribution of this area is uniform, which is not conducive to white balance processing. Choose to ignore this area.

(4) Regardless of the areas that are not conducive to white balance processing in the previous step, calculate the mean value AvrH _all , AvrS _all , AvrCb _all , AvrCr _all and variance DevH _all , DevS _all , DevCb _all , DevCr of the entire image except these areas _all .

(5) Determine the candidate white point, the candidate white point meets the following requirements,

(6) Select the reference white point. The selection method of the reference white point is: arrange the luminance values of the candidate white point pixels from high to low, and select the white point whose luminance value is in the top 10% in HSV space and YCbCr space as the reference white point. point.

(7) The reference white point is fused with the luminance information of HSV space and YCbCr space to complete the white balance calculation. The calculation formula is as follows,

Wherein, V _max is the maximum value of the brightness of all images in the image, Y _max is the maximum value of the brightness of all images in the image, θ is the multi-color domain fusion degree, and θ=0.4 can be selected.

After the white balance processing of the video, it is stored in the memory DDR SDRAM, and at the same time, it is decoded and sent to display through the output processing unit S8.

Through the above method, it is possible to obtain images in the furnace through the endoscope in an extremely weak light environment. The image quality of the video stream is improved by three-bit comb filtering. Then, the video stream is highlighted by multi-frame overlay to optimize the video stream. Then white balance processing is performed on the high-brightness video stream to get close to the real color temperature in the furnace to obtain high-quality images in the furnace. Finally, the video stream obtained in the furnace is stored and sent to the upper computer S9 for display.

(4) DDR SDRAM storage module S5

It is mainly responsible for storing the collected and decoded video stream image frame information, the image frame information processed by the preset task algorithm, and the relevant information in the task operation processing process, and provides a data exchange platform for other modules.

(5) Output processing unit S8

The highlighted video stream re-synthesized by the image processing task is output to the upper computer S9 through video coding (VENC) and on-screen display (OSD).

In this embodiment, based on the above steps, image frame processing and high-brightness video synthesis analysis is performed on the captured video stream, and the specific workflow of the image brightness processing module U5 is as shown in Figure 6: first, the video stream captured by the CCD is decoded , and stored in DDR SDRAM; secondly, the signal processing trigger module judges the number of image frames, and sends the image preset task processing trigger signal to MicroBlaze, and at the same time configures the task in advance and manages the addition and deletion of data resources; then, when MicroBlaze receives the trigger signal, takes out image frame data from DDR SDRAM, performs three-dimensional digital comb filtering, video highlight synthesis based on time-space domain adaptation and white balance based on multi-color gamut algorithm, and finally generates a highlighted video stream and The output is sent to display.

The low-light high-temperature industrial endoscope image capture provided by the embodiment of the present invention obtains digital images in industrial furnaces and kilns to form an original video stream, performs three-dimensional comb filtering on the original video stream, and performs three-dimensional comb filtering on the original video stream, and performs three-dimensional comb filtering on the video after three-dimensional comb filtering. The high-brightness video stream is synthesized to obtain the high-brightness video stream and perform white balance processing on the image frames in the high-brightness video stream, so as to obtain high-quality furnace images, which solves the problem of industrial furnace images under extremely low light conditions. The technical problems of insufficient imaging brightness, fuzzy details, and unclear outlines, and the designed fidelity function, consistency function, and regularization function use the strong correlation between the optimized image and the original image to achieve the goal of ensuring the brightness of the enhanced image. Simultaneously, the purpose of eliminating the double image appearing in the enhanced image effectively improves the image quality of the enhanced image.

Specifically, in this embodiment, firstly, the image quality of the video stream is improved through three-dimensional comb filtering, and then a method for highlighting video stream synthesis based on multi-frame synthesis and maximum a posteriori probability model is designed to perform highlight processing on the video stream, and finally Perform white balance processing on the high-brightness video stream to get close to the real color temperature in the furnace, obtain high-quality images in the furnace, and provide information collection for image and video streams in industrial kilns in extremely low-light environments (illuminance below 0.0001Lux). A feasible and reliable imaging device is proposed.

Embodiment two

After the equipment in Embodiment 2 of the present invention is installed, start to collect video information, the specific steps are as follows:

(1) In order to ensure that the Youguang high-temperature endoscope can work stably in the high-temperature environment in the furnace, the air-cooled-water-cooled dual cooling system is first turned on, and the cold air and cold air are sent into the equipment through the air inlet and water inlet respectively to continuously cool down the equipment .

(2) After power-on, the device starts to collect image information inside the industrial furnace stably.

(3) Adjust the focusing lens group M4 to focus on the interior of the industrial furnace to make the image of the endoscope clear.

(4) The original image inside the industrial kiln is collected by a reverse telephoto objective lens group with a large field of view and a large amount of incoming light, and is transmitted to the CCD imaging chip without optical loss through the Hopkins rod lens M3 that can offset transmission distortion and deviation U3, realizing low-light level imaging of industrial endoscopes.

(5) The CCD imaging chip U3 converts the image signal into an electrical signal, which is used as the input of the image brightness processing module U5.

(6) Perform three-dimensional digital comb filter processing on the input image frame.

(7) Using the adaptive video highlight synthesis algorithm in the space-time domain, video stream synthesis is performed on multiple image frames to effectively improve video brightness.

(8) Based on the white balance of the multi-color dynamic domain algorithm, the white balance processing is performed on the image frame to realize the high-quality low-light-level highlight video stream output of the industrial endoscope.

(9) Through the output processing unit S8, the video stream after the video encoding and screen display module is transmitted to the host computer S9, so as to realize the output and display of the high-brightness video stream of the low-light industrial endoscope.

The above are only preferred embodiments of the present invention, and are not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and changes. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principles of the present invention shall be included within the protection scope of the present invention.

Claims (10)

1. A pyloric high temperature industrial endoscope, comprising a housing assembly, a pyloric-level optical imaging system mounted within the housing assembly, and an image brightness processing module (U5), wherein:

the image capturing system comprises an imaging system, a light level optical imaging system and a light level optical imaging system, wherein the imaging system is used for acquiring an image video stream in a furnace;

and the image brightness processing module (U5) is connected with the faint light level optical imaging system and is used for receiving the in-furnace image video stream output by the faint light level optical imaging system and carrying out brightness processing on the in-furnace image video stream so as to obtain a high-brightness and clear in-furnace image video stream.

2. The pyloric high temperature industrial endoscope of claim 1, wherein the housing assembly includes a front end housing assembly and a back end housing assembly, wherein:

the front-end housing assembly is of a double-layer structure, an air-cooling circulation channel (U1) is arranged on the outer layer of the double-layer structure, and a double-spiral wound water-cooling circulation flow channel (U2) is arranged on the outer side of the inner layer;

the rear end shell component is used for installing a CCD imaging chip (U3), an imaging driving circuit and an image brightness processing module (U5), and an air cooling air inlet (U4), a water cooling water inlet (U6) and a water cooling water outlet (U7) are formed in the rear end shell component.

3. The pyloric high temperature industrial endoscope of claim 2, further comprising a cooling system, wherein:

the cooling system adopts an air cooling-water cooling double cooling system, cold air in the air cooling-water cooling double cooling system enters an air cooling circulation channel (U1) from an air cooling air inlet, and air circulates in the air cooling circulation channel (U1) to form an air cooling effect;

water flow in the air-cooling and water-cooling double-cooling system enters the water-cooling circulation flow channel (U2) from the water-cooling water inlet (U6) and is discharged from the water-cooling water outlet (U7), and the discharged water flow can flow in again after being treated, so that the water-cooling effect is formed by circulation.

4. The pyloric high-temperature industrial endoscope according to claim 1, wherein the pyloric optical imaging system comprises an image capturing objective lens group (M1), a relay lens group (M2), a focusing lens group (M4), and a CCD imaging chip (U3) connected in sequence, wherein:

the image capturing objective lens group (M1) comprises a negative power lens group and a positive power lens group, the negative power lens group is used for scattering incident light, the acquired optical information with a large field angle and a large light incoming amount is uniformly scattered to the positive power lens group, initial conditions are provided for long-distance parallel transmission of the optical information, and the positive power lens group is used for re-gathering the scattered light to form light beams parallel to an optical axis;

the relay lens group (M2) is used for transmitting the in-furnace image acquired by the image taking objective lens group (M1) to a CCD imaging chip (U3);

the focusing mirror group (M4) is used for adjusting the focal length when the CCD imaging chip (U3) is blurred, so that the imaging is clear;

and the CCD imaging chip (U3) is used for converting the clear image transmitted by the focusing mirror group (M4) into a digital image.

5. A claustrophobic high temperature industrial endoscope according to any one of claims 1-4, wherein the image brightness processing module (U5) comprises an input processing unit (S7) and an FPGA video stream image processing unit (S1) and an output processing unit (S8) connected in sequence with the input processing unit (S7), wherein:

the input processing unit (S7) is used for decoding the in-furnace image video stream output by the optical imaging system at the dim level and storing the decoded in-furnace image video stream into a DDR SDRAM storage module (S5);

the FPGA video stream image processing unit (S1) is used for carrying out image preprocessing and brightness synthesis on the decoded in-furnace image video stream, and specifically comprises a MicroBlaze processing module (S4), a signal processing task triggering module (S3), a preset processing task module (S2) and a DDR SDRAM storage module (S5), wherein the signal processing task triggering module (S3) is connected with the MicroBlaze processing module (S4), and the FPGA video stream image processing unit (S1) is used for carrying out image preprocessing and brightness synthesis on the decoded in-furnace image video stream, and comprises:

the signal processing task triggering module (S3) is used for judging whether the number of image frames in the furnace image video stream meets the algorithm requirement in the preset processing task module (S2) or not according to the furnace image video stream stored by the DDR SDRAM storage module (S5), and sending a task command to the MicroBlaze processing module (S4) to realize task triggering, task configuration and task management;

the preset processing task module (S2) is used for pre-storing an algorithm of an image preprocessing and brightness synthesis processing task;

the MicroBlaze processing module (S4) is used for receiving task triggering, task configuration and task management commands sent by the signal processing task triggering module (S3), and carrying out image preprocessing and brightness synthesis on the decoded in-furnace image video stream according to an algorithm of an image preprocessing and brightness synthesis processing task stored in the preset processing task module (S2) in advance;

the DDR SDRAM storage module (S5) is used for storing the in-furnace image video stream decoded by the input processing unit (S7) and performing data interaction with the signal processing task triggering module (S3) and the MicroBlaze processing module (S4);

and the output processing unit (S8) is used for outputting the in-furnace image video stream processed by the FPGA video stream image processing unit (S1).

6. The hyperspectral high-temperature industrial endoscope of claim 5, wherein the preset processing task module (S2) comprises a three-dimensional digital comb filtering algorithm sub-module, a video highlight synthesis algorithm sub-module based on time-space domain adaptation, and a white balance algorithm sub-module based on a multi-color dynamic domain algorithm, which are connected in sequence, wherein:

the three-dimensional digital comb filtering algorithm submodule is used for carrying out three-dimensional digital comb filtering on the decoded in-furnace video stream;

the video highlight synthesis algorithm submodule based on the time-space domain self-adaption is used for carrying out video highlight synthesis on the video stream in the furnace after the three-dimensional digital comb filtering;

and the white balance algorithm submodule based on the multi-color dynamic domain algorithm is used for carrying out white balance processing on the in-furnace video stream after the video brightness is synthesized.

7. The pyloric high-temperature industrial endoscope of claim 6, wherein the video highlight synthesis algorithm sub-module based on time-space domain adaptation comprises a video frame acquisition sub-module, a non-linear transformation sub-module, an overlay sub-module and a synthesized video frame acquisition sub-module which are connected in sequence, wherein:

the video frame acquisition submodule is used for acquiring video frames in the three-dimensional comb-shaped filtered video stream, wherein the video frames comprise a current video frame and a video frame adjacent to the current video frame;

the nonlinear transformation submodule is used for carrying out nonlinear transformation on the video frame;

the superposition submodule is used for superposing the video frames subjected to the nonlinear transformation to obtain an enhanced image;

and the composite video frame acquisition submodule is used for optimizing the enhanced image by adopting a maximum posterior model to obtain a composite video frame so as to obtain a highlight video stream.

8. The pyloric high-temperature industrial endoscope of claim 7, wherein the composite video frame acquisition sub-module comprises a fidelity function calculation sub-module, a consistency function calculation sub-module, a regularization function calculation sub-module, a posterior probability calculation sub-module, and a composite video frame calculation sub-module, which are connected in sequence, wherein:

the fidelity function calculating submodule is used for calculating the fidelity function, wherein the fidelity function has a calculation formula as follows:

wherein,ψ(O _t ,Y _j ) Is a fidelity function and is used for measuring the optimized image O at the time t _t And j time instant enhanced image Y _j The similarity of (a) to (b) is,

is O _t And Y _j B and D represent the blur matrix and the down-sampled matrix, respectively, S represents the covariance matrix,

optimized image O representing time t _t Enhanced image Y to time j _j The motion compensation matrix of (a);

the consistency function calculation submodule is used for calculating a consistency function, wherein the calculation formula of the consistency function is as follows:

wherein,

representing the output optimized image at time t-1,

an optimized image O representing a consistency function and measuring the t moment _t And output optimized image at time t-1

The similarity between the two or more of the images,

is O _t And

the variance of the joint distribution of (a),

is the optimized image O at time t _t Output optimized image to time t-1

The motion compensation matrix of (a);

the regularization function calculation sub-module is used for calculating a regularization function, wherein a calculation formula of the regularization function is as follows:

wherein, L (O) _t ) For the regularization function, P is the size of the moving window, α is used to constrain the smoothness of the optimized image, H _l And V _j Indicating that the time t is optimized to the image O _t Operators that move the l and j pixels in the horizontal and vertical directions, respectively;

the posterior probability calculation submodule is used for obtaining an optimized image O according to the fidelity function, the consistency function and the regularization function _t And the calculation formula of the posterior probability is as follows:

wherein, p (O) _t ) Is to optimize the image O _t R is the video image frame number of the original video used by the output optimized image before the time t, b is the video image frame number of the original video used by the output optimized image after the time t;

the composite video frame calculation submodule is used for obtaining a composite video frame according to the posterior probability of the output optimized image, and the specific calculation formula is as follows:

9. the deep light industrial endoscope of claim 8, wherein the white balance algorithm sub-module comprises a space conversion sub-module, an absolute deviation mean calculation sub-module, a region to be processed acquisition sub-module, a mean and deviation calculation sub-module, a candidate white point acquisition sub-module and a fusion sub-module, which are connected in sequence, wherein:

the space conversion sub-module is used for converting image frames in the highlight video stream from an RGB space to an HSV space and a YCbCr space respectively to obtain an H channel, an S channel, a V channel, a Y channel, a Cb channel and a Cr channel;

the absolute deviation mean value calculation submodule is used for dividing the image frame into a preset number of areas and calculating the absolute deviation mean values of the H channel, the S channel, the Cb channel and the Cr channel of each area respectively;

the to-be-processed area acquisition submodule is used for acquiring the to-be-processed area needing white balance processing according to the absolute deviation mean value of the H channel, the S channel, the Cb channel and the Cr channel;

the mean value and deviation calculation submodule is used for calculating the mean values and deviations of an H channel, an S channel, a Cb channel and a Cr channel of the region to be processed;

the candidate white point obtaining submodule is used for obtaining candidate white points according to the mean values and the deviations of the H channel, the S channel, the Cb channel and the Cr channel of the area to be processed, and the calculation formula specifically comprises the following steps:

wherein, H (i, j), S (i, j), cb (i, j) and Cr (i, j) are respectively the i pixel and the j pixel of the H channel, the S channel, the Cb channel and the Cr channel, and the AvrH _all 、AvrS _all 、AvrCb _all And AvrCr _all Mean values of H channel, S channel, cb channel and Cr channel, devH _all 、DevS _all 、DevCb _all And DevcCr _all Variances of an H channel, an S channel, a Cb channel and a Cr channel respectively;

the fusion submodule is used for selecting a reference white point from the candidate white points and fusing the reference white point with the brightness information of the HSV space and the YCbCr space, and the calculation formula specifically comprises the following steps:

wherein img represents a fused image obtained by fusing brightness information of a reference white point and HSV space and YCbCr space, imgR, imgG and imgB represent channel images corresponding to an R channel, a G channel and a B channel respectively, theta represents multi-color-domain fusion degree, and V is _max And Y _max Respectively, the maximum value of lightness and the maximum value of brightness, R _avgw 、G _avgw 、B _avgw The average pixel gray scale values of the R channel, the G channel and the B channel are respectively.

10. The pyloric high-temperature industrial endoscope according to claim 9, further comprising a display module for transmitting the highlight clear in-furnace image video stream to an upper computer (S9) for display.

Images