CN-121982357-A - Method and system for identifying polar surface defects of electric dust remover

CN121982357ACN 121982357 ACN121982357 ACN 121982357ACN-121982357-A

Abstract

The embodiment of the invention provides a method and a system for identifying polar surface defects of an electric dust collector, belonging to the technical field of electric dust collection. The method comprises the steps of collecting an original digital color image of a polar line in a corona discharge state, generating a gray image based on the original digital color image, constructing a binary image used for representing a brightness area according to a preset threshold value, constructing a denoising binary image based on the binary image by sequentially performing median filtering and mean filtering, performing Canny edge detection on the denoising binary image to generate a halo edge image of a barbed sharp point, extracting corona discharge positions and brightness information of two sharp points of each barbed point based on the halo edge image, and outputting a polar line surface defect identification result according to a preset normal distribution mode. According to the scheme, the corona discharge characteristics can be stably extracted in complex imaging environments such as high dust, weak light and the like, so that the reliable identification of polar line surface defects is realized.

Inventors

ZHONG JIANFENG
YAO SHUYONG
YUAN BIN
ZHOU YUCHENG
ZHAO CHEN
ZHAO HAIBAO
WANG ZEDONG

Assignees

浙江菲达环保科技股份有限公司

Dates

Publication Date: 20260505
Application Date: 20251205

Claims (10)

1. The method for identifying the polar line surface defects of the electric dust remover is characterized by comprising the following steps: collecting an original digital color image of an electrode wire in a corona discharge state; generating a gray image based on the original digital color image, and constructing a binary image for representing a brightness area according to a preset threshold value; Sequentially performing median filtering and mean filtering on the binary image to construct a denoising binary image, and performing Canny edge detection on the denoising binary image to generate a halo edge image of the barbed point; and extracting corona discharge position and brightness information of two sharp points of each bur based on the halation edge image, and outputting an polar line surface defect identification result according to a preset normal distribution mode.
2. The method of claim 1, wherein capturing the raw digital color image of the epipolar line in the corona discharge state comprises: determining a shooting visual angle and an imaging range of a camera based on the relative posture of an image acquisition module fixed on the side edge of the polar line system of the electric dust collector; When the polar line is in a corona discharge state, performing continuous imaging on the polar line area with preset exposure time, preset sensitivity and preset frame rate to obtain multi-frame digital color images covering the whole polar line; And carrying out inter-frame consistency screening on the multi-frame digital color image based on the brightness histogram difference value of corona halo pixels in each frame image, and obtaining the frame image meeting the preset consistency condition as an original digital color image.
3. The method of claim 1, wherein generating a grayscale image based on the original digital color image comprises: extracting red, green and blue components of each pixel in the original digital color image; Performing weighting operation on the red component, the green component and the blue component based on a preset brightness conversion coefficient, and constructing a gray matrix for representing brightness values of each pixel; And writing all brightness values in the gray matrix into an image buffer area according to the matrix position sequence, and generating a gray image corresponding to the original digital color image.
4. A method according to claim 3, wherein constructing a binary image for characterizing the luminance region in accordance with a preset threshold comprises: the brightness value of each pixel in the gray image is read, and each brightness value is compared with a preset brightness threshold value; Writing the position of the corresponding pixel into a highlight value unit in the binary matrix when the brightness value is larger than a preset brightness threshold value, and writing the position of the corresponding pixel into a low-brightness value unit in the binary matrix when the brightness value is not larger than the preset brightness threshold value so as to obtain a binary image for representing the brightness region.
5. The method of claim 1, wherein constructing a denoised binary image based on the binary image by sequentially performing median filtering and mean filtering, comprises: extracting a first neighborhood window of a first preset size taking the target pixel as a center from each target pixel in the binary image, and calculating the median value of all pixel values in each first neighborhood window; Replacing the target pixel value of the corresponding first neighborhood window with the median value in each first neighborhood window to generate a median filtering image; Extracting a second neighborhood window of a second preset size taking the target pixel as a center from each target pixel in the median filtering image, and calculating the average value of all pixel values in each second neighborhood window; And replacing the target pixel value of the corresponding second neighborhood window with the average value in each second neighborhood window to generate a denoising binary image.
6. The method of claim 5, wherein performing Canny edge detection on the denoised binary image to generate a halo edge image of a barbed point comprises: Respectively calculating a row-direction gradient component obtained by a first-order difference along a row direction and a column-direction gradient component obtained by a first-order difference along a column direction in the denoising binary image, and respectively constructing a gradient amplitude matrix and a gradient direction matrix based on the row-direction gradient component and the column-direction gradient component; performing non-maximum suppression processing on the gradient magnitude matrix according to the gradient direction matrix; comparing the gradient amplitude subjected to non-maximum value inhibition treatment with a preset amplitude high threshold value and a preset amplitude low threshold value by double threshold values, and respectively marking strong edge points and weak edge points; connectivity determination is performed on the adjacent weak edge points based on the strong edge points, and a halation edge image for characterizing the corona discharge position of the barbed point is generated according to the determination result.
7. The method of claim 6, wherein performing connectivity decisions on adjacent weak edge points based on strong edge points and generating a halo edge image characterizing a barbed spike corona discharge location from the decisions comprises: Searching weak edge points consistent with the gradient direction of the strong edge points in the continuous neighborhood of the strong edge points and the gradient direction of the strong edge points, and marking the weak edge points with the gradient direction deviation of the gradient direction of the strong edge points not exceeding a preset angle threshold value as connection candidate points; Constructing candidate edge chains extending along the gradient direction based on the connection candidate points, and screening target edge chains meeting preset matching conditions according to matching results of gradient amplitude sequences of pixels in the candidate edge chains and preset brightness change modes; Writing all pixels in the target edge chain into an edge pixel set, and clearing weak edge points which do not belong to the target edge chain; Generating a halation edge image for representing the corona discharge positions of the two sharp points of the bur according to the pixel sequence based on the edge pixel set.
8. The method of claim 1, wherein extracting corona discharge position and brightness information of two points of each bur based on the halo edge image and outputting a polar line surface defect recognition result according to a preset normal distribution mode comprises: Separating each halation region in the halation edge image based on the spatial communication relation of edge pixels, and calculating the geometric centroid position of each halation region and the brightness average value of the edge pixels to respectively serve as corona discharge position and brightness information corresponding to the barbed points; calculating the pitch of the sharp points based on two corona discharge positions of the same bur, and comparing the pitch of the sharp points with the pitch of the sharp points in a preset normal distribution mode; Calculating a brightness difference value based on the brightness average value of two corona discharge of the same bur, and comparing the brightness difference value with a preset normal brightness difference value range; marking the corresponding barbed as a sharp defect when the sharp pitch is greater than the preset sharp pitch threshold or the brightness difference is greater than the preset brightness difference threshold; Marking the corresponding polar line area as a polar line integral defect under the condition that any corona discharge position is not obtained by a plurality of continuous barbs; and outputting a pole line surface defect identification result based on the marking result of the sharp point defect and/or the pole line overall defect.
9. An electrostatic precipitator polar line surface defect identification system, characterized in that the system comprises: the acquisition unit is used for acquiring an original digital color image of the polar line in a corona discharge state; The construction unit is used for generating a gray image based on the original digital color image and constructing a binary image for representing a brightness area according to a preset threshold value; The processing unit is used for sequentially executing median filtering and mean filtering on the basis of the binary images to construct a denoising binary image, and executing Canny edge detection on the denoising binary image to generate a halation edge image of the barbed point; And the result output unit is used for extracting corona discharge positions and brightness information of two sharp points of each bur based on the halation edge image and outputting an identification result of the polar line surface defects according to a preset normal distribution mode.
10. A computer readable storage medium having instructions stored thereon, which when run on a computer causes the computer to perform the method for identifying surface defects of an electric precipitator pole line according to any of claims 1-8.

Description

Method and system for identifying polar surface defects of electric dust remover Technical Field The invention relates to the technical field of electric dust removal, in particular to an electric dust remover polar line surface defect identification method and an electric dust remover polar line surface defect identification system. Background As an industrial flue gas treatment device, the electric dust collector is used for trapping particles in flue gas by means of a high-voltage electric field, and the running state of the electric dust collector has direct influence on the emission control of atmospheric pollutants. Under long-term operation conditions, due to the factors of equipment age increase, operation condition fluctuation, high temperature and high voltage bearing of internal structures and the like, faults such as short circuit, electrode deformation, electrode wire connecting piece breakage, partial discharge caused by foreign matter lap joint and the like easily occur in an electric field in the electric precipitator. Upon such failure, the faulty electric field may trigger an automatic shutdown, affecting the overall dust removal performance and in certain cases creating a risk of emissions overrun. At present, the overhaul of the electrode wires inside the electric dust remover mainly depends on the manual entry of the electric field cavity for inspection. However, the scaffold or the lifting platform cannot be arranged in the electric dust collector, personnel must carry out suspension type operation through the safety belt and the anti-falling device, the operation space is narrow, the posture is limited, overhaul blind areas exist, and detection quality is difficult to guarantee. Therefore, in order to advance the automation of maintenance operation, a technical scheme capable of automatically identifying the surface defects of the electrode wires of the electric dust remover is needed. However, the internal detection environment of the electric dust remover is poor, the dust content of the flue gas is large, the illumination is insufficient, and the image acquired by the camera is often low in resolution and high in noise. Meanwhile, the polar line has a complex structure and numerous barbs, the traditional image analysis method based on geometric outline or shape features is difficult to accurately separate the appearance features of the cathode line and the barbs, is easy to be misjudged due to noise and shielding influence, and is difficult to realize reliable polar line defect identification. Disclosure of Invention The embodiment of the invention aims to provide a method and a system for identifying polar line surface defects of an electric dust remover, which at least solve the problem that polar line and barbed surface defects are difficult to accurately identify in complex environments such as high dust, weak light and the like in the prior art. In order to achieve the above purpose, the first aspect of the invention provides a method for identifying polar line surface defects of an electric precipitator, which comprises the steps of collecting polar line original digital color images in a corona discharge state, generating gray images based on the original digital color images, constructing binary images used for representing brightness areas according to preset thresholds, sequentially performing median filtering and mean filtering on the binary images to construct denoising binary images, performing Canny edge detection on the denoising binary images to generate halo edge images of barbed points, extracting corona discharge positions and brightness information of the two points of each barbed point based on the halo edge images, and outputting polar line surface defect identification results according to a preset normal distribution mode. The method comprises the steps of acquiring original digital color images of polar lines in a corona discharge state, determining a shooting visual angle and an imaging range of a camera based on relative postures of an image acquisition module fixed on the side edge of an electric dust collector polar line system, continuously imaging polar line areas with preset exposure time, preset light sensitivity and preset frame rate when the polar lines are in the corona discharge state to obtain multi-frame digital color images covering the whole polar lines, and carrying out inter-frame consistency screening on the multi-frame digital color images based on brightness histogram difference values of corona halo pixels in each frame image to obtain frame images meeting preset consistency conditions as original digital color images. Optionally, generating a gray scale image based on the original digital color image includes extracting a red component, a green component and a blue component of each pixel in the original digital color image, performing a weighting operation on the red component, the green component and the blue component based on a pr