CN-121998910-A - Quick earth wire detection method based on image edge gradient distribution

CN121998910ACN 121998910 ACN121998910 ACN 121998910ACN-121998910-A

Abstract

The specification discloses a rapid earth wire detection method based on image edge gradient distribution, and relates to the technical field of photoelectric guidance. The method utilizes an improved median filtering program to filter image noise and eliminate the influence of the image noise on edge extraction, utilizes an improved edge extraction algorithm to realize stable extraction of image edges and textures, utilizes texture distribution characteristics of an aerial unmanned plane, a cloud layer and the like to discriminate whether the texture distribution characteristics are space-earth demarcation points or not to realize robust extraction of space-earth wires, and combines image edge gradient information with edge characteristics of ground object background to realize rapid detection and identification of space-earth demarcation lines and promote real-time and accuracy of identification. The problem of prior art when surveying, because the existence of ground object target, the ground object target is detected, the false alarm rate is higher is solved.

Inventors

ZHANG TING
CUI JINGJING
ZHANG DONGFANG
Zong Jiaqing

Assignees

北京遥感设备研究所

Dates

Publication Date: 20260508
Application Date: 20251225

Claims (10)

1. The rapid antenna and ground wire detection method based on image edge gradient distribution is characterized by comprising the following steps of: performing improved median filtering on the image to be identified, and removing Gaussian white noise introduced by detector imaging; Using an improved sobel edge extraction algorithm to adaptively select an edge threshold value, and extracting edges and textures of an image; According to the image edge and texture structure, extracting the world demarcation points by utilizing edge gradient distribution and combining the feature background edge distribution characteristics; All the space-ground demarcation points are connected to form a space-ground demarcation line.
2. The method of claim 1, wherein the improved median filtering of the image to be identified to remove gaussian white noise introduced by detector imaging comprises: judging 3*3 cross filtering areas, judging whether the central pixel is higher than or lower than 20 gray scales of eight neighborhood pixels, if so, considering Gaussian noise, and using a median filtering method to replace the central pixel value with the ordered intermediate value pixel gray scale; if the center pixel is not satisfied that all are above or all are below 20 gray levels of the eight neighborhood pixels, then the center pixel is considered not gaussian noise, which is preserved.
3. The method according to claim 2, wherein the processing mode of the modified sobel edge extraction algorithm includes: gradient differentiation is performed on the image, and when the difference is greater than a given edge threshold, the point is considered to be an edge point.
4. A method according to claim 3, wherein the selection of the edge threshold is adaptively selected in a manner comprising: And calculating the gray average value of the image and the difference value between each pixel and the gray average value, taking the absolute value of the difference value, summing all the absolute values, and finally calculating the average value of the absolute difference value, wherein half of the average value is used as the edge threshold value of the frame image.
5. The method of claim 4, wherein extracting edges and textures of an image comprises: Starting from the upper left corner of the image to be identified after median filtering, convolving the image with the image by utilizing an edge structure operator, calculating the gradient value of the pixel point, setting the gray value to 255 if the gradient value is greater than an edge threshold value, otherwise setting the gray value to 0; traversing the complete frame image to obtain the edge and texture of the frame image.
6. The method of claim 5, wherein the edge structure operator is: Lx=∑I(i-1,j+1)+2*I(i,j+1)+I(i+1,j+1)-∑I(i-1,j-1)-2*I(i,j-1)-I(i+1,j-1) Ly=∑I(i-1,j-1)+2*I(i-1,j)+I(i-1,j+1)-∑I(i+1,j-1)-2*I(i+1,j)-I(i+1,j+1) Where Lx is the azimuth edge calculation formula, ly is the elevation edge calculation formula, and I (I, j) is the pixel gray value at the image (I, j) coordinates.
7. The method of claim 6, wherein the extracting the world demarcation point based on the image edge, the texture, and the edge gradient distribution in combination with the feature background edge distribution characteristics comprises: judging whether the gray value of each pixel is larger than 0 from the first pixel of each column of the image edge, if so, continuously judging 15 pixels from the second pixel of the column, and if 10 pixels are larger than 0, judging the second pixel of the column as the space-earth demarcation point of the column; If no 10 pixels are larger than 0, starting at the second pixel in the row, and repeating the judgment until the world demarcation point is found; If the last pixel of the column of the image is reached, and still no pixel point meeting the requirement is still available, the column is judged to have no space-earth demarcation point.
8. The method of claim 7, wherein the extracting the world demarcation point based on the image edge, the texture, and the edge gradient distribution in combination with the feature background edge distribution characteristics further comprises: Considering that the air possibly has unmanned aerial vehicle or interference of strong cloud layer edges, when a demarcation point of a suspected antenna and ground wire is found, continuously counting the number of continuous equal to 0 pixel values from the demarcation point downwards, if the number equal to 0 exceeds 10 pixels, considering the row, and if the found antenna and ground wire demarcation point is an error point, eliminating the antenna and ground demarcation point obtained by the row, and re-solving the demarcation point until reaching the lower edge of an image; If no demarcation point has been found, the column is considered to have no world demarcation point, and the lower edge of the column image is taken as the world demarcation point.
9. The rapid antenna and ground wire detection device based on image edge gradient distribution is applied to the method of any one of claims 1 to 8, and is characterized by comprising a noise removal module, an edge extraction module, a demarcation point extraction module and a demarcation line determination module, wherein: the noise removing module is used for carrying out improved median filtering on the image to be identified and removing Gaussian white noise introduced by the imaging of the detector; the edge extraction module is used for adaptively selecting an edge threshold value by utilizing an improved sobel edge extraction algorithm and extracting the edge and texture of the image; The demarcation point extraction module is used for extracting the world demarcation points by utilizing edge gradient distribution and combining ground object background edge distribution characteristics according to the image edges and the texture structures; and the demarcation line determination module is used for connecting all the world demarcation points to form a world demarcation line.
10. A computer readable storage medium storing a computer program, the computer program being capable of being invoked by a processor to perform the steps of a fast earth wire detection method based on image edge gradient distribution as claimed in claim 1.

Description

Quick earth wire detection method based on image edge gradient distribution Technical Field The document relates to the technical field of photoelectric detection, in particular to a rapid earth wire detection method based on image edge gradient distribution. Background The infrared detection of weak and small targets brings great difficulty to detection due to long distance and weak targets. When land object targets such as mountain bodies and buildings exist in the detection view field, in order to detect weak and small targets, the detection threshold value is set to be low, so that branches, cables and the like in the land object background can be detected as false alarms, and the probability of the false alarms is high. The traditional false alarm removing method adopts a method of using candidate target tracks, wherein the tracks cannot be associated or track coordinates are fixed, and the false targets are considered to be removed. The method aims at the ground object background, the ground object target is motionless, but the relative motion of the static object is caused by the motion of the detector, so that the false alarm rejection method is invalid, and under the ground object background, some moving objects also cause interference to detection. Therefore, a detection method for infrared remote detection of weak and small air targets is needed, and the problem that the false alarm rate is high because the ground object targets are detected when the ground object targets exist in the detection in the prior art is solved. Disclosure of Invention The specification provides a rapid antenna and ground wire detection method based on image edge gradient distribution, which is used for solving the problem that the false alarm rate is high because of the existence of ground object targets when the prior art is used for detection. In a first aspect, the present disclosure provides a method for detecting a fast antenna and ground line based on gradient distribution of an image edge, including: performing improved median filtering on the image to be identified, and removing Gaussian white noise introduced by detector imaging; Using an improved sobel edge extraction algorithm to adaptively select an edge threshold value, and extracting edges and textures of an image; According to the image edge and texture structure, extracting the world demarcation points by utilizing edge gradient distribution and combining the feature background edge distribution characteristics; All the space-ground demarcation points are connected to form a space-ground demarcation line. In a second aspect, the present disclosure provides a rapid antenna and ground line detection device based on image edge gradient distribution, which includes a noise removal module, an edge extraction module, a demarcation point extraction module, and a demarcation line determination module, wherein: the noise removing module is used for carrying out improved median filtering on the image to be identified and removing Gaussian white noise introduced by the imaging of the detector; the edge extraction module is used for adaptively selecting an edge threshold value by utilizing an improved sobel edge extraction algorithm and extracting the edge and texture of the image; The demarcation point extraction module is used for extracting the world demarcation points by utilizing edge gradient distribution and combining ground object background edge distribution characteristics according to the image edges and the texture structures; and the demarcation line determination module is used for connecting all the world demarcation points to form a world demarcation line. In a third aspect, the present description also provides a computer-readable storage medium storing a computer program that is capable of being invoked by a processor to perform the steps of a fast earth wire detection method based on image edge gradient distribution. The beneficial effects of the invention are as follows: The specification discloses a rapid earth and sky line detection method based on image edge gradient distribution, which comprises the steps of carrying out improved median filtering on an image to be identified, removing salt and pepper noise introduced by detector imaging, extracting edges of the image by using an improved edge extraction algorithm, and extracting an earth and sky line according to the image edges and the edge gradient distribution. The problem of prior art when surveying, because the existence of ground object target, the ground object target is detected, the false alarm rate is higher is solved. The method utilizes an improved median filtering program to filter image noise and eliminate the influence of the image noise on edge extraction, utilizes an improved edge extraction algorithm to realize stable extraction of image edges and textures, utilizes texture distribution characteristics of an aerial unmanned plane, a cloud layer and the like to discriminate