CN110782471A - Multi-scale SAR image edge detection method - Google Patents

Multi-scale SAR image edge detection method Download PDF

Info

Publication number
CN110782471A
CN110782471A CN201910981758.0A CN201910981758A CN110782471A CN 110782471 A CN110782471 A CN 110782471A CN 201910981758 A CN201910981758 A CN 201910981758A CN 110782471 A CN110782471 A CN 110782471A
Authority
CN
China
Prior art keywords
image
gradient
pixel
pixels
value
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Pending
Application number
CN201910981758.0A
Other languages
Chinese (zh)
Inventor
郎丰铠
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
China University of Mining and Technology CUMT
Original Assignee
China University of Mining and Technology CUMT
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by China University of Mining and Technology CUMT filed Critical China University of Mining and Technology CUMT
Priority to CN201910981758.0A priority Critical patent/CN110782471A/en
Publication of CN110782471A publication Critical patent/CN110782471A/en
Pending legal-status Critical Current

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/10—Segmentation; Edge detection
    • G06T7/13—Edge detection
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/10—Segmentation; Edge detection
    • G06T7/136—Segmentation; Edge detection involving thresholding
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/10—Image acquisition modality
    • G06T2207/10032—Satellite or aerial image; Remote sensing
    • G06T2207/10044—Radar image

Landscapes

  • Engineering & Computer Science (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Radar Systems Or Details Thereof (AREA)

Abstract

The invention discloses a multi-scale SAR image edge detection method, which is characterized in that edge gradients in different directions in an image are calculated by using windows with different sizes, then the maximum value and the corresponding direction information in all the gradients with different sizes and different directions are retained, and finally an edge image with the final width of one pixel is obtained through non-maximum value inhibition.

Description

Multi-scale SAR image edge detection method
Technical Field
The invention belongs to the technical field of image processing, and particularly relates to a multi-scale SAR image edge detection method.
Background
Edge detection is one of the most basic problems in the field of image processing, because edges in an image are the most basic and unchangeable features in the image, are the places where information is most concentrated in the image, and are the basis for processing such as image feature extraction, target recognition, image segmentation, and the like. However, edge information in an image is represented in the image in various ways due to the influence of factors such as a sensor, an imaging principle, an imaging position, and an imaging object in the imaging process, which makes it difficult to accurately detect the edge.
The current image edge detection algorithm can be roughly classified into: 1) an edge detection algorithm based on spatial gradient; 2) wavelet edge detection algorithm based on frequency domain; 3) an edge detection algorithm based on machine learning; 4) and (3) an edge detection algorithm based on other theoretical technologies. Among these methods, the spatial gradient-based edge detection algorithm is the most classical and most influential, and Canny edge detection algorithm is the most typical. However, the detection windows of these algorithms are all fixed in size and perform poorly when the image noise level is high.
In the field of Synthetic Aperture Radar (SAR) image processing, since speckle noise is strong, a general edge detection algorithm is difficult to achieve an ideal effect, and a commonly used edge detector is a Ratio of average (ROA) detector. The detector is a spatial gradient-based detector, the detection window and the direction of the detector can be manually adjusted as parameters, and the detector has better noise resistance because a region mean value is adopted to replace a single pixel value. However, the window size of the detector is still single, whereas for an image the appropriate detection window size may be different in different regions. For example, in an area rich in detail information, it is generally more suitable to use a small window for detection, because when a larger window is used for edge detection, the edge positions detected around some targets such as points, lines, etc. have larger deviations; areas with weaker edge information and more homogeneous pixels are generally suitable for detection with a large window, because a small window is sensitive to noise and has more detected false edges. Therefore, a single size detection window cannot meet the requirements at the same time.
Disclosure of Invention
The invention aims to overcome the defect that the conventional ROA edge detector cannot meet the requirement by using a detection window with a single size, and provides a multi-scale SAR image edge detection method comprehensively using detection windows with different sizes.
The technical scheme is as follows: in order to realize the purpose of the invention, the technical scheme adopted by the invention is as follows: a multi-scale SAR image edge detection method comprises the following steps:
step 1: setting a maximum detection window W maxAnd a minimum detection window W minObtaining N W＝(W max–W min) 2+1 windows W of different sizes;
step 2: for N WSetting the parameters l, W, d and theta of the detection template for each window with different sizes, wherein each window with the size of W can beObtaining N θPi/theta templates, then obtaining N-N W*N θTemplates with different sizes W and different directions theta; wherein l is the template length, w is the template width, d is the template region R 1And R 2The distance between the two plates, theta is the angle corresponding to the template;
and step 3: for each pixel p, calculating a template region R by using N templates with different sizes W and different directions theta 1And R 2Gradient D between W(θ), obtaining N gradient values and corresponding angle values in total;
and 4, step 4: selecting the maximum value of the N gradient values, recording the value and the corresponding angle thereof, and continuously processing the next pixel until all the pixels are processed to obtain a gradient image and an angle image;
and 5: carrying out non-maximum suppression in the gradient direction to obtain a refined gradient image; the method comprises the following specific steps:
for each pixel p in the gradient image, comparing the magnitude of adjacent pixel values in the direction normal to the angle indicated in the corresponding angle image; if the pixel p is smaller than the adjacent pixels, the gradient value of the pixel p is set to 0, otherwise, the original value of the pixel p is reserved, and a refined gradient image is obtained after all the pixels are processed;
step 6: and carrying out binarization processing on the thinned gradient image to obtain a final edge image with a single-pixel width.
Further, in step 3, for the SAR image, the gradient calculation formula is:
Figure BDA0002235413880000021
wherein x is 1、x 2Respectively representing the template areas R in the direction theta 1And region R 2The SAR image amplitude or intensity;
for a polarized SAR image, the formula for the gradient calculation is:
D W(θ)＝2ln|X 1+X 2|-ln|X 1|-ln|X 2|+2qln2
wherein, X 1、X 2Respectively representing the template areas R in the direction theta 1And region R 2Q × q, q representing the matrix dimension.
Further, in step 6, an entropy threshold method is used for binarization processing, which specifically includes:
step 6.1: and carrying out image enhancement processing on the refined gradient image, wherein the method comprises the following steps:
firstly, carrying out logarithmic transformation on the refined gradient image, wherein the calculation formula is as follows: x is 10 log 10(x) Wherein x represents a pixel value;
then, maximum value and minimum value stretching is carried out, and the processing method comprises the following steps: sorting image pixel values in a descending order, taking the pixel value at the a% position as a minimum value min, and taking the pixel value at the b% position as a maximum value max;
and finally, carrying out normalization processing by using the maximum value and the minimum value, wherein the normalization formula is as follows: x ═ x-min)/(max-min), where x denotes the pixel value;
step 6.2: dividing the pixels in the gradient image processed in the step 6.1 into n gray levels, and calculating the ratio p of the pixels in each gray level i to the total number of pixels i；
Step 6.3: let the gray level s divide the image into two parts a and B, the entropy of each part is defined as follows:
Figure BDA0002235413880000031
Figure BDA0002235413880000032
Figure BDA0002235413880000033
the sum of the entropies of a and B is then: h sum＝H A+H BCalculating H corresponding to each gray level i sum；
Step 6.4: finding the largest H among n gray levels sumAnd the corresponding gray level s is obtained, s is the segmentation threshold, the pixels less than or equal to s are set as 0, and the pixels more than s are set as 1, so that the binary image is obtained.
Has the advantages that: compared with the prior art, the technical scheme of the invention has the following beneficial technical effects:
the invention provides a method for detecting the edge of an image by utilizing window templates with different sizes, and integrates the edge information obtained by each window to obtain a final edge detection result, thereby overcoming the defect that the conventional single window for edge detection can not consider both a texture rich area and a weak edge area, not only detecting possible edges to the maximum extent, but also ensuring the accuracy of edge detection.
Drawings
FIG. 1 is a flow chart of the present invention;
FIG. 2 is a diagram of an edge detection template in an embodiment of the invention.
Detailed Description
The technical solution of the present invention is further described below with reference to the accompanying drawings and examples.
As shown in fig. 1, the method for detecting the edge of the multi-scale SAR image according to the present invention includes the following steps:
step 1: setting a maximum detection window W maxAnd a minimum detection window W minObtaining N W＝(W max–W min) 2+1 windows W of different sizes;
the present embodiment is set up W max＝7，W minWhen it is 3, it has N WThe 3 windows with different sizes, namely W is 3, 5 and 7.
Step 2: setting template parameters l, w, d and theta of 3 windows with different sizes respectively, as shown in FIG. 2; n is available for each window of size W θPi/theta templates, then obtaining N-N W*N θTemplates with different sizes W and different directions theta; wherein l is the template length, w is the template width, d is the template region R 1And R 2Is a distance between the two templates, theta is a template pairThe corresponding angle; the template is generally square, i.e., l 2 × w + d;
in the present embodiment, when the window size is 3, l is 3, w is 1, d is 1, and θ is 45 °; when the window size is 5, setting l to 5, w to 2, d to 1, and θ to 45 °; when the window size is 7, setting l to 7, w to 3, d to 1, and θ to 45 °; there are 12 templates of different sizes and different orientations, N being 3 (180 °/45 °).
And step 3: for each pixel p, a template region R is calculated by using 12 templates with different sizes and different directions 1And R 2Gradient D between W(θ), obtaining 12 gradient values and corresponding angle values in total;
in this embodiment, for the SAR image, the calculation formula of the gradient in step 3 is:
wherein x is 1、x 2Respectively representing the template areas R in the direction theta 1And region R 2The SAR image amplitude or intensity;
in this embodiment, for a polarized SAR image, the gradient calculation formula in step 3 is:
D W(θ)＝2ln|X 1+X 2|-ln|X 1|-ln|X 2|+2qln2
wherein, X 1、X 2Respectively representing the template areas R in the direction theta 1And region R 2Q × q, q representing the matrix dimension.
And 4, step 4: calculating the maximum value of the 12 gradient values, recording the value and the corresponding angle, and continuously processing the next pixel until all the pixels are processed, and obtaining a gradient image and an angle image after all the pixels are processed.
And 5: carrying out non-maximum suppression in the gradient direction to obtain a refined gradient image; the method comprises the following specific steps:
and for each pixel p in the gradient image, comparing the values of adjacent pixels along the normal direction of the angle indicated in the corresponding angle image, if the pixel p is smaller than the adjacent pixels, setting the gradient value of the pixel p to be 0, otherwise, keeping the original value of the pixel p, and obtaining a refined gradient image after all pixels are processed.
Step 6: carrying out binarization processing on the thinned gradient image to obtain a final edge image with a single-pixel width;
in this embodiment, the binarization processing is performed by using an entropy threshold method, which specifically includes:
step 6.1: and carrying out image enhancement processing on the refined gradient image, wherein the method comprises the following steps:
firstly, carrying out logarithmic transformation on the refined gradient image, wherein the calculation formula is as follows: x is 10 log 10(x) Wherein x represents a pixel value;
then, maximum value and minimum value stretching is carried out, and the processing method comprises the following steps: sorting the image pixel values in the order from small to large, taking the pixel value at the 6.25% position as the minimum value min, and taking the pixel value at the 93.75% position as the maximum value max;
and finally, carrying out normalization processing by using the maximum value and the minimum value, wherein the normalization formula is as follows: x ═ x-min)/(max-min), where x denotes the pixel value;
step 6.2: dividing the pixels in the gradient image processed in the step 6.1 into n gray levels, and calculating the ratio p of the pixels in each gray level i to the total number of pixels i(ii) a In this example, n is 256;
step 6.3: assuming that the gray level s divides the image into two parts a and B, the entropy of each part is defined as follows:
Figure BDA0002235413880000042
Figure BDA0002235413880000043
the sum of the entropies of a and B is then: h sum＝H A+H BCalculating H corresponding to each gray level i sum；
Step 6.4: finding the largest H among n gray levels sumAnd the corresponding gray level s is obtained, s is the segmentation threshold, the pixels less than or equal to s are set as 0, and the pixels more than s are set as 1, so that the binary image is obtained.
The specific embodiments described herein are merely illustrative of the spirit of the invention. Various modifications or additions may be made to the described embodiments or alternatives in a similar manner to those skilled in the art to which the present invention pertains. If the template parameter theta can be set to other values; the gradient may be replaced by other formulas; other methods can be selected for the binarization method to replace the entropy threshold method; and so on. This is not to be regarded as a departure from the spirit of the invention, or beyond the scope of the appended claims.

Claims (6)

1. A multi-scale SAR image edge detection method is characterized in that: the method comprises the following steps:
step 1: setting a maximum detection window W maxAnd a minimum detection window W minObtaining N W＝(W max–W min) 2+1 windows W of different sizes;
step 2: for N WSetting detection template parameters l, W, d and theta of windows with different sizes, and obtaining N for each window with the size of W θPi/theta templates, then obtaining N-N W*N θTemplates with different sizes W and different directions theta; wherein l is the template length, w is the template width, d is the template region R 1And R 2The distance between the two plates, theta is the angle corresponding to the template;
and step 3: for each pixel p, calculating a template region R by using N templates with different sizes W and different directions theta 1And R 2Gradient D between W(θ), obtaining N gradient values and corresponding angle values in total;
and 4, step 4: selecting the maximum value of the N gradient values, recording the value and the corresponding angle thereof, and continuously processing the next pixel until all the pixels are processed to obtain a gradient image and an angle image;
and 5: carrying out non-maximum suppression in the gradient direction to obtain a refined gradient image;
step 6: and carrying out binarization processing on the thinned gradient image to obtain a final edge image with a single-pixel width.
2. The edge detection method of the multi-scale SAR image according to claim 1, characterized in that: in step 3, for the SAR image, a gradient calculation formula is:
Figure FDA0002235413870000011
wherein x is 1、x 2Respectively representing the template areas R in the direction theta 1And region R 2The SAR image amplitude or intensity;
for a polarized SAR image, the formula for the gradient calculation is:
D W(θ)＝2ln|X 1+X 2|-ln|X 1|-ln|X 2|+2q ln2
wherein, X 1、X 2Respectively representing the template areas R in the direction theta 1And region R 2Q × q, q representing the matrix dimension.
3. The edge detection method of the multi-scale SAR image according to claim 1, characterized in that: in the step 5, performing non-maximum suppression in the gradient direction to obtain a refined gradient image; the method comprises the following specific steps:
for each pixel p in the gradient image, comparing the magnitude of adjacent pixel values in the direction normal to the angle indicated in the corresponding angle image; if the pixel p is smaller than the adjacent pixels, the gradient value of the pixel p is set to be 0, otherwise, the original value of the pixel p is reserved, and the refined gradient image is obtained after all the pixels are processed.
4. The edge detection method for the multi-scale SAR image according to any one of claims 1-3, characterized in that: in the step 6, an entropy threshold method is adopted for binarization processing, which specifically includes:
step 6.1: carrying out image enhancement processing on the refined gradient image;
step 6.2: dividing the pixels in the gradient image processed in the step 6.1 into n gray levels, and calculating the ratio p of the pixels in each gray level i to the total number of pixels i；
Step 6.3: the image is divided into two parts A and B by a gray level s, and the entropy of each part is respectively H AAnd H B(ii) a The sum of the entropies of a and B is then: h sum＝H A+H BCalculating H corresponding to each gray level i sum；
Step 6.4: finding the largest H among n gray levels sumAnd the corresponding gray level s is obtained, s is the segmentation threshold, the pixels less than or equal to s are set as 0, and the pixels more than s are set as 1, so that the binary image is obtained.
5. The edge detection method of the multi-scale SAR image according to claim 4, characterized in that: the step 6.1 is to perform image enhancement processing on the refined gradient image, and the method comprises the following steps:
firstly, carrying out logarithmic transformation on the refined gradient image, wherein the calculation formula is as follows: x is 10 log 10(x) Wherein x represents a pixel value;
then, maximum value and minimum value stretching is carried out, and the processing method comprises the following steps: sorting image pixel values in a descending order, taking the pixel value at the a% position as a minimum value min, and taking the pixel value at the b% position as a maximum value max;
and finally, carrying out normalization processing by using the maximum value and the minimum value, wherein the normalization formula is as follows: x is (x-min)/(max-min), where x represents the pixel value.
6. The edge detection method of the multi-scale SAR image according to claim 4, characterized in that: in step 6.3, the entropy of the two parts a and B corresponding to the gray level s is defined as follows:
Figure FDA0002235413870000021
Figure FDA0002235413870000022
Figure FDA0002235413870000023
CN201910981758.0A 2019-10-16 2019-10-16 Multi-scale SAR image edge detection method Pending CN110782471A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201910981758.0A CN110782471A (en) 2019-10-16 2019-10-16 Multi-scale SAR image edge detection method

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201910981758.0A CN110782471A (en) 2019-10-16 2019-10-16 Multi-scale SAR image edge detection method

Publications (1)

Publication Number Publication Date
CN110782471A true CN110782471A (en) 2020-02-11

Family

ID=69385591

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201910981758.0A Pending CN110782471A (en) 2019-10-16 2019-10-16 Multi-scale SAR image edge detection method

Country Status (1)

Country Link
CN (1) CN110782471A (en)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113298839A (en) * 2021-05-18 2021-08-24 中国矿业大学 Semi-automatic threshold extraction method for SAR (synthetic aperture radar) image

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103778613A (en) * 2014-02-21 2014-05-07 武汉大学 Polarization SAR image filtering method for window self-adaptation
CN103886562A (en) * 2014-04-14 2014-06-25 苏州大学 SAR image edge detection method and system
CN105930847A (en) * 2016-03-31 2016-09-07 中国人民解放军空军航空大学 Combined edge detection-base SAR image linear feature extraction method
CN106683107A (en) * 2016-11-08 2017-05-17 成都理工大学 SAR image edge detection method based on ROEWA improvement
CN108961290A (en) * 2018-07-10 2018-12-07 中国计量大学 A kind of Ratio operator adapting to image edge detection method based on Otsu

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103778613A (en) * 2014-02-21 2014-05-07 武汉大学 Polarization SAR image filtering method for window self-adaptation
CN103886562A (en) * 2014-04-14 2014-06-25 苏州大学 SAR image edge detection method and system
CN105930847A (en) * 2016-03-31 2016-09-07 中国人民解放军空军航空大学 Combined edge detection-base SAR image linear feature extraction method
CN106683107A (en) * 2016-11-08 2017-05-17 成都理工大学 SAR image edge detection method based on ROEWA improvement
CN108961290A (en) * 2018-07-10 2018-12-07 中国计量大学 A kind of Ratio operator adapting to image edge detection method based on Otsu

Non-Patent Citations (3)

* Cited by examiner, † Cited by third party
Title
JESPER SCHOU等: "CFAR Edge Detector for Polarimetric SAR Images", 《IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING》 *
沈夏炯 等: "一种面向SAR图像的多窗口道路边缘检测算法", 《计算机工程与应用》 *
郎丰铠: "极化SAR影像滤波及分割方法研究", 《中国优秀博硕学位论文全文数据库（博士）基础科学辑》 *

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113298839A (en) * 2021-05-18 2021-08-24 中国矿业大学 Semi-automatic threshold extraction method for SAR (synthetic aperture radar) image
CN113298839B (en) * 2021-05-18 2023-08-15 中国矿业大学 SAR image semiautomatic threshold extraction method

Similar Documents

Publication Publication Date Title
CN107301661B (en) High-resolution remote sensing image registration method based on edge point features
US9361672B2 (en) Image blur detection
CN111080661B (en) Image-based straight line detection method and device and electronic equipment
CN102156996B (en) Image edge detection method
CN108898132B (en) Terahertz image dangerous article identification method based on shape context description
CN103778613B (en) A kind of adaptive polarization SAR images filter method of window
CN114399522A (en) High-low threshold-based Canny operator edge detection method
Krishnan et al. A survey on different edge detection techniques for image segmentation
CN109242032B (en) Target detection method based on deep learning
Han et al. Research and implementation of an improved canny edge detection algorithm
Reddy et al. Comparative analysis of common edge detection algorithms using pre-processing technique
Perumal et al. Multilevel morphological fuzzy edge detection for color images (MMFED)
CN114943744A (en) Edge detection method based on local Otsu thresholding
Othman et al. Road crack detection using adaptive multi resolution thresholding techniques
Chen et al. An improved edge detection in noisy image using fuzzy enhancement
CN110782471A (en) Multi-scale SAR image edge detection method
CN110321828B (en) Front vehicle detection method based on binocular camera and vehicle bottom shadow
Sun Image edge detection based on relative degree of grey incidence and Sobel operator
Sornam et al. Hysteresis thresholding based edge detectors for inscriptional image enhancement
CN108961290B (en) Adaptive image edge detection method based on Ratio operator of Otsu
Janoriya et al. Critical review on edge detection techniques in spatial domain on low illumination images
Guo et al. An adaptive soft-morphological-gradient-filter-for-edge-detection
Yuan et al. Automatic detection of solar features in HSOS full-disk solar images using guided filter
SKiran A framework in shadow detection and compensation of images
Mustafa et al. Effect of direct statistical contrast enhancement technique on document image binarization

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination