CN113393395A - High-dynamic infrared image segmentation threshold self-adaptive calculation method - Google Patents
High-dynamic infrared image segmentation threshold self-adaptive calculation method Download PDFInfo
- Publication number
- CN113393395A CN113393395A CN202110671261.6A CN202110671261A CN113393395A CN 113393395 A CN113393395 A CN 113393395A CN 202110671261 A CN202110671261 A CN 202110671261A CN 113393395 A CN113393395 A CN 113393395A
- Authority
- CN
- China
- Prior art keywords
- histogram
- gray level
- infrared image
- high dynamic
- segmentation threshold
- 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.)
- Granted
Links
- 238000004364 calculation method Methods 0.000 title claims abstract description 27
- 238000003709 image segmentation Methods 0.000 title claims abstract description 27
- 238000000034 method Methods 0.000 claims abstract description 30
- 230000011218 segmentation Effects 0.000 claims description 23
- 238000001514 detection method Methods 0.000 claims description 6
- 230000008569 process Effects 0.000 abstract description 3
- 230000000694 effects Effects 0.000 description 5
- 230000003044 adaptive effect Effects 0.000 description 4
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000013459 approach Methods 0.000 description 1
- 229910002056 binary alloy Inorganic materials 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 238000000605 extraction Methods 0.000 description 1
- 230000006870 function Effects 0.000 description 1
- 230000008303 genetic mechanism Effects 0.000 description 1
- 238000003711 image thresholding Methods 0.000 description 1
- 238000013139 quantization Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/90—Dynamic range modification of images or parts thereof
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/40—Image enhancement or restoration using histogram techniques
-
- 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
- G06T2207/10048—Infrared image
-
- 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
- G06T2207/20—Special algorithmic details
- G06T2207/20172—Image enhancement details
- G06T2207/20208—High dynamic range [HDR] image processing
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Image Analysis (AREA)
Abstract
The invention discloses a high dynamic infrared image segmentation threshold value self-adaptive computing method, which comprises the following steps: the method comprises the following steps: counting a high dynamic infrared image gray level histogram; step two: projecting the gray level histogram into a significant feature histogram; step three: calculating the zero crossing point gray level of the histogram of the significant features; step four: calculating the mean value and standard deviation of positive value intervals of the histogram of the significant features; step five: and calculating a high dynamic infrared image segmentation threshold. Compared with other methods of compressing the gray level of the high dynamic image and then calculating the threshold, the method provided by the invention directly processes on the original gray level histogram, and the threshold calculation is more accurate.
Description
Technical Field
The invention belongs to the technical field of target detection and tracking, and relates to a high dynamic infrared image segmentation threshold value self-adaptive computing method.
Background
Image segmentation is a key step of target detection and tracking, and the result of the image segmentation directly influences target feature extraction and description, and further detection and tracking capability, multi-target judgment, true and false target judgment, memory tracking and the like.
Image segmentation methods can be classified into threshold-based segmentation, region-based segmentation, boundary-based segmentation, and the like. The segmentation method based on threshold calculation is low in complexity, easy to process in real time and wide in application in project engineering. The purpose of the threshold segmentation is to find a suitable segmentation threshold by which the whole image or a local region can be divided into two parts, namely a foreground part and a background part. The gray level histogram of the image can well reflect gray level distribution information in an image, so that the gray level histogram of the image is an important reference basis for threshold value selection.
The gray level histogram adopted in the existing threshold segmentation method mainly comprises a one-dimensional gray level histogram and a two-dimensional gray level histogram. The one-dimensional gray level histogram only contains image gray level information, and the threshold value can be calculated through contrast difference or entropy algorithm, so that the method is generally suitable for the condition that the image foreground and background are large in area and the gray level is uniformly distributed. The two-dimensional gray level histogram increases neighborhood average gray level information on the basis of the one-dimensional gray level histogram to form a gray level-neighborhood average two-dimensional gray level histogram. On the basis of a two-dimensional gray histogram, a segmentation threshold can be calculated by adopting a region division method or an entropy algorithm, the method utilizes space information in different forms, the threshold segmentation effect is good, the robustness is high, but the utilization of the space information is limited, and the image information cannot be completely reflected, so that the adaptability of different images is poor. A periodical ' robust image thresholding based on GLGM histogram ' J ' published by XiaoY et al, Pattern registration Letters,2014,40:47-55, a GLGM (gray level and gradient-amplitude) histogram is proposed as a new threshold segmentation image histogram, the GLGM histogram simultaneously displays gray level occurrence probability and spatial distribution characteristics, the space information is effectively represented by using Fibonacci quantization gradient amplitude, the image threshold is calculated by adopting the maximum entropy, the effectiveness and robustness of the threshold method are verified, a large number of discrete points need to be measured, the precision of a solution vector is high, and the reliability is good; the invention discloses an image segmentation method combining a binary algorithm and a two-dimensional linear cross entropy, which is invented by Hoffonius and the like and is named as CN201410172544.6, and the image segmentation method adopts a gray level-neighborhood average two-dimensional gray histogram of an image to construct the two-dimensional linear cross entropy, uses a straight line perpendicular to a main diagonal line to segment the image into a target part and a background part, then uses the binary system to code individuals, uses the two-dimensional linear cross entropy as a fitness function, provides a binary bee colony algorithm self-adaptive threshold acquisition method based on a genetic mechanism, and has good convergence performance and the capability of quickly, stably and accurately acquiring the segmentation threshold. However, the high dynamic infrared image has a wide gray distribution range and a large number of gray levels belonging to long-tail distribution, so that the entropy calculation of each gray level takes a long time, and the difference between some gray levels is small, so that it is difficult to determine a stable threshold. Therefore, more stable, effective and real-time technical approaches are needed to be researched and searched.
Disclosure of Invention
Objects of the invention
The purpose of the invention is: the high-dynamic infrared image segmentation threshold value self-adaptive calculation method is provided to realize high-dynamic infrared image segmentation, is suitable for detecting weak and small targets in air scenes such as a transparent background, a cloud background and the like, is stable in segmentation threshold value and segmentation effect, is self-adaptive in parameters, does not need parameter adjustment, is low in complexity, and can meet real-time processing requirements.
(II) technical scheme
In order to solve the above technical problem, the present invention provides a high dynamic infrared image segmentation threshold adaptive calculation method, which projects a gray level histogram into a significant feature histogram, and finds a segmentation threshold in a feature value space, and mainly includes the following steps:
the method comprises the following steps: counting a high dynamic infrared image gray level histogram;
step two: projecting the gray level histogram into a significant feature histogram;
step three: calculating the zero crossing point gray level of the histogram of the significant features;
step four: calculating the mean value and standard deviation of positive value intervals of the histogram of the significant features;
step five: and calculating a high dynamic infrared image segmentation threshold.
(III) advantageous effects
The high dynamic infrared image segmentation threshold self-adaptive calculation method provided by the technical scheme has the following advantages:
(1) the gray level histogram is projected to be the significant feature histogram, so that the difference between the target gray level and the background is enhanced, the stable threshold value is convenient to calculate, and meanwhile, the segmentation capability of the weak and small target with small gray level difference is also improved.
(2) The one-dimensional histogram is adopted for statistics and calculation, and the method is more suitable for processing high-dynamic images. If a two-dimensional histogram is used, the amount of calculation increases geometrically. Compared with other methods of compressing the gray level of the high dynamic image and then calculating the threshold, the method provided by the invention directly processes on the original gray level histogram, and the threshold calculation is more accurate.
Drawings
FIG. 1 is a high dynamic long wave infrared raw image.
Fig. 2 is a schematic diagram of a grayscale histogram projected as a histogram of salient features.
FIG. 3 is a flowchart of a high dynamic infrared image segmentation threshold adaptive calculation method according to the present invention.
Detailed Description
In order to make the objects, contents, and advantages of the present invention clearer, the following detailed description of the embodiments of the present invention will be made in conjunction with the accompanying drawings and examples.
The invention mainly aims to provide an image segmentation threshold value self-adaptive calculation method to realize high-dynamic infrared image segmentation, is suitable for detecting weak and small targets in air scenes such as a transparent background, a cloud background and the like, has stable segmentation threshold value and segmentation effect, is adaptive to parameters, does not need to adjust parameters, has low complexity and can be used for real-time target detection and tracking.
Referring to fig. 1 to 3, the high dynamic infrared image segmentation threshold value self-adaptive calculation method of the present invention includes the following steps:
the first step is as follows: counting a gray level histogram;
for a high dynamic infrared image I, the gray level number is L (2)1416384), the image size is M × N, and the statistical grayscale histogram H is H ═ l0,l1,…,lL}。
In this embodiment, the resolution of the high-dynamic long-wave infrared image is 1024 × 768, and the bit width of the image data is 14 bits.
The second step is that: projecting the gray level histogram into a significant feature histogram;
traversing the gray level histogram H, calculating the gray level by gray level to obtain a significant feature histogram S which is S ═ S0,s1,…,sLThe calculation method is gray level by gray level
Wherein lkThe number of points of the k-th gray level of the histogram H, and (k-n) the difference of the gray levels.
The third step: calculating the zero-crossing gray level of the histogram of the significant features:
since the significant feature histogram S is a monotone increasing curve from negative to positive, the zero-crossing gray level z can be calculated. The calculation method is to traverse the histogram S of the significant features and find the histogram satisfying Sz-1< 0 and szA gray level z of > 0.
The fourth step: calculating the mean and standard deviation of the positive interval of the histogram of the significant features:
taking the zero-crossing gray level z as a starting point, and intercepting a positive value interval S of the significant feature histogram from the significant feature histogram Sp＝{sz,sz+1,sz+2,…,sLAnd calculate SpMean value of Sp-aveAnd standard deviation Sp-var. The calculation formula is as follows:
the fifth step: calculating a high dynamic infrared image segmentation threshold value:
calculating a positive value interval S of the histogram of the salient featurespThe calculation formula of the segmentation threshold value T is as follows:
T＝Sp-ave+a*Sp-var
traversing positive interval S of histogram of significant featurespFind out the satisfaction sth-1< T and sthThe gray level th is more than or equal to T, namely the high dynamic infrared image segmentation threshold lthAnd a represents a variance coefficient of the histogram of the salient features. In this embodiment, a is preferably 5.
According to the technical scheme, the method and the device can realize high-dynamic infrared image segmentation, can be suitable for segmentation detection of weak and small targets in an air scene, are stable in segmentation threshold value and segmentation effect, adaptive to parameters, free of parameter adjustment and low in complexity, and can meet real-time processing requirements, so that the method and the device have good application prospects. The invention can also be applied to processing of highly dynamic visible light images.
The above description is only a preferred embodiment of the present invention, and it should be noted that, for those skilled in the art, several modifications and variations can be made without departing from the technical principle of the present invention, and these modifications and variations should also be regarded as the protection scope of the present invention.
Claims (10)
1. A high dynamic infrared image segmentation threshold value self-adaptive calculation method is characterized by comprising the following steps:
the method comprises the following steps: counting a high dynamic infrared image gray level histogram;
step two: projecting the gray level histogram into a significant feature histogram;
step three: calculating the zero crossing point gray level of the histogram of the significant features;
step four: calculating the mean value and standard deviation of positive value intervals of the histogram of the significant features;
step five: and calculating a high dynamic infrared image segmentation threshold.
2. The method as claimed in claim 1, wherein in the first step, for one high dynamic infrared image I, the number of gray levels is L, the image size is mxn, and the statistical histogram of gray levels H is H ═ L0,l1,…,lL}。
3. The method as claimed in claim 2, wherein in the second step, the histogram H is traversed, and the calculation is performed on a gray level by gray level basis to obtain a histogram S of significant features (S ═ S)0,s1,…,sLThe calculation method is gray level by gray level
Wherein lkThe number of points of the k-th gray level of the histogram H, and (k-n) the difference of the gray levels.
4. The method according to claim 3, wherein in step three, the zero-crossing gray level z is calculated by traversing a histogram S of salient features and finding a value satisfying Sz-1< 0 and szA gray level z of > 0.
5. The method according to claim 4, wherein in the fourth step, the zero-crossing gray level z is used as a starting point, and a positive interval S of the histogram S of the salient features is extracted from the histogram S of the salient featuresp＝{sz,sz+1,sz+2,…,sLAnd calculate SpMean value of Sp-aveAnd standard deviation Sp-var(ii) a The calculation formula is as follows:
6. the high dynamic infrared image segmentation threshold value self-adaptive computing method of claim 5, characterized in that in the fifth step, a positive value interval S of a significant feature histogram is computedpThe calculation formula of the segmentation threshold value T is as follows:
T＝Sp-ave+a*Sp-var
traversing positive interval S of histogram of significant featurespFind out the satisfaction sth-1< T and sthThe gray level th is more than or equal to T, namely the high dynamic infrared image segmentation threshold lthAnd a represents the variance system of the histogram of the significant featuresAnd (4) counting.
7. The method for adaptively calculating the segmentation threshold of the high dynamic infrared image as claimed in claim 6, wherein in the first step, the resolution of the high dynamic long wave infrared image is 1024 x 768, and the bit width of the image data is 14 bits.
8. The method according to claim 7, wherein in the first step, L is 214＝16384。
9. The method according to claim 8, wherein in the step five, a is 5.
10. Application of the high dynamic infrared image segmentation threshold value self-adaptive computing method based on any one of claims 1-9 in the technical field of target detection and tracking.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110671261.6A CN113393395B (en) | 2021-06-17 | 2021-06-17 | Adaptive calculation method for high-dynamic infrared image segmentation threshold |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110671261.6A CN113393395B (en) | 2021-06-17 | 2021-06-17 | Adaptive calculation method for high-dynamic infrared image segmentation threshold |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN113393395A true CN113393395A (en) | 2021-09-14 |
| CN113393395B CN113393395B (en) | 2023-10-31 |
Family
ID=77621681
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110671261.6A Active CN113393395B (en) | 2021-06-17 | 2021-06-17 | Adaptive calculation method for high-dynamic infrared image segmentation threshold |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113393395B (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114937055A (en) * | 2022-03-31 | 2022-08-23 | 江苏益捷思信息科技有限公司 | Image self-adaptive segmentation method and system based on artificial intelligence |
Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20060170769A1 (en) * | 2005-01-31 | 2006-08-03 | Jianpeng Zhou | Human and object recognition in digital video |
| CN104700416A (en) * | 2015-03-23 | 2015-06-10 | 河海大学常州校区 | Image segmentation threshold determination method based on visual analysis |
| CN105844603A (en) * | 2016-04-04 | 2016-08-10 | 上海大学 | High dynamic range infrared image dynamic scope compression method |
| CN106157323A (en) * | 2016-08-30 | 2016-11-23 | 西安工程大学 | The insulator division and extracting method that a kind of dynamic division threshold value and block search combine |
| US20180033125A1 (en) * | 2016-07-27 | 2018-02-01 | Sharp Laboratories Of America, Inc. | System for conversion of low dynamic range images to high dynamic range images |
| CN111062954A (en) * | 2019-12-30 | 2020-04-24 | 中国科学院长春光学精密机械与物理研究所 | Infrared image segmentation method, device and equipment based on difference information statistics |
| CN111179295A (en) * | 2019-12-23 | 2020-05-19 | 青海大学 | Improved two-dimensional Otsu threshold image segmentation method and system |
| WO2021000524A1 (en) * | 2019-07-03 | 2021-01-07 | 研祥智能科技股份有限公司 | Hole protection cap detection method and apparatus, computer device and storage medium |
-
2021
- 2021-06-17 CN CN202110671261.6A patent/CN113393395B/en active Active
Patent Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20060170769A1 (en) * | 2005-01-31 | 2006-08-03 | Jianpeng Zhou | Human and object recognition in digital video |
| CN104700416A (en) * | 2015-03-23 | 2015-06-10 | 河海大学常州校区 | Image segmentation threshold determination method based on visual analysis |
| CN105844603A (en) * | 2016-04-04 | 2016-08-10 | 上海大学 | High dynamic range infrared image dynamic scope compression method |
| US20180033125A1 (en) * | 2016-07-27 | 2018-02-01 | Sharp Laboratories Of America, Inc. | System for conversion of low dynamic range images to high dynamic range images |
| CN106157323A (en) * | 2016-08-30 | 2016-11-23 | 西安工程大学 | The insulator division and extracting method that a kind of dynamic division threshold value and block search combine |
| WO2021000524A1 (en) * | 2019-07-03 | 2021-01-07 | 研祥智能科技股份有限公司 | Hole protection cap detection method and apparatus, computer device and storage medium |
| CN111179295A (en) * | 2019-12-23 | 2020-05-19 | 青海大学 | Improved two-dimensional Otsu threshold image segmentation method and system |
| CN111062954A (en) * | 2019-12-30 | 2020-04-24 | 中国科学院长春光学精密机械与物理研究所 | Infrared image segmentation method, device and equipment based on difference information statistics |
Non-Patent Citations (3)
| Title |
|---|
| 刘媛;赵敏;: "野外大空间火焰目标红外图像的分割算法研究", 青岛大学学报(工程技术版), no. 01 * |
| 周武根: "动态环境下稠密视觉同时定位与地图构建方法研究", 中国优秀博士学位论文全文数据库-信息科技辑 * |
| 郭飞: "视觉假体中视觉图像信息感知关键技术研究", 中国优秀博士学位论文全文数据库信息科技辑 * |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114937055A (en) * | 2022-03-31 | 2022-08-23 | 江苏益捷思信息科技有限公司 | Image self-adaptive segmentation method and system based on artificial intelligence |
| CN114937055B (en) * | 2022-03-31 | 2024-05-03 | 厦门市虹约产品设计有限公司 | Image self-adaptive segmentation method and system based on artificial intelligence |
Also Published As
| Publication number | Publication date |
|---|---|
| CN113393395B (en) | 2023-10-31 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111144213B (en) | Object detection method and related equipment | |
| CN109086724B (en) | Accelerated human face detection method and storage medium | |
| Zuo et al. | Explicit edge inconsistency evaluation model for color-guided depth map enhancement | |
| CN103337077B (en) | A kind of based on multi-scale division and the visible ray of SIFT and the method for infrared image registration | |
| CN111882586B (en) | Multi-actor target tracking method oriented to theater environment | |
| CN109472267A (en) | Picture similarity alignment algorithm | |
| CN105160649A (en) | Multi-target tracking method and system based on kernel function unsupervised clustering | |
| CN107403451B (en) | Self-adaptive binary characteristic monocular vision odometer method, computer and robot | |
| CN107301649B (en) | Regional merged SAR image coastline detection algorithm based on superpixels | |
| CN113034383B (en) | Method for acquiring video image based on improved grid motion statistics | |
| CN104715251A (en) | Salient object detection method based on histogram linear fitting | |
| CN111709893B (en) | ORB-SLAM2 improved algorithm based on information entropy and sharpening adjustment | |
| CN111881837B (en) | Shadow extraction-based video SAR moving target detection method | |
| Sun et al. | Fusing Kalman filter with TLD algorithm for target tracking | |
| CN110097585B (en) | SAR image matching method and system based on SIFT algorithm | |
| CN112183434A (en) | Building change detection method and device | |
| Prakoso et al. | Vehicle detection using background subtraction and clustering algorithms | |
| CN113393395A (en) | High-dynamic infrared image segmentation threshold self-adaptive calculation method | |
| CN113160097A (en) | Infrared image quantization method based on histogram transformation | |
| CN113570554A (en) | Single image visibility detection method based on scene depth | |
| CN108985292A (en) | A kind of SAR image CFAR object detection method and system based on multi-scale division | |
| CN110502968B (en) | Method for detecting infrared small and weak moving target based on track point space-time consistency | |
| CN116678418A (en) | Improved laser SLAM quick loop-back detection method | |
| CN108010076B (en) | End face appearance modeling method for intensive industrial bar image detection | |
| CN107067411B (en) | Mean-shift tracking method combined with dense features |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |