CN108364299B - Automatic threshold segmentation method for low-quality shoe print image - Google Patents
Automatic threshold segmentation method for low-quality shoe print image Download PDFInfo
- Publication number
- CN108364299B CN108364299B CN201810039047.7A CN201810039047A CN108364299B CN 108364299 B CN108364299 B CN 108364299B CN 201810039047 A CN201810039047 A CN 201810039047A CN 108364299 B CN108364299 B CN 108364299B
- Authority
- CN
- China
- Prior art keywords
- image
- segmentation method
- shoe print
- threshold segmentation
- segmentation
- 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.)
- Active
Links
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
- G06T7/136—Segmentation; Edge detection involving thresholding
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/149—Segmentation; Edge detection involving deformable models, e.g. active contour models
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/194—Segmentation; Edge detection involving foreground-background segmentation
Landscapes
- Engineering & Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Software Systems (AREA)
- Image Analysis (AREA)
- Editing Of Facsimile Originals (AREA)
- Character Input (AREA)
Abstract
The invention discloses an automatic threshold segmentation method for low-quality shoe print images. Firstly, eliminating the influence of poor exposure or equipment artifacts by adopting morphological corrosion and expansion operation; then carrying out gray scale conversion to enhance the image contrast; and finally, obtaining a final shoe printing area by using morphological noise reduction. The method does not need manual interaction, can segment the shoe print images in a large scale and in a full-automatic manner, is obviously superior to a manual segmentation method and an iterative segmentation method in time efficiency, and has higher segmentation accuracy than the traditional global threshold segmentation method and the traditional iterative segmentation method.
Description
Technical Field
The invention relates to the field of digital image processing, in particular to an automatic shoe print image segmentation method which can effectively improve the phenomenon of mistaken segmentation of the traditional global threshold segmentation under the condition of low acquisition quality, and has high segmentation accuracy and good real-time property.
Background
Shoe printing is one of the most common traces on crime scenes, and is also one of the important bases for the reconnaissance to find out clues to solve a case, reveal and confirm the crime. In the shoe print image processing, people often only need to extract a specific shoe print part, and hardly pay attention to the background area. In this case, it is necessary to segment the shoe print portion from the image.
A case will contact two types of shoe print images during the reconnaissance process:
crime showingThe extracted pattern remains as an image of the shoe print left behind by the perpetrator. The shoe print texture pattern in the image is greatly influenced by the ground texture, the difficulty of automatic segmentation by a computer is high, and at present, a manual segmentation mode is mainly adopted.
And (4) locking shoe print images of criminal suspects in the process of solving a crime. The image is imaged by a pressing die, the contrast between the texture of the shoe print and the background is obvious, and the automatic segmentation by a computer is convenient.
Segmentation based on global threshold is one of the most classical and popular image segmentation methods, and is also the simplest one. The basic idea is to find one or several gray threshold values according to the gray histogram of the image, divide the gray level of the image into several levels, and consider the pixels at the same gray level to belong to the same object. The threshold segmentation is particularly suitable for images with different gray level ranges of the target and the background, is simple in calculation, does not need manual intervention, and is widely applied to application occasions with high requirements on operation efficiency. However, for a class of images with small gray scale difference, small difference between the gray scale values of the target and the background, and complicated illumination conditions, the segmentation result is not satisfactory. Especially, when the die imaging device is aged or used for too many times, the phenomenon of underexposure, overexposure or imprint residue occurs in the shoe print image, and the typical automatic threshold segmentation algorithm cannot accurately separate the shoe print pattern from the background.
Although a plurality of image segmentation algorithms are proposed at present, a general image segmentation algorithm suitable for complex illumination conditions does not exist, and particularly, an automatic shoe print image segmentation algorithm which is high in accuracy, good in real-time performance and free of manual interaction is lacked.
Disclosure of Invention
The invention aims to solve the technical problems in the prior art and provides the automatic shoe print segmentation method which is oriented to the low-quality shoe print image, high in segmentation accuracy and good in real-time performance.
The technical solution of the invention is as follows: an automatic threshold segmentation method for low-quality shoe print images is characterized by comprising the following steps:
step 1, inputting an imageIIs converted into HSI and its luminance component is preservedH；
Step 2. use radius ofRThe flat disk-shaped structural element of (2) performs morphological etching treatment on the H to obtain an imageI 1；
Step 3, using the same structural elements as step 2 to imageI 1Performing morphological dilation to obtain an estimated value of the image backgroundB；
Step 4, inputting the imageISubtracting the background estimateBObtaining a shoe print image without the influence of illuminationI 2；
Step 5, statisticsI 2The minimum value of the gray value range of (2) is set toI minMaximum value ofI maxAccording to the formula (1) toI 2Performing linear gray scale stretching to stretch the pixel gray scale range to 0-255:
wherein the content of the first and second substances,xrepresenting the input pixel value to be stretched,yrepresenting the stretched output pixel value. Let the enhanced image beI 3；
Step 6, adopt the traditional Otsu method to rightI 3Performing global threshold segmentation to obtain a segmented binary imageI 4；
Step 7, removing by using a morphological noise reduction methodI 4Medium area is less thanA8-connected foreground regions of individual pixels, thereby eliminatingI 4The separated foreground of the point distribution finally obtains the segmented shoe print image
。
Compared with the prior art, the invention has the technical characteristics that: firstly, the illumination influence caused by overexposure or underexposure and the false prospect caused by imprint residue caused by equipment aging can be eliminated, and the prospect target, namely the shoe mark, is well reserved; secondly, the shoe print images can be automatically processed in batch without manual interaction, the time efficiency is obviously superior to that of a manual shoe print image segmentation method and an iterative segmentation method (such as an image segmentation method based on a partial differential equation), and the segmentation accuracy is higher than that of a traditional global threshold segmentation method and an iterative segmentation method.
Detailed Description
An automatic threshold segmentation method for low-quality shoe print images is characterized by comprising the following steps of;
step 1, inputting an imageIIs converted into HSI and its luminance component is preservedH；
Step 2. use radius ofRThe flat disk-shaped structural element of (2) performs morphological etching treatment on the H to obtain an imageI 1Wherein, getR=28；
Step 3, using the same structural elements as step 2 to imageI 1Performing morphological dilation to obtain an estimated value of the image backgroundB；
Step 4, inputting the imageISubtracting the background estimateBObtaining a shoe print image without the influence of illuminationI 2；
Step 5, statisticsI 2The minimum value of the gray value range of (2) is set toI minMaximum value ofI maxAccording to the formula (1) toI 2Performing linear gray scale stretching to stretch the pixel gray scale range to 0-255:
wherein the content of the first and second substances,xrepresenting the input pixel value to be stretched,yrepresenting the stretched output pixel value. Let the enhanced image beI 3；
Step 6, adopt the traditional Otsu method to rightI 3Performing global threshold segmentation to obtain a segmented binary imageI 4；
Step 7, removing by using a morphological noise reduction methodI 4Medium area is less thanA8-connected foreground regions of individual pixels, thereby eliminatingI 4The separated foreground of the point distribution finally obtains the segmented shoe print image
Wherein, getA=100。
The 300 shoe print images are used for testing, and Jaccard Similarity (JS for short) is used for evaluation, and the closer the evaluation score is to 1, the closer the segmentation result is to the manual segmentation result. The average scores and standard deviations of the traditional Dajin segmentation method, the segmentation method based on the traditional active contour model and the algorithm are shown in Table 1 (the hardware environment of the embodiment is that a CPU is an Intel (R) core (TM) i5-3470 @3.20GHz dual core, the memory is 4GB, and the software environment is a Window 7 operating system and a Matlab 2016 scientific computing platform).
TABLE 1 Objective segmentation quality and segmentation time comparison
Claims (1)
1. An automatic threshold segmentation method for low-quality shoe print images is characterized by comprising the following steps:
step 1, inputting an imageIIs converted into HSI and its luminance component is preservedH；
Step 2. use radius ofRThe flat disk-shaped structural element of (2) performs morphological etching treatment on the H to obtain an imageI 1；
Step 3, using the same structural elements as step 2 to imageI 1Performing morphological dilation to obtain an estimated value of the image backgroundB；
Step 4, will loseInput imageISubtracting the background estimateBObtaining a shoe print image without the influence of illuminationI 2；
Step 5, statisticsI 2The minimum value of the gray value range of (2) is set toI minMaximum value ofI maxAccording to the formula (1) toI 2Performing linear gray scale stretching to stretch the pixel gray scale range to 0-255:
wherein the content of the first and second substances,xrepresenting the input pixel value to be stretched,yrepresenting the stretched output pixel values, let the enhanced image beI 3；
Step 6, adopt the traditional Otsu method to rightI 3Performing global threshold segmentation to obtain a segmented binary imageI 4；
Step 7, removing by using a morphological noise reduction methodI 4Medium area is less thanA8-connected foreground regions of individual pixels, thereby eliminatingI 4The separated foreground of the point distribution finally obtains the segmented shoe print image
。
Applications Claiming Priority (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2017114215322 | 2017-12-25 | ||
| CN201711421532 | 2017-12-25 |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN108364299A CN108364299A (en) | 2018-08-03 |
| CN108364299B true CN108364299B (en) | 2021-07-06 |
Family
ID=63006215
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810039047.7A Active CN108364299B (en) | 2017-12-25 | 2018-01-16 | Automatic threshold segmentation method for low-quality shoe print image |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108364299B (en) |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104156914A (en) * | 2014-07-18 | 2014-11-19 | 北京海鑫科金高科技股份有限公司 | Method for automatically processing footprint image |
| CN104794476A (en) * | 2015-04-21 | 2015-07-22 | 杭州创恒电子技术开发有限公司 | Personnel trace extraction method |
| CN105374045A (en) * | 2015-12-07 | 2016-03-02 | 湖南科技大学 | Morphology-based image specific shape dimension objet rapid segmentation method |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2015143948A1 (en) * | 2014-03-27 | 2015-10-01 | 大连恒锐科技股份有限公司 | Extraction method and extraction device for crime scene footprint through photographing |
-
2018
- 2018-01-16 CN CN201810039047.7A patent/CN108364299B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104156914A (en) * | 2014-07-18 | 2014-11-19 | 北京海鑫科金高科技股份有限公司 | Method for automatically processing footprint image |
| CN104794476A (en) * | 2015-04-21 | 2015-07-22 | 杭州创恒电子技术开发有限公司 | Personnel trace extraction method |
| CN105374045A (en) * | 2015-12-07 | 2016-03-02 | 湖南科技大学 | Morphology-based image specific shape dimension objet rapid segmentation method |
Non-Patent Citations (2)
| Title |
|---|
| Segmentation and Enhancement of Low Quality Fingerprint Images;Fleyeh, Hasan;《WEB INFORMATION SYSTEMS ENGINEERING - WISE 2016》;20161102;全文 * |
| 低质量指纹图像分割算法研究;周海徽等;《计算机技术与发展》;20110831;第21卷(第8期);全文 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN108364299A (en) | 2018-08-03 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109872285B (en) | Retinex low-illumination color image enhancement method based on variational constraint | |
| US9251614B1 (en) | Background removal for document images | |
| CN109978848B (en) | Method for detecting hard exudation in fundus image based on multi-light-source color constancy model | |
| CN106709964B (en) | Sketch generation method and device based on gradient correction and multidirectional texture extraction | |
| Niu et al. | Research and analysis of threshold segmentation algorithms in image processing | |
| CN107038416B (en) | Pedestrian detection method based on binary image improved HOG characteristics | |
| CN108898132B (en) | Terahertz image dangerous article identification method based on shape context description | |
| JP2013200872A (en) | System and method for processing image for identifying alphanumeric characters present in series | |
| CN107610144B (en) | Improved infrared image segmentation method based on maximum inter-class variance method | |
| CN109544583B (en) | Method, device and equipment for extracting interested area of leather image | |
| CN111598886B (en) | Pixel-level transmittance estimation method based on single image | |
| CN107256539B (en) | Image sharpening method based on local contrast | |
| CN115272362A (en) | Method and device for segmenting effective area of digital pathology full-field image | |
| CN103177244B (en) | Method for quickly detecting target organisms in underwater microscopic images | |
| CN110807747B (en) | Document image noise reduction method based on foreground mask | |
| CN110766614B (en) | Image preprocessing method and system of wireless scanning pen | |
| Valliammal et al. | A hybrid method for enhancement of plant leaf recognition | |
| CN108364299B (en) | Automatic threshold segmentation method for low-quality shoe print image | |
| CN117496109A (en) | Image comparison and analysis method and device, electronic equipment and storage medium | |
| CN109191481B (en) | Rapid threshold segmentation method for shoe print image under poor exposure condition | |
| CN112308793A (en) | Novel method for enhancing contrast and detail of non-uniform illumination image | |
| CN109241865B (en) | Vehicle detection segmentation algorithm under weak contrast traffic scene | |
| Blajovici et al. | Shadow detection and removal from a single image | |
| Somasekar et al. | Contrast-enhanced microscopic imaging of malaria parasites | |
| CN108961290B (en) | Adaptive image edge detection method based on Ratio operator of Otsu |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| 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 |