CN-121998841-A - Method for highlighting texture features based on multi-focus image

Abstract

The invention provides a method for highlighting texture features based on multi-focusing images, which comprises the steps of 1, collecting multi-layer focusing sequence images of fiber fragments, 2, accumulating pixels at corresponding positions of the collected multi-layer focusing sequence images, averaging to obtain an average image, 3, calculating the definition of each layer of images, 4, calculating the sum of the definition of each multi-layer image interval, finding the optimal multi-layer image interval with the maximum value, 5, subtracting the pixels at the corresponding positions of the average image from each layer of images in the multi-layer image interval and taking absolute values to obtain a difference image sequence of each image to the average image, and 6, summing and normalizing all the difference images to obtain a final texture enhancement image. The method has the technical effects of good texture highlighting effect, good certainty, high efficiency and good robustness on the fiber image in the microscopic imaging state, and can be used for the image preprocessing stage of a qualitative and quantitative analysis system for the textile fiber type components.

Inventors

• PENG XIANGZHOU
• ZHANG BAOGUO
• LIU DAN
• WANG LEI
• XIAO XIONG
• LV XIN
• LI WENBANG
• DING MAOLUAN
• HUANG ZHIHONG
• YAN CHUNHONG
• LIANG YUYAN
• PENG JUN

Assignees

• 北京和众视野科技有限公司

Dates

Publication Date

20260508

Application Date

20260122

Claims (4)

1. 1. A method for highlighting texture features based on a multi-focus image, comprising the steps of: step 1, acquiring a multi-layer focusing sequence image of a fiber segment; step 2, carrying out pixel accumulation on the corresponding positions of the acquired multi-layer focusing sequence images, and then averaging to obtain an average image; Step3, calculating the definition of each layer of image; step 4, calculating the sum of the definition of each multi-layer graph interval, and finding the optimal multi-layer graph interval with the largest value; Step 5, each layer of images in the multi-layer image interval is respectively subtracted by pixels at positions corresponding to the average images and absolute values are taken, so that a difference image sequence of each image to the average images is obtained; and 6, adding and normalizing all the difference images to obtain a final texture enhancement image.
2. 2. The method for highlighting texture features based on a multi-focused image as recited in claim 1, wherein said step 1 comprises: And controlling a Z-axis focusing motor by using an image system with an electric control automatic object stage, and collecting multi-layer focusing sequence images { M 1 ,M 2 ,… M n } within a focusing stroke interval of 200-400 um, wherein the distance between the collecting layers is kept to be 5-10um, and the collecting states of a plurality of sequential sections of blurring-clear-blurring are included.
3. 3. The method of claim 1, wherein the sharpness is obtained by summing pixel gradients in step 3.
4. 4. The method of claim 1, wherein subtracting the average map corresponding position pixels from each layer of images in the multi-layer map interval and taking absolute values in step 5 comprises setting a sensitivity threshold as a control parameter, and setting the pixel value smaller than the threshold to zero, thereby controlling whether the weak difference is removed as a disturbance.

Description

Method for highlighting texture features based on multi-focus image Technical Field The invention belongs to the technical field of textile fiber microscopic image texture feature enhancement, and particularly relates to a method for highlighting texture features based on a multi-focus image. Background Fiber content testing is a routine item in textile testing. Wherein textiles made of natural fibers (animal fibers, plant fibers) occupy a large specific gravity. The detection of such products is usually performed by manual microscopic examination or semi-automatic image analysis, and mainly depends on the classification and identification capability of detection engineers on microscopic amplified fiber fragment images, and has quite high subjectivity. In recent years, with the continuous development of digital image processing technology and artificial intelligence technology, the detection industry increasingly expects automated objective detection, and machine vision-based detection is one of the most potential ways. The detection based on vision (image) generally comprises links such as image acquisition, pretreatment, extraction, identification, measurement, result statistical analysis and the like, and each step before the result has factors which can influence the accuracy and representativeness of the result. In particular, a preprocessing method with good texture saliency effect, good certainty, high efficiency and good robustness is needed. Disclosure of Invention The invention aims to provide a method for highlighting texture features based on multi-focus images, which is used for providing images with salient features for subsequent identification work based on fiber surface texture features in multi-layer focus image salient images and can be used for an image preprocessing stage of a fiber component qualitative and quantitative analysis system. The invention provides a method for highlighting texture features based on multi-focus images, which comprises the following steps: step 1, acquiring a multi-layer focusing sequence image of a fiber segment; step 2, carrying out pixel accumulation on the corresponding positions of the acquired multi-layer focusing sequence images, and then averaging to obtain an average image; Step3, calculating the definition of each layer of image; step 4, calculating the sum of the definition of each multi-layer graph interval, and finding the optimal multi-layer graph interval with the largest value; Step 5, each layer of images in the multi-layer image interval is respectively subtracted by pixels at positions corresponding to the average images and absolute values are taken, so that a difference image sequence of each image to the average images is obtained; and 6, adding and normalizing all the difference images to obtain a final texture enhancement image. Further, the step 1 includes: And controlling a Z-axis focusing motor by using an image system with an electric control automatic object stage, and collecting multi-layer focusing sequence images { M 1,M2,… Mn } within a focusing stroke interval of 200-400 um, wherein the distance between the collecting layers is kept to be 5-10um, and the collecting states of a plurality of sequential sections of blurring-clear-blurring are included. Further, the sharpness in step 3 is obtained by calculating the sum of pixel gradients. Further, the step 5 of subtracting the pixels at the positions corresponding to the average map from the images at each layer in the multi-layer map interval and taking the absolute value includes setting a sensitive threshold as a control parameter, and setting the pixel value smaller than the threshold to zero, thereby controlling whether the weak difference is removed as interference. Compared with the prior art, the invention has the beneficial effects that: 1) The effect of the prominent texture is good, similar to the imaging effect of an electron microscope. For shallow and sparsely populated texture features enhanced by accumulation of multi-layer graphs, non-feature graded content is suppressed and non-location-preserving interfering content is suppressed. Providing a strong feature map basis for subsequent extraction or recognition of texture features based on these highlights. 2) The algorithm has good certainty, and the texture change caused by different focusing conditions like a conventional single-layer graph can not influence the subsequent recognition. 3) The main process of the algorithm is matrix addition operation, so that the algorithm is very efficient. 4) The algorithm has good robustness, the illumination adaptation range of the acquired image is larger, and the illumination uniformity in the monoscopic field is not high. Drawings FIG. 1 is a flow chart of a method of highlighting texture features based on a multi-focused image in accordance with the present invention; FIG. 2 is a texture enhancement diagram according to an embodiment of the present invention; FIG. 3 is