WO2021248915A1 - 一种清水混凝土色差分析/检测方法及*** - Google Patents

一种清水混凝土色差分析/检测方法及*** Download PDF

Info

Publication number
WO2021248915A1
WO2021248915A1 PCT/CN2021/073124 CN2021073124W WO2021248915A1 WO 2021248915 A1 WO2021248915 A1 WO 2021248915A1 CN 2021073124 W CN2021073124 W CN 2021073124W WO 2021248915 A1 WO2021248915 A1 WO 2021248915A1
Authority
WO
WIPO (PCT)
Prior art keywords
color
image
fair
area
color difference
Prior art date
Application number
PCT/CN2021/073124
Other languages
English (en)
French (fr)
Inventor
范宏
郭思瑶
张鹏
金祖权
万小梅
Original Assignee
青岛理工大学
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 青岛理工大学 filed Critical 青岛理工大学
Publication of WO2021248915A1 publication Critical patent/WO2021248915A1/zh

Links

Images

Classifications

    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06TIMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00Image analysis
    • G06T7/90Determination of colour characteristics
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06TIMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00Image analysis
    • G06T7/0002Inspection of images, e.g. flaw detection
    • G06T7/0004Industrial image inspection
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06TIMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00Image analysis
    • G06T7/10Segmentation; Edge detection
    • G06T7/11Region-based segmentation
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06TIMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00Image analysis
    • G06T7/60Analysis of geometric attributes
    • G06T7/62Analysis of geometric attributes of area, perimeter, diameter or volume
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06TIMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00Indexing scheme for image analysis or image enhancement
    • G06T2207/30Subject of image; Context of image processing
    • G06T2207/30108Industrial image inspection
    • G06T2207/30132Masonry; Concrete

Definitions

  • the invention belongs to the field of civil engineering fair-faced concrete appearance quality evaluation, and specifically relates to a color difference evaluation method for color images.
  • Fair-faced concrete is the concrete that is formed at one time without any decoration. It uses the natural texture of the concrete itself and the natural state formed by the combination of carefully designed open seams, zen seams and tension bolt holes as the architectural expression of the decorative surface. It is widely used In industrial buildings, civil engineering high-rise buildings, public buildings and municipal bridges.
  • Fair-faced concrete construction originated from abroad and has been widely used in Japan, Europe and the United States and other countries, becoming a new architectural genre. my country's fair-faced concrete technology began to be tried in the late 1980s, and its development process can be summarized in four stages: original fair-faced concrete, fair-faced concrete, mirror-finished fair-faced concrete and colored fair-faced concrete.
  • fair-faced concrete is uniform, smooth, and beautiful in color, without any modification, with accurate cross-sectional dimensions, rounded edges and corners, smooth lines, and natural transition between layers.
  • the visual quality of fair-faced concrete has reached a higher artistic level, without any decoration, and the plaster layer and surface layer are eliminated.
  • Significant environmental benefits fair-faced concrete technology eliminates plastering and wet work, improves the level of civilized construction on site, reduces winter construction, and also reduces construction waste.
  • the fair-faced concrete technology eliminates the plastering layer, the common quality problems of plastering engineering that are easy to hollow, fall off and cracks are eliminated.
  • Fair-faced concrete requires that the surface of the concrete is smooth and smooth, with uniform color and no damage or pollution.
  • the setting of the tension bolts and construction joints should be neat and beautiful, and the common quality defects of ordinary concrete are not allowed. But in fact, due to the complex construction process of fair-faced concrete, the level of the construction team is different, and there are no strict quality acceptance specifications and technical standards to follow in China. It is very easy to cause defects in the appearance of fair-faced concrete, among which color defects are the most common. Therefore, the evaluation of chromatic aberration is of great significance to the construction of fair-faced concrete. There are two commonly used methods for evaluating the color difference of fair-faced concrete:
  • the manual evaluation method is the method adopted in my country's standard fair-faced concrete application technical regulations (JGJ 169-2009), and is currently the most widely used method.
  • the specific method is: randomly select three to five trained color difference inspectors, stand at a designated distance, and score and evaluate the selected concrete surface area. Finally, the average score of all color difference inspectors is taken to evaluate the color difference defects.
  • Artificial evaluation methods are more susceptible to the influence of people's subjective consciousness, resulting in evaluation results that are not objective and accurate.
  • Gray-scale image evaluation method Z. Zhu (Zhu Z, Brilakis I. Detecting air pockets for architectural quality assessment using visual sensing[J]. Electronic Journal of Information Technology in Construction, 2008, 13: 86-102.) proposed a gray-scale image based In the color difference evaluation method, the specific steps are to obtain a concrete image, perform gray-scale conversion on the image, and calculate the standard deviation of the gray-scale image. Then the chromatic aberration defects are simply classified as qualified and unqualified. Peng Haitao (Peng Haitao, Su Jie, Fang Zhi, et al. Detection and evaluation of concrete surface chromatic aberration based on image analysis technology[J]. Highway Engineering, 2012(5), 19-22) On the basis of Z. Zhu, it is also calculated image The standard deviation of gray scale, but introduces the concepts of human eye brightness contrast threshold, viewing angle size, etc., and fuzzy evaluation of color difference defects based on the graded membership function.
  • the above gray-scale image evaluation methods are all based on the gray-scale image method, and the color difference is evaluated based on the standard deviation. Converting an RGB image into a grayscale image according to a certain algorithm, but the conversion into a grayscale image has a major disadvantage. After the conversion, the color information of the image will be lost to a greater extent, resulting in inaccurate evaluation.
  • RGB color mode is a color standard in the industry. It obtains a variety of colors by changing the three color channels of red (R), green (G), and blue (B) and superimposing them with each other. Yes, RGB is the color representing the three channels of red, green, and blue. This standard includes almost all the colors that human vision can perceive, and it is currently one of the most widely used color systems.
  • Each of the three color channels of red, green and blue is divided into 255 levels of brightness. At 0, the "light” is the weakest and turned off, and at 255, the "light” is the brightest. When the three color values are the same, it is a colorless grayscale color, and when the three colors are all 255, it is the brightest white, and when the three colors are all 0, it is black.
  • the three color channel values corresponding to each typical color are shown in Table 1.
  • RGB images can be converted into grayscale images. There are many ways to convert RGB values and grayscales. In fact, it is the conversion of the human eye's perception of color to brightness. This is a psychological problem. The commonly used conversion has the following formula:
  • the conversion of gray-scale images is based on the visual sense of the human eye, for different color components, taking different weights, the R component is 0.229, the G component is 0.587, and the B component is 0.114. Then add up to get the gray value.
  • the three color components are R:100G:0B:0 (brown), R:0G:40B:0 (blue), and R:0G:0B:200 (dark green).
  • their gray scale values are all approximately equal to 23. It can be seen that the grayscale conversion of different color images loses image information and cannot reflect the human eye's perception of color differences.
  • the present invention proposes a new method of color difference evaluation method based on RGB color image, which converts the RGB color image into RGB three color channel values, calculates the standard deviation for each color channel, and then The standard deviation of the three color channels is analyzed to realize the accurate evaluation of the color difference analysis of fair-faced concrete.
  • the relative area the ratio of the color difference area to the entire image
  • the degree of difference the size of the color difference between the color difference defect area and the non-color difference defect area
  • Image segmentation is used to determine the relative area
  • the color difference formula is used to determine the degree of difference.
  • the color image-based chromatic aberration defect evaluation method avoids the subjectivity of artificial chromatic aberration defect evaluation and avoids the loss of image information based on gray-scale image chromatic aberration defect evaluation. The specific process is shown in Figure 1.
  • the first aspect of the present invention provides a method for analyzing the color difference of fair-faced concrete, including:
  • the present invention converts the RGB color image into RGB three color channel values, calculates the standard deviation of each color channel, and then analyzes the standard deviation of the three color channels to realize the accurate evaluation of the color difference analysis of fair-faced concrete.
  • the second aspect of the present invention provides a fair-faced concrete color difference analysis system, including:
  • a module for evaluating the color consistency of concrete appearance by using the gray standard deviation value is a module for evaluating the color consistency of concrete appearance by using the gray standard deviation value.
  • the third aspect of the present invention provides a method for detecting the color difference of fair-faced concrete, including:
  • Image segmentation is performed by combining region merging and quadtree segmentation, and the chromatic aberration area and the non-chromatic aberration area are segmented;
  • the chromatic aberration area is rated with reference to the quality evaluation table.
  • the present invention uses image segmentation to determine the relative area, and uses a color difference formula to determine the degree of difference.
  • the color image-based chromatic aberration defect evaluation method avoids the subjectivity of artificial chromatic aberration defect evaluation, and avoids the loss of image information based on gray-scale image chromatic aberration defect evaluation, and the evaluation result is more accurate.
  • the fourth aspect of the present invention provides a system for detecting the color difference of fair-faced concrete, which includes:
  • a module for grading the color difference area A module for grading the color difference area.
  • the present invention converts the RGB color image into three RGB color channel values, calculates the standard deviation of each color channel, and then analyzes the standard deviation of the three color channels to realize the evaluation of the color difference analysis of fair-faced concrete , The evaluation result is clearer and more accurate.
  • the present invention uses image segmentation to determine the relative area, and uses a color difference formula to determine the degree of difference.
  • the color image-based chromatic aberration defect evaluation method avoids the subjectivity of artificial chromatic aberration defect evaluation, and avoids the loss of image information based on gray-scale image chromatic aberration defect evaluation.
  • FIG. 1 is a flowchart of the present invention
  • Figure 2 is a standard image taken in Embodiment 1;
  • Fig. 3 is a diagram showing the decomposition of the image into three color channels in embodiment 1;
  • Embodiment 4 is a distribution diagram of pixel values of three color channels in Embodiment 1;
  • Figure 5 is a standard image taken in Embodiment 2.
  • FIG. 6 is a diagram of the image decomposed into three color channels in Embodiment 2;
  • FIG. 7 is a distribution diagram of pixel values of three color channels in Embodiment 2.
  • FIG. 8 is a diagram of the quadtree segmentation of Embodiment 3.
  • Figure 9 is a color difference defect diagram of Example 3.
  • Fig. 10 is a Lab color space diagram of embodiment 3.
  • FIG. 11 is a color difference defect segmentation diagram of Example 3.
  • the chromatic aberration area can be regarded as the target object with the normal concrete surface as the background. Due to the intricacies of environmental conditions, shooting conditions (such as environmental brightness, camera resolution, and shooting distance, etc.) have an important influence on the results of image analysis. In order to avoid the influence of environmental brightness, camera resolution and shooting distance on the standard deviation of image gray levels, to obtain standard images, the collected images should meet the following conditions:
  • the photographing distance should be 4 to 6 meters.
  • the length of the photograph corresponding to the size of the subject is 0.9 to 1.1 meters and the width is 0.5 to 0.7 meters.
  • Achieve white balance that is, "the white object can be restored to white regardless of the light source.”
  • AFB automatic white balance
  • most digital SLR cameras support custom white balance, and in the custom white balance operation, it is a better practice to use a standard gray card.
  • the method to achieve white balance is as follows:
  • Standard gray card Use the white surface, enter the "custom white balance mode" to take a picture of the white surface of the standard gray card, and DC will know what is white under the light conditions.
  • RGB images are also called full-color images. There are three channels: R (red), G (green), and B (blue). Use image software to separate the standard RGB image into three channels, and obtain the three-channel gray value. After separation, three matrices representing R (red), G (green), and B (blue) are obtained, and the values of the matrices are distributed in the range of 0 to 255.
  • M and N respectively represent the number of rows and columns of the color channel image
  • Gray (i, j) represents the gray value of each pixel
  • Gray represents the average gray value of the entire image.
  • the gray standard deviation value is used as the evaluation standard for the consistency of concrete appearance color, as shown in Table 2. Take the maximum value of the standard deviations of the three color channels, and evaluate the gray-scale standard deviation less than 6 as the first-class appearance color consistency, the gray-scale standard deviation less than 10 as qualified, and the gray-scale standard deviation greater than 10. In order to fail to meet the requirements for the appearance and color consistency of fair-faced concrete, it is unqualified.
  • the present invention proposes a method for evaluating the color difference of fair-faced concrete, including:
  • the color space of the RGB image is converted into a color space Lab space which is closer to the color recognition mechanism of the human eye.
  • the RGB color space is the most commonly used color space. Electronic devices such as digital cameras and scanners all use the RGB color space to represent colors. Therefore, the RGB color space is also referred to as the device-related color space.
  • Lab was established on the basis of the international standard for color measurement established by the International Commission on Illumination (CIE) in 1931. It is a device-independent color system as well as a color system based on physiological characteristics. This also means that it uses digital methods to describe human visual perception.
  • the Lab color gamut is wide, not only includes all the color gamuts of RGB, but also expresses the colors they cannot express. The colors that the human eye can perceive can be expressed through the Lab model, which makes up for the lack of uneven RGB color distribution, and In the Lab color space, the space coordinate distance between two points can be used to express the color difference ⁇ .
  • XYZ color space is required as a transition.
  • the XYZ color space is also a color space introduced by the International Commission on Illumination (CIE) on the basis of RGB.
  • CIE International Commission on Illumination
  • a new color space is established with three imaginary primary colors X, Y, and Z.
  • X n , Y n , and Z n are the tristimulus values of the CIE standard illuminator on the complete diffuse reflector, and then through the complete diffuse reflector to the observer's eye.
  • the values are usually 95, 100, 108, respectively.
  • Image segmentation can better identify the chromatic aberration defects on the surface of the concrete, and lay the foundation for the subsequent evaluation of chromatic aberration defects.
  • the segmentation based on the color image ensures the completeness of the information of the concrete surface image.
  • the histogram threshold method is a widely used segmentation method for grayscale images, but it will appear when it is applied to a color image, and the region obtained by segmentation may be incomplete; the histogram of a color image is a three-dimensional array and There may not necessarily be obvious valleys, which are used for thresholding segmentation; there are no problems such as using local spatial information.
  • Edge detection is also a widely used technique for grayscale image segmentation.
  • the present invention systematically analyzes and researches the existing graphics segmentation method. Aiming at the characteristics of fair-faced concrete surface smoothness, high color saturation, and good overall uniformity, it proposes the use of region merging and quadtree segmentation.
  • the combined image segmentation method effectively identifies the chromatic aberration defect area on the surface of fair-faced concrete, ensures the information integrity of the concrete surface image, and has high accuracy in the evaluation of chromatic aberration defects.
  • Region merging is to classify images according to certain characteristics, different categories are classified into different sets, and the same categories are classified into the same set. Combine the pixels of the same set to make them into a whole.
  • Quadtree segmentation is to divide the image into four rectangular regions of the same size, and set the threshold.
  • the rectangular area that meets the threshold requirement is no longer segmented; the rectangular area that does not meet the threshold requirement, the quadtree segmentation continues. Repeat this way, until the segmented area is a single pixel or when the segmented area meets the threshold requirement, the quadtree segmentation is stopped, as shown in FIG. 8.
  • the specific steps are: 1. Perform sub-block segmentation on the image. 2. Perform area merging on the divided pure color sub-blocks. 3. Perform quadtree division on the non-pure color sub-blocks, and then merge the regions. 4. Perform region merging on the regions divided by the pure color sub-block and the non-pure color sub-block.
  • the image is divided into sub-blocks, and the image is divided into m ⁇ n sub-blocks. It should be noted that the values of m and n can be adjusted according to the actual size of the image.
  • x ij represents a vector composed of the L, a, and b values of the pixel in the i-th row and the j-th column in a sub-block of pixels in the c row and the d column.
  • the variance vector is:
  • the pure-color sub-block has a small variance because the inner color is uniform, and the non-pure-color sub-block has a larger variance because it contains a color difference area. Thresholds can be set according to the situation to distinguish between pure-color sub-blocks and non-pure-color sub-blocks.
  • the pure color sub-block is divided into two parts, one part is the color difference area, and the other part is the non-color difference area.
  • the non-pure color sub-blocks are divided by a quadtree method.
  • the non-pure color sub-block is divided into four small sub-blocks of the same size and shape.
  • step (3) Repeat the above step (2) until all sub-blocks are pure color sub-blocks or one pixel. End the quadtree split.
  • the first uniform block or uniform pixel in the quadtree segmentation process is taken as the set P 1 .
  • the non-pure color sub-block is also divided into two parts, the color difference area and the non-color difference area.
  • Color difference is the description of people's perception of different colors.
  • the space coordinate distance between two points is used to express the color difference ⁇ , as shown in formulas (9) to (12).
  • NBS color difference unit 1
  • the human eye recognition result is basically no color difference; 3-6
  • NBS there is a big difference; when there are more than 12 NBS, it will be recognized as different colors by the human eye. Details are shown in Table 3.
  • the degree of difference is expressed as the degree of color difference between the chromatic aberration defect area and the entire concrete shooting surface. The greater the color difference, the serious color difference on the surface of the concrete.
  • the color distance ⁇ is used to indicate the degree of difference.
  • B Indicates the size of the chromatic aberration defect area relative to the entire image area.
  • B is used to represent the relative area
  • S O represents the area of the chromatic aberration defect area
  • S represents the area of the entire concrete shooting surface. Note that S O and S here are in pixels. Then B is:
  • the concrete color difference area is graded and evaluated.
  • the evaluation form is established here in accordance with the "Code for Acceptance of Construction Quality of Concrete Structure Engineering" (GB50204-2015).
  • the table is divided into five levels, I, II, III, IV, and V. The larger the level, the more serious the chromatic aberration defect, as shown in Table 4.
  • the specific usage method is to evaluate the color difference defect based on the relative area and the degree of difference. When the two indicators belong to different levels, the higher level is used as the evaluation result.
  • Figure 3 Decomposes the image into three color channels
  • Formula (2) is used to calculate the standard deviation S of the three color channels, and the calculation results are shown in Table 5.
  • the gray standard deviation value is used as the evaluation standard of the color consistency of the concrete appearance.
  • the standard deviations of the three color channels are the red channel: 41.58, the blue channel: 66.07, the green channel: 50.96, and the maximum of the three is 66.07 , Compare the maximum value of 66.07 with the color difference judgment standard in Table 2.
  • the gray scale standard deviation is greater than 10, and it is judged as unqualified.
  • the fair-faced concrete image in a subway station is used for analysis.
  • the exposed concrete room is illuminated by light with sufficient light.
  • the photographing distance should be 5 meters, and the photographed fair-faced concrete should be 1 meter long and 0.65 meters wide.
  • white balance is realized, and standard images of fair-faced concrete are taken. As shown in Figure 5.
  • Formula (2) is used to calculate the standard deviation S of the three color channels, and the calculation results are shown in Table 6.
  • the gray standard deviation value is used as the evaluation standard of the color consistency of the concrete appearance.
  • the standard deviations of the three color channels are the red channel: 11.55, the blue channel: 15.51, the green channel: 14.31, and the maximum of the three is 15.51 , Compare the maximum value of 15.51 with the color difference judgment standard in Table 2.
  • the gray scale standard deviation is greater than 10, and it is judged as unqualified.
  • the bridge piers are poured with concrete, due to inadequate maintenance. Some areas on the piers have begun to produce chromatic aberration defects. As shown in Figure 9.
  • Figure 9(a) is rust. Because the upper metal members of the pillars are rusted, the lower pillars have chromatic aberration defects;
  • Figure 9(b) is the stain, which is caused by insufficient protection of the concrete surface.
  • Figure 9(c) shows the oil stains, due to improper use of the release agent when the concrete is being formed, resulting in chromatic aberration defects.
  • the method mentioned in the present invention is used to evaluate the chromatic aberration of the image.
  • the steps are:
  • Step 1 Take a picture of the area that needs to be evaluated for chromatic aberration, and it is required to ensure that the lens is parallel to the shooting surface as much as possible.
  • Step 2 Use formulas (3) ⁇ (6) to convert the RGB color space of the image to Lab color space as shown in Figure 10.
  • Step 3 Perform image segmentation on the color difference defect area of the image, and identify the color difference defect area as shown in Figure 11.
  • Step 4 Find the relative area of one of the evaluation indicators.
  • Step 5 Calculate the degree of difference between the color difference defect area and the non-color difference defect area.
  • the specific method is to use formulas (7) to (8) to find the color average vector of the color difference area and the non-color difference area. Then use equations (9) to (12) to find the color distance ⁇ , and use ⁇ to judge the degree of difference.
  • Step 6 Refer to Table 3 to evaluate the chromatic aberration defect area. The results are shown in Table 7.

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • General Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Quality & Reliability (AREA)
  • Geometry (AREA)
  • Image Analysis (AREA)
  • Image Processing (AREA)

Abstract

提供一种清水混凝土色差分析/检测方法及***。该分析方法包括以下步骤:采集待测混凝土表面的标准图像;将所述标准图像转化为RGB图像,并进行三个颜色通道分离,得到三个颜色通道的灰度值;分别计算三个颜色通道的标准差,得到三个颜色通道的标准差;取三个颜色通道的标准差的最大值作为灰度标准偏差,将灰度标准偏差小于6的评定为外观色泽一致性的一等,将灰度标准偏差小于10的评定为合格,将灰度标准偏差大于10的评定为不合格。该方法避免了人工色差缺陷评价的主观性以及基于灰度图像的色差缺陷评价的图像信息丢失。

Description

一种清水混凝土色差分析/检测方法及*** 技术领域
本发明属于土木工程清水混凝土外观质量评价领域,具体涉及一种于彩色图像的色差评价方法。
背景技术
公开该背景技术部分的信息仅仅旨在增加对本发明的总体背景的理解,而不必然被视为承认或以任何形式暗示该信息构成已经成为本领域一般技术人员所公知的现有技术。
清水混凝土即一次成型,不做任何装饰的混凝土,其以混凝土本身的自然质感与精心设计的明缝、禅缝和对拉螺栓孔组合形成的自然状态作为装饰面的建筑表现形式,广泛应用于工业建筑、民用工程中高层、公共建筑以及市政桥梁中。
清水混凝土建筑起源于国外,在日本以及欧美等国家得到了广泛的应用,成为了一种新的建筑流派。我国清水混凝土技术于20世纪80年代后期开始试用,其发展过程可概括为原始清水混凝土、清水混凝土、镜面清水混凝土和彩色清水混凝土四个阶段。
清水混凝土备受青睐和广泛应用的原因主要有以下几个方面:(1)清水混凝土不做任何修饰,颜色均匀、光滑、美观,截面尺寸准确,棱角倒圆,线条顺畅,层间过渡自然。(2)清水混凝土的观感质量达到较高的艺术境界,不做任何装饰,取消了抹灰层和面层。(3)环境效益显著:清水混凝土技术既取消抹灰又取消了湿作业,提高了现场文明施工程度,减少了冬季施工,同时也减少了建筑垃圾产生。(4)清水混凝土技术由于取消了抹灰层而消除了抹灰工程易空鼓、脱落和裂缝的质量通病。
清水混凝土要求混凝土表面平整光滑、色泽均匀,无碰损和污染,对拉螺栓及施工缝的设置应整齐美观,且不允许出现普通混凝土的质量通病。但实际上,由于清水混凝土的施工工艺复杂,而施工队伍水平的高低不一并且国内尚无严格的质量验收规范和技术标准可遵循。极易造成清水混凝土外观质量缺陷,这其中以色差缺陷最为常见。因此色差的评价对清水混凝土的施工有着重大的意义。目前常用的清水混凝土色差评价方法有以下两种:
1、人工评价法。人工评价法是我国标准清水混凝土应用技术规程(JGJ 169-2009)采用的方法,是目前采用最广泛的方法。具体做法为:随机抽取三到五名经过培训的色差检验员,站在指定距离处,对选定的混凝土表面区域进行打分评价。最后取所有色差检验员的平均分,来对色差缺陷进行评价。基于人工的评价方法较容易受人主观意识的影响,造成评价结果不 够客观、准确。
2、灰度图像评价方法。Z.Zhu(Zhu Z,Brilakis I.Detecting air pockets for architectural concrete quality assessment using visual sensing[J].Electronic Journal of Information Technology in Construction,2008,13:86-102.)提出了一种基于灰度图像的色差评价方法,具体步骤为获取混凝土图像,对图像进行灰度转换,计算灰度图像的标准差。然后将色差缺陷简单的区分为合格与不合格。彭海涛(彭海涛,苏捷,方志,等.基于图像分析技术的混凝土表面色差检测及评定[J].公路工程,2012(5),19-22)在Z.Zhu的基础上,同样是计算图像的灰度标准差,但引入了人眼亮度对比阈值、视角大小等概念,依据分级隶属函数,对色差缺陷进行模糊评价。
以上的灰度图像评价方法均基于灰度图像法,依据标准差进行色差评价。将RGB图像按照一定的算法转化为灰度图像,但转换成灰度图像有个较大的缺点,转换之后会在较大的程度上丢失图像的色彩信息,造成评价不准确。
RGB色彩模式是工业界的一种颜色标准,是通过对红(R)、绿(G)、蓝(B)三个颜色通道的变化以及它们相互之间的叠加来得到各式各样的颜色的,RGB即是代表红、绿、蓝三个通道的颜色,这个标准几乎包括了人类视力所能感知的所有颜色,是目前运用最广的颜色***之一。
红、绿、蓝三个颜色通道每种颜色各分为255阶亮度,在0时"灯"最弱,是关掉的,而在255时"灯"最亮。当三色数值相同时为无色彩的灰度色,而三色都为255时为最亮的白色,都为0时为黑色。各个典型颜色对应的三个颜色通道值如表1。
表1 RGB图像典型颜色的三通道颜色值
Figure PCTCN2021073124-appb-000001
Figure PCTCN2021073124-appb-000002
RGB图像可转化为灰度图像,RGB值和灰度的转换,有很多方法,实际上是人眼对于彩色的感觉到亮度感觉的转换,这是一个心理学问题,常用转换有下列一个公式:
Grey=0.299×R+0.587×G+0.114×B公式(1)
根据这个公式(1),依次读取每个像素点的R,G,B值,计算灰度值,将灰度值赋值给新图像的相应位置,所有像素点遍历一遍后完成转换。
灰度图像的转换就是根据人眼的视觉感官,针对不同的颜色分量,取不同的权重,R分量取0.229,G分量取0.587,B分量取0.114。之后再相加,得出灰度值。
因为RGB三个颜色分量权重固定,造成某些情况下图像的重要信息丢失。例如三个颜色分量分别为R:100G:0B:0(褐色),R:0G:40B:0(蓝色),R:0G:0B:200(墨绿色)。将三个图像进行灰度转换之后,其灰度值都约等于23。可见不同的彩色图像进行灰度转换,丢失了图像信息,不能反映人眼对色彩差异的感知。
发明内容
针对传统色差缺陷评价的缺点,本发明提出了一种新方法基于RGB彩色图像的色差评价方法,将RGB彩色图像转化为RGB三个颜色通道值,分别对每个颜色通道计算标准差,然后对三个颜色通道的标准差进行分析,实现对清水混凝土的色差分析的准确评价。同时,还采用相对面积(色差区域占整个图像的比率大小)和差异程度(色差缺陷区域和非色差缺陷区域的颜色差异程度的大小)两个指标来对色差缺陷进行评价。采用图像分割确定相对面积,采用色差公式确定差异程度。基于彩色图像的色差缺陷评价方法,避免了人工色差缺陷评价的主观性,又避免了基于灰度图像色差缺陷评价的图像信息丢失,具体流程如图1。
为实现上述技术目的,本发明采用如下技术方案:
本发明的第一个方面,提供了一种清水混凝土色差分析方法,包括:
采集待测混凝土表面的标准图像;
将所述标准图像转化为RGB图像,并进行三通道分离,得到三通道灰度值;
分别计算三个颜色通道的标准差,得到三个颜色通道的标准差S;
取三个颜色通道标准差的最大值,将灰度标准偏差小于6的评定为外观色泽一致性的一等,将灰度标准偏差小于10的评定为合格,将灰度标准偏差大于10的评定为不满足清水混凝土外观色泽一致性要求,即不合格。
本发明将RGB彩色图像转化为RGB三个颜色通道值,分别对每个颜色通道计算标准差, 然后对三个颜色通道的标准差进行分析,实现对清水混凝土的色差分析的准确评价。
本发明的第二个方面,提供了一种清水混凝土色差分析***,包括:
用于采集待测混凝土表面的标准图像的模块;
用于将标准图像转化为RGB图像的模块;
用于对RGB图像进行三通道分离的模块;
用于计算三个颜色通道的标准差的模块;
用灰度标准偏差值进行混凝土外观色泽一致性的评价的模块。
本发明的第三个方面,提供了一种清水混凝土色差的检测方法,包括:
采集待测混凝土表面的标准图像,并转化为RGB图像;
将RGB图像的颜色空间转换为Lab空间;
采用区域合并与四叉树分割相结合的方法进行图像分割,分割出了色差区域与非色差区域;
在lab颜色空间的基础上采用色差公式,对色差缺陷区域的差异程度进行量化;
采用相对面积和色差差异程度两个指标,参照质量评价表对色差区域进行评级。
本发明采用图像分割确定相对面积,采用色差公式确定差异程度。基于彩色图像的色差缺陷评价方法,避免了人工色差缺陷评价的主观性,又避免了基于灰度图像色差缺陷评价的图像信息丢失,评价结果更为准确。
本发明的第四个方面,提供了一种清水混凝土色差的检测***,包括:
用于采集待测混凝土表面的标准图像的模块;
用于将标准图像转化为RGB图像的模块;
用于将RGB图像的颜色空间转换为Lab空间的模块;
用于图像分割的模块;
用于对在lab颜色空间上色差缺陷区域的差异程度进行量化的模块;
用于对色差区域进行评级的模块。
本发明的有益效果在于:
(1)本发明则将RGB彩色图像转化为RGB三个颜色通道值,分别对每个颜色通道计算标准差,然后对三个颜色通道的标准差进行分析,实现对清水混凝土的色差分析的评价,评价结果更为清晰、准确。
(2)本发明采用图像分割确定相对面积,采用色差公式确定差异程度。基于彩色图像 的色差缺陷评价方法,避免了人工色差缺陷评价的主观性,又避免了基于灰度图像色差缺陷评价的图像信息丢失。
(3)本发明的分析评价方法简单、准确、实用性强,易于推广。
附图说明
构成本发明的一部分的说明书附图用来提供对本发明的进一步理解,本发明的示意性实施例及其说明用于解释本发明,并不构成对本发明的不当限定。
图1是本发明的流程图;
图2是实施例1中拍摄的标准图像;
图3是实施例1中将图像分解为三个颜色通道图;
图4是实施例1中三个颜色通道像素值分布图;
图5是实施例2中拍摄的标准图像;
图6是实施例2中图像分解为三个颜色通道图;
图7是实施例2中三个颜色通道像素值分布图;
图8是实施例3的四叉树分割图;
图9是实施例3的色差缺陷图;
图10是实施例3的Lab颜色空间图;
图11是实施例3的色差缺陷分割图。
具体实施方式
应该指出,以下详细说明都是例示性的,旨在对本发明提供进一步的说明。除非另有指明,本发明使用的所有技术和科学术语具有与本发明所属技术领域的普通技术人员通常理解的相同含义。
需要注意的是,这里所使用的术语仅是为了描述具体实施方式,而非意图限制根据本发明的示例性实施方式。如在这里所使用的,除非上下文另外明确指出,否则单数形式也意图包括复数形式,此外,还应当理解的是,当在本说明书中使用术语“包含”和/或“包括”时,其指明存在特征、步骤、操作、器件、组件和/或它们的组合。
一种新方法,基于RGB彩色图像的色差评价方法:
1.采集标准图像
对于物体表面色差,可以认为色差区域就是以正常混凝土表面为背景的目标物。由于环境条件错综复杂,因此拍摄条件(如环境亮度、相机分辨率和拍摄距离等)对图像分析结果 的影响至关重要。为避免环境亮度、相机分辨率和拍摄距离对图像灰度标准差的影响,获得标准图像,采集图像应满足下列条件:
(1)拍照时应光照充分,若自然光充足,被拍对象应光照均匀;若无自然光,或自然光不够充足,应有辅助光源。
(2)拍照距离应为4至6米,拍摄的照片对应被拍摄对象尺寸长为0.9至1.1米,宽为0.5至0.7米。
(3)实现白平衡,即在“不管在任何光源下,都能将白色物体还原为白色”。一般使用时选择相机的自动白平衡(AWB)就足够了,但在特定条件下如果色调不理想,可以选择使用其他的各种白平衡选项。目前,绝大部分数码单反相机都支持自定义白平衡,而在自定义白平衡操作中,使用标准灰卡是一种较好的做法。实现白平衡的方法如下:
1)自动白平衡。依赖数码相机里的测色温***调整。
2)标准灰卡。使用白面,进入“自定白平衡模式”对标准灰卡的白面,拍摄一张相片,DC便会得知在该光线条件下,什么才是白色。
3)拍摄保存RAW格式图像后期处理。
2.对RGB图像进行三个颜色通道分离
RGB图像也叫全彩图。其有三个通道,分别为:R(red),G(green),B(blue)。用图像软件对标准RGB图像进行三通道分离,并获取三通道灰度值。分离后,分别得到代表R(red),G(green),B(blue)的三个矩阵,矩阵的值分布在0到255的范围里。
3.分别计算三个颜色通道的标准差
采用下式计算三个颜色通道的标准差:
Figure PCTCN2021073124-appb-000003
其中M和N分别代表颜色通道图像的行数和列数,Gray(i,j)代表各个像素点的灰度值,Gray代表整幅图像的灰度均值。按上式计算会得到三个颜色通道的标准差S。
4.对清水混凝土色差进行评价
用灰度标准偏差值作为混凝土外观色泽一致性的评价标准,如表2所示。取三个颜色通道标准差的最大值,将灰度标准偏差小于6的评定为外观色泽一致性的一等,将灰度标准偏差小于10的评定为合格,将灰度标准偏差大于10的评定为不满足清水混凝土外观色泽一致性要求,即不合格。
表2 清水混凝土色差评价标准
Figure PCTCN2021073124-appb-000004
另一方面,针对目前基于灰度图像分析时依据标准差进行色差评价的方法存在转换成灰度图像有个较大的缺点,转换之后会在较大的程度上丢失图像的信息,造成评价不准确的问题,本发明提出了一种清水混凝土色差评价方法,包括:
一、将彩色图像的颜色空间转换为Lab空间。
将RGB图像的颜色空间转换为一种与人眼识别颜色机理更加接近的颜色空间Lab空间。RGB颜色空间是最常用的颜色空间,数码相机、扫描仪等电子设备,都采用RGB颜色空间来表示颜色,因此RGB颜色空间又被称为与设备有关的颜色空间。Lab是在1931年国际照明委员会(CIE)制定的颜色度量国际标准的基础上建立起来的。它是一种设备无关的颜色***,也是一种基于生理特征的颜色***。这也就意味着,它是用数字化的方法来描述人的视觉感应。Lab色域开阔,不仅包含了RGB的所有色域,还能表现它们不能表现的色彩,人的肉眼能感知的色彩,都能通过Lab模型表现出来,弥补了RGB色彩分布不均的不足,并且在Lab颜色空间中可以用两点之间的空间坐标距离来表示色差ΔΕ。
从RGB颜色空间转换到Lab颜色空间,需要XYZ颜色空间作为过渡。XYZ颜色空间也是由国际照明委员会(CIE)在RGB基础上引入的颜色空间,用三个假想的原色X、Y、Z建立了一个新的颜色空间。首先将RGB颜色空间转换为XYZ颜色空间。如式(3)所示:
Figure PCTCN2021073124-appb-000005
然后就是XYZ颜色空间转换为Lab颜色空间。如公式(4)~(6)。
Figure PCTCN2021073124-appb-000006
Figure PCTCN2021073124-appb-000007
Figure PCTCN2021073124-appb-000008
X n、Y n、Z n为CIE标准照明体在完全漫反射体上,再经完全漫反射体到观察者眼中的三刺激值。其值通常分别为95、100、108。
二、采用四叉树分割与区域合并相结合的算法思想,对图像进行分割。
图像分割可以较好的识别出混凝土表面色差缺陷部位,为后续的色差缺陷评价奠定基础。基于彩色图像的分割,保证了混凝土表面图像的信息完整。
发明人研究发现:虽然针对图像区域分割的具体要求已经建立了许多算法,然而至今尚无统一的理论,不能找到通用的方法能够适合于所有类型的图像。现有的大多数图像分割算法主要是针对灰度图像的,算法也相对较为成熟。彩色图像分割与灰度图像分割的算法相比,大部分算法在分割思想上是一致的。但彩色图像包含更丰富的信息,并有多种颜色空间的表达方式,分割算法因此有所不同。
特别是对于清水混凝土,其表面平滑度好,整体均匀性好,现有的分割方法往往难以保证混凝土表面图像的信息完整,从而导致后续色差缺陷评价存在一定偏差。例如:直方图阈值法是灰度图像广泛使用的一种分割方法,但当其应用于彩色图像时会出现,分割得到的区域可能是不完整的;彩色图像的直方图是一个三维数组中并不一定存在明显的谷,用来进行阈值化分割;没有利用局部空间信息等问题。而边缘检测也是灰度图像分割广泛使用的一种技术,它是基于在区域边缘上的像素灰度变化比较剧烈,通过检测不同区域的边缘来解决图像分割问题。当彩色图像中区域对比明显时,分割效果较好,反之,效果较差,因此,也不适用于清水混凝土的图像的分割。
为了解决上述问题,本发明对现有图形分割方法进行了***的分析和研究,针对清水混凝土表面平滑度好、颜色饱和度高、整体均匀性好的特点,提出采用区域合并与四叉树分割相结合的图像分割方法,有效地识别出了清水混凝土表面的色差缺陷区域,保证了混凝土表面图像的信息完整性,色差缺陷评价的准确性高。
区域合并即对图像按照某种特性进行分类,不同类别归入不同的集合,相同类别归入同一集合。对同一集合的像素进行合并,使得其成为一个整体。
四叉树分割即将图像分割成大小相同的四个矩形区域,设定阈值。满足阈值要求的矩形区域,则不再进行分割;不满足阈值要求的矩形区域,则继续进行四叉树分割。如此循环,直到分割的区域为单个像素点时或分割区域满足阈值要求时停止四叉树分割,如图8所示。
对斜线阴影部分进行四叉树分割,首先分成四部分,如若所分割的这部分没有斜线阴影区域则停止分割如图8中的b所示;如若分割区域内有斜线阴影区域则继续进行四叉树分割如图8中c、d、e所示。
具体步骤为:1.对图像进行子块分割。2.对分割的纯色子块进行区域合并。3.对非纯色 子块进行四叉树分割,之后进行区域合并。4.对纯色子块与非纯色子块所分割的区域进行区域合并。
1.对图像进行子块分割
1)对图像进行子块分割,将图像分割成m×n的子块。需要注意的是m和n的取值可根据图像的实际大小进行调整。
2)对子块进行分类,分为内部只有一种颜色的纯色子块和内部含有色差区域的非纯色子块。求出每个子块的颜色均值也就是颜色平均矢量为:
Figure PCTCN2021073124-appb-000009
其中x ij代表在一个c行像素和d列像素的子块中,第i行第j列的像素点的L、a、b值组成的矢量。
方差矢量为:
Figure PCTCN2021073124-appb-000010
纯色子块因为内部颜色均匀因此方差较小,非纯色子块因为内部含有色差区域因此方差较大。可以根据情况设定阈值,对纯色子块与非纯色子块进行区分。
2.对分割的纯色子块进行区域合并
(1)对于每一个纯色子块设立一个集合S i(i=1,2,3…k)。同时将每一个子块的颜色平均矢量
Figure PCTCN2021073124-appb-000011
设定为子块代表值,用C i(i=1,2,3…k)来表示。
(2)如果两个纯色子块的颜色距离ΔΕ≤6,则两个子块的集合S i、S j就合并到新的集合S v。新集合子块代表值C v=(C i+C j)/2。
(3)重复上一个步骤,直到所有子块集合的颜色距离ΔΕ≥6。停止集合合并。
(4)对集合内的所有子块进行区域合并,并进行颜色标记。
至此就完成了对纯色子块分割完毕了。纯色子块被分割为两部分,一部分为色差区域,另一部分为非色差区域。
3.对非纯色子块进行四叉树分割,之后进行区域合并
(1)对非纯色子块采用四叉树的方法进行分割。将非纯色子块分割成为大小形状相同的四个小子块。
(2)对四个小子块进行颜色均值与方差的计算。采用之前的阈值对四个小子块进行判断,如果判断为纯色的则停止四叉树分割;如果为非纯色,则继续进行四叉树分割。
(3)重复上(2)步骤,直到所有子块都为纯色子块或者为一个像素点时。结束四叉树分割。
(4)将四叉树分割过程中的第一个均匀块或者均匀的像素点作为集合P 1
(5)计算每一个经过四叉树分割过后子块与集合P 1的颜色距离ΔΕ。如果颜色距离ΔΕ≤6,则将新的子块并入P 1中;如果颜色距离ΔΕ≥6,则将子块并入新的集合P i(i=2,3,4…)中。重复本步骤,直到所有的子块均并入集合当中。
(6)比较除P 1之外的任意两个集合,如果颜色距离ΔΕ≤6,则将两个集合合并为一个集合,重新计算颜色均值。如果颜色距离ΔΕ≥6,则不进行合并。
重复(1)~(6)步骤,直到所有的非纯色子块处理完毕。
至此就完成了对非纯色子块的分割。非纯色子块也被分成了两部分,色差区域和非色差区域。
4.对纯色子块与非纯色子块所分割的区域进行区域合并。
(1)抽取像素所属于的集合,如果颜色距离ΔΕ≥6,则不进行合并。如果颜色距离ΔΕ≤6,则进行合并。
(2)对完成合并的色差区域与非色差区域,采用不同颜色进行标记。
至此就完成了对色差缺陷图像的分割,分割出了色差区域与非色差区域,
三、在lab颜色空间的基础上采用色差公式(7)(8),量化色差缺陷区域的差异程度。识别出色差区域的面积,量化色差缺陷区域的相对面积。
为了能够准确反映混凝土外观质量中的色差缺陷,规定了两个评价指标相对面积和色差差异程度。
1、色差度量
色差即人们在对于不同颜色感知的描述。在Lab颜色空间中用两点之间的空间坐标距离来表示色差ΔΕ,如式(9)~(12)。
△L=L 1-L 2公式(9)
△a=a 1-a 2公式(10)
△b=b 1-b 2公式(11)
Figure PCTCN2021073124-appb-000012
令△E ab=1时称之为一个NBS色差单位,根据Y.H.Gong等 [4]人研究发现,当图像的计算结果为小于3个NBS时,人眼识别结果为基本无色差;3~6个NBS时,有较大差别;大于12个NBS时,会被人眼识别为不同颜色。具体如表3。
表3 人类视觉与NBS颜色距离度量的对应关系
NBS 0~1.5 1.5~3 3~6 6~12 >12
人类视觉 相同 细微差别 较大差别 显著差别 不同颜色
差异程度表示为色差缺陷区域与整个混凝土的拍摄面颜色相差程度大小。颜色相差越大,则混凝土表面色差严重。采用颜色距离ΔΕ来表示差异程度。
2、相对面积
表示色差缺陷区域相对于整个图像区域的大小程度。这里采用B来表示相对面积,S O表示色差缺陷区域的面积,S表示整个混凝土拍摄面的面积。注意这里的S O与S,单位为像素。则B为:
Figure PCTCN2021073124-appb-000013
四、结合两个评价指标,参照质量评价表对色差区域进行评级。
评价指标确定后,对混凝土色差区域进行分级评价。此处依据《混凝土结构工程施工质量验收规范》(GB50204-2015),建立评价表。表格分为Ⅰ、Ⅱ、Ⅲ、Ⅳ、Ⅴ五级,级别越大代表色差缺陷越严重,如表4所示。其具体使用方法为根据相对面积和差异程度两个指标对色差缺陷进行评价。当两个指标所属的等级不同时,采用级别较大的级别作为评价结果。
表4 色差缺陷评价表
Figure PCTCN2021073124-appb-000014
下面结合具体的实施例,对本发明做进一步的详细说明,应该指出,所述具体实施例是对本发明的解释而不是限定。
实施例1
1.采集标准图像
首先在标准条件下,通过相机测温***调节,实现白平衡,拍摄标准图像。如图2。
2.对RGB图像进行三个颜色通道分离
将图2的RGB图像进行三个通道分离为:R(red),G(green),B(blue)三个通道,并获取三通道灰度值。分离后,分别得到代表R(red),G(green),B(blue)的三个矩阵,矩阵的值分布在0到255的范围里,对应的图像如图3。
图3将图像分解为三个颜色通道
3.分别计算三个颜色通道的标准差
三个颜色通道像素值的分布如图4,可以看出不同颜色通道的灰度值的分布有很大差异。
采用式(2)计算三个颜色通道的标准差S,计算结果如表5。
表5 图像的三个颜色通道的标准差
颜色通道 均值 标准差 变异系数
红色通道 114.19 41.58 0.364
蓝色通道 109.26 66.07 0.604
绿色通道 107.47 50.96 0.474
4.对清水混凝土色差进行评价
用灰度标准偏差值作为混凝土外观色泽一致性的评价标准,本例中三个颜色通道的标准差分别为红色通道:41.58,蓝色通道:66.07,绿色通道:50.96,三者最大值为66.07,将最大值66.07与表2的色差判断标准进行比较,灰度标准偏差大于10,判别为不合格。
实施例2
为了较好地说明本方法,采用某地铁站点内的清水混凝土图像进行分析。
1.采集标准图像
拍照时,被拍摄清水混凝土室内采用灯光照明,光线充足。拍照距离应为5米,拍摄的清水混凝土尺寸长为1米,宽为0.65米。通过相机测温***调节,实现白平衡,拍摄清水混凝土标准图像。如图5。
2.对RGB图像进行三个颜色通道分离
将图5的RGB图像进行三个通道分离为:R(red),G(green),B(blue)三个通道,并获取三通道灰度值。分离后,分别得到代表R(red),G(green),B(blue)的三个矩阵,矩阵的值分布在0到255的范围里,对应的图像如图6。
3.分别计算三个颜色通道的标准差
三个颜色通道像素值的分布如图7,可以看出红色通道的灰度值集中,差异小,蓝色通道的灰度值差异最大。
采用式(2)计算三个颜色通道的标准差S,计算结果如表6。
表6 图像的三个颜色通道的标准差
颜色通道 均值 标准差 变异系数
红色通道 119.48 11.55 0.097
蓝色通道 112.86 15.51 0.137
绿色通道 116.15 14.31 0.123
4.对清水混凝土色差进行评价
用灰度标准偏差值作为混凝土外观色泽一致性的评价标准,本例中三个颜色通道的标准差分别为红色通道:11.55,蓝色通道:15.51,绿色通道:14.31,三者最大值为15.51,将最大值15.51与表2的色差判断标准进行比较,灰度标准偏差大于10,判别为不合格。
实施例3
选取某桥墩进行色差缺陷的检测。
桥墩采用混凝土进行浇筑,由于养护不到位。某些桥墩上的区域已经开始产生色差缺陷。如图9所示。
图9中的三个图中,图9(a)是锈迹,由于柱子的上部金属构件生锈,导致下部柱子产生色差缺陷;图9(b)是污点,由于混凝土表面保护不到位,产生的色差区域;图9(c)是油渍,由于混凝土在成型时,脱模剂的使用不当,导致的色差缺陷。
采用本发明所提到的方法对图像进行色差评价。步骤为:
步骤一:对所需进行色差评价的区域进行拍摄,要求尽量保证镜头与拍摄面的平行。
步骤二:使用式(3)~(6)将图像的RGB颜色空间转换为Lab颜色空间如图10。
步骤三:对图像色差缺陷区域进行图像分割,识别出色差缺陷区域如图11。
步骤四:求出评价指标之一的相对面积。
步骤五:对色差缺陷区域与非色差缺陷区域进行评价指标差异程度的计算,具体做法为采用式(7)~(8),分别求出色差区域与非色差区域的颜色平均矢量
Figure PCTCN2021073124-appb-000015
再运用式(9)~(12)求出颜色距离ΔΕ,运用ΔΕ来判断差异程度。
步骤六:参照表3对色差缺陷区域进行级别的评价。结果如表7。
表7 评价结果
Figure PCTCN2021073124-appb-000016
Figure PCTCN2021073124-appb-000017
最后应该说明的是,以上所述仅为本发明的优选实施例而已,并不用于限制本发明,尽管参照前述实施例对本发明进行了详细的说明,对于本领域的技术人员来说,其依然可以对前述实施例所记载的技术方案进行修改,或者对其中部分进行等同替换。凡在本发明的精神和原则之内,所作的任何修改、等同替换、改进等,均应包含在本发明的保护范围之内。上述虽然对本发明的具体实施方式进行了描述,但并非对本发明保护范围的限制,所属领域技术人员应该明白,在本发明的技术方案的基础上,本领域技术人员不需要付出创造性劳动即可做出的各种修改或变形仍在本发明的保护范围以内。

Claims (10)

  1. 一种清水混凝土色差分析方法,其特征在于,包括:
    采集待测混凝土表面的标准图像;
    将所述标准图像转化为RGB图像,并进行三通道分离,得到三通道灰度值;
    分别计算三个颜色通道的标准差,得到三个颜色通道的标准差S;
    取三个颜色通道标准差的最大值,将灰度标准偏差小于6的评定为外观色泽一致性的一等,将灰度标准偏差小于10的评定为合格,将灰度标准偏差大于10的评定为不满足清水混凝土外观色泽一致性要求,即不合格。
  2. 如权利要求1所述的清水混凝土色差分析方法,其特征在于,色差区域是以正常混凝土表面为背景的目标物。
  3. 如权利要求1所述的清水混凝土色差分析方法,其特征在于,采集标准图像的过程中,拍照距离为4至6米,拍摄的照片对应被拍摄对象尺寸长为0.9至1.1米,宽为0.5至0.7米;
    先实现白平衡后再进行图像采集。
  4. 如权利要求1所述的清水混凝土色差分析方法,其特征在于,所述标准差的计算公式为
    Figure PCTCN2021073124-appb-100001
    其中,M和N分别代表颜色通道图像的行数和列数,Gray(i,j)代表各个像素点的灰度值,Gray代表整幅图像的灰度均值。
  5. 一种清水混凝土色差分析***,其特征在于,包括:
    用于采集待测混凝土表面的标准图像的模块;
    用于将标准图像转化为RGB图像的模块;
    用于对RGB图像进行三通道分离的模块;
    用于计算三个颜色通道的标准差的模块;
    用灰度标准偏差值进行混凝土外观色泽一致性的评价的模块。
  6. 一种清水混凝土色差的检测方法,其特征在于,包括:
    采集待测混凝土表面的标准图像,并转化为RGB图像;
    将RGB图像的颜色空间转换为Lab空间;
    采用区域合并与四叉树分割相结合的方法进行图像分割,分割出了色差区域与非色差区域;
    在lab颜色空间的基础上采用色差公式,对色差缺陷区域的差异程度进行量化;
    采用相对面积和色差差异程度两个指标,参照质量评价表对色差区域进行评级。
  7. 如权利要求6所述的清水混凝土色差的检测方法,其特征在于,所述四叉树分割时,若所分割的这部分没有斜线阴影区域则停止分割;若分割区域内有斜线阴影区域则继续进行四叉树分割。
  8. 如权利要求6所述的清水混凝土色差的检测方法,其特征在于,所述图像分割的具体步骤为:对图像进行子块分割;对分割的纯色子块进行区域合并;对非纯色子块进行四叉树分割,之后进行区域合并;对纯色子块与非纯色子块所分割的区域进行区域合并。
  9. 如权利要求6所述的清水混凝土色差的检测方法,其特征在于,采用颜色距离ΔΕ来表示色差差异程度,其中,
    △L=L 1-L 2公式(9)
    △a=a 1-a 2公式(10)
    △b=b 1-b 2公式(11)
    Figure PCTCN2021073124-appb-100002
    采用B来表示相对面积,SO表示色差缺陷区域的面积,S表示整个混凝土拍摄面的面积;其中:
    Figure PCTCN2021073124-appb-100003
  10. 一种清水混凝土色差的检测***,其特征在于,包括:
    用于采集待测混凝土表面的标准图像的模块;
    用于将标准图像转化为RGB图像的模块;
    用于将RGB图像的颜色空间转换为Lab空间的模块;
    用于图像分割的模块;
    用于对在lab颜色空间上色差缺陷区域的差异程度进行量化的模块;
    用于对色差区域进行评级的模块。
PCT/CN2021/073124 2020-06-10 2021-01-21 一种清水混凝土色差分析/检测方法及*** WO2021248915A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN202010523766.3 2020-06-10
CN202010523766.3A CN111681289A (zh) 2020-06-10 2020-06-10 一种清水混凝土色差分析/检测方法及***

Publications (1)

Publication Number Publication Date
WO2021248915A1 true WO2021248915A1 (zh) 2021-12-16

Family

ID=72454928

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2021/073124 WO2021248915A1 (zh) 2020-06-10 2021-01-21 一种清水混凝土色差分析/检测方法及***

Country Status (2)

Country Link
CN (1) CN111681289A (zh)
WO (1) WO2021248915A1 (zh)

Cited By (15)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114462543A (zh) * 2022-02-15 2022-05-10 河南四通实业有限公司 再生沥青混凝土生产用筛分方法
CN114543840A (zh) * 2022-02-25 2022-05-27 四川大学 环境因素对空三摄影影响实验装置和测评方法
CN114708623A (zh) * 2022-03-23 2022-07-05 北京航空航天大学 一种基于图像处理的金属腐蚀等级评价方法
CN114862778A (zh) * 2022-04-25 2022-08-05 河南农业大学 一种基于ImagePy的数字土壤形态测量方法
CN114972354A (zh) * 2022-08-02 2022-08-30 济宁金筑新型建材科技有限公司 基于图像处理的蒸压加气混凝土砌块生产控制方法及***
CN115060665A (zh) * 2022-08-16 2022-09-16 君华高科集团有限公司 一种食品安全自动巡检***
CN115690249A (zh) * 2022-11-03 2023-02-03 武汉纺织大学 一种纺织面料数字化色彩体系构建方法
CN116485797B (zh) * 2023-06-25 2023-08-29 潍坊亚贝涂料有限公司 基于人工智能的涂料色差快速检测方法
CN116805312A (zh) * 2023-08-21 2023-09-26 青岛时佳汇服装有限公司 一种基于图像处理的针织式面料质量检测方法
CN116883403A (zh) * 2023-09-07 2023-10-13 山东国宏生物科技有限公司 一种基于机器视觉的大豆品质检测方法
CN116993716A (zh) * 2023-09-22 2023-11-03 山东力加力钢结构有限公司 一种混凝土搅拌过程视觉检测方法
CN117095067A (zh) * 2023-10-17 2023-11-21 山东虹纬纺织有限公司 基于人工智能的纺织品色差检测方法
CN117351433A (zh) * 2023-12-05 2024-01-05 山东质能新型材料有限公司 基于计算机视觉的胶固砂浆饱满度监控***
CN117593303A (zh) * 2024-01-18 2024-02-23 浙江锦德光电材料有限公司 一种量子点光学膜的缺陷检测方法及***
CN117764938A (zh) * 2023-12-19 2024-03-26 北方工业大学 一种基于墨水染色实验的中间包死区评价方法及***

Families Citing this family (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111681289A (zh) * 2020-06-10 2020-09-18 青岛理工大学 一种清水混凝土色差分析/检测方法及***
CN112200242B (zh) * 2020-10-09 2023-06-20 陕西科技大学 绘画唐卡图像的颜色分类方法、计算机设备、终端及介质
CN112611754B (zh) * 2020-10-26 2023-12-29 中国水利水电第三工程局有限公司 一种清水混凝土外观质量评价方法
CN113838081A (zh) * 2021-09-15 2021-12-24 浙江中烟工业有限责任公司 一种基于机器视觉判别烤烟烟叶颜色均匀度的方法和装置
CN115219033B (zh) * 2022-07-18 2023-12-22 旭密林绿色建筑科技有限公司 一种彩色混凝土颜色的确定方法
CN115147410B (zh) * 2022-08-30 2023-12-05 南通夏克塑料包装有限公司 一种多色块包装覆膜缺陷检测方法
CN116787737B (zh) * 2023-07-06 2024-01-26 明大瑞天智能装备(浙江)有限公司 中空成型机挤吹质量检测***

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2017103982A1 (ja) * 2015-12-14 2017-06-22 株式会社 ニコン・トリンブル 欠陥検出装置及びプログラム
CN107730493A (zh) * 2017-10-24 2018-02-23 广东天机工业智能***有限公司 产品颜色差异检测方法、装置、介质以及计算机设备
CN110706294A (zh) * 2019-09-10 2020-01-17 武汉纺织大学 一种色纺织物颜色差异度检测方法
CN111179362A (zh) * 2019-12-14 2020-05-19 西安交通大学 一种基于动态光照校正算法的试纸颜色均匀性检测方法
CN111681289A (zh) * 2020-06-10 2020-09-18 青岛理工大学 一种清水混凝土色差分析/检测方法及***

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2017103982A1 (ja) * 2015-12-14 2017-06-22 株式会社 ニコン・トリンブル 欠陥検出装置及びプログラム
CN107730493A (zh) * 2017-10-24 2018-02-23 广东天机工业智能***有限公司 产品颜色差异检测方法、装置、介质以及计算机设备
CN110706294A (zh) * 2019-09-10 2020-01-17 武汉纺织大学 一种色纺织物颜色差异度检测方法
CN111179362A (zh) * 2019-12-14 2020-05-19 西安交通大学 一种基于动态光照校正算法的试纸颜色均匀性检测方法
CN111681289A (zh) * 2020-06-10 2020-09-18 青岛理工大学 一种清水混凝土色差分析/检测方法及***

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
LIU, HAIBAO: "Study on Key Techonoioies of Fair-faced Concrete in Subway Station", CHINESE MASTER'S THESES FULL-TEXT DATABASE, ENGINEERING SCIENCE & TECHNOLOGY II, 15 May 2018 (2018-05-15), pages 1 - 95, XP055878745, ISSN: 1674-0246 *

Cited By (24)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114462543B (zh) * 2022-02-15 2024-05-03 河南四通实业有限公司 再生沥青混凝土生产用筛分方法
CN114462543A (zh) * 2022-02-15 2022-05-10 河南四通实业有限公司 再生沥青混凝土生产用筛分方法
CN114543840A (zh) * 2022-02-25 2022-05-27 四川大学 环境因素对空三摄影影响实验装置和测评方法
CN114543840B (zh) * 2022-02-25 2023-07-25 四川大学 环境因素对空三摄影影响实验装置和测评方法
CN114708623A (zh) * 2022-03-23 2022-07-05 北京航空航天大学 一种基于图像处理的金属腐蚀等级评价方法
CN114862778A (zh) * 2022-04-25 2022-08-05 河南农业大学 一种基于ImagePy的数字土壤形态测量方法
CN114862778B (zh) * 2022-04-25 2024-02-27 河南农业大学 一种基于ImagePy的数字土壤形态测量方法
CN114972354A (zh) * 2022-08-02 2022-08-30 济宁金筑新型建材科技有限公司 基于图像处理的蒸压加气混凝土砌块生产控制方法及***
CN115060665A (zh) * 2022-08-16 2022-09-16 君华高科集团有限公司 一种食品安全自动巡检***
CN115690249A (zh) * 2022-11-03 2023-02-03 武汉纺织大学 一种纺织面料数字化色彩体系构建方法
CN116485797B (zh) * 2023-06-25 2023-08-29 潍坊亚贝涂料有限公司 基于人工智能的涂料色差快速检测方法
CN116805312A (zh) * 2023-08-21 2023-09-26 青岛时佳汇服装有限公司 一种基于图像处理的针织式面料质量检测方法
CN116805312B (zh) * 2023-08-21 2024-01-05 青岛时佳汇服装有限公司 一种基于图像处理的针织式面料质量检测方法
CN116883403B (zh) * 2023-09-07 2023-11-14 山东国宏生物科技有限公司 一种基于机器视觉的大豆品质检测方法
CN116883403A (zh) * 2023-09-07 2023-10-13 山东国宏生物科技有限公司 一种基于机器视觉的大豆品质检测方法
CN116993716B (zh) * 2023-09-22 2023-12-19 山东力加力钢结构有限公司 一种混凝土搅拌过程视觉检测方法
CN116993716A (zh) * 2023-09-22 2023-11-03 山东力加力钢结构有限公司 一种混凝土搅拌过程视觉检测方法
CN117095067A (zh) * 2023-10-17 2023-11-21 山东虹纬纺织有限公司 基于人工智能的纺织品色差检测方法
CN117095067B (zh) * 2023-10-17 2024-02-02 山东虹纬纺织有限公司 基于人工智能的纺织品色差检测方法
CN117351433A (zh) * 2023-12-05 2024-01-05 山东质能新型材料有限公司 基于计算机视觉的胶固砂浆饱满度监控***
CN117351433B (zh) * 2023-12-05 2024-02-23 山东质能新型材料有限公司 基于计算机视觉的胶固砂浆饱满度监控***
CN117764938A (zh) * 2023-12-19 2024-03-26 北方工业大学 一种基于墨水染色实验的中间包死区评价方法及***
CN117593303A (zh) * 2024-01-18 2024-02-23 浙江锦德光电材料有限公司 一种量子点光学膜的缺陷检测方法及***
CN117593303B (zh) * 2024-01-18 2024-04-09 浙江锦德光电材料有限公司 一种量子点光学膜的缺陷检测方法及***

Also Published As

Publication number Publication date
CN111681289A (zh) 2020-09-18

Similar Documents

Publication Publication Date Title
WO2021248915A1 (zh) 一种清水混凝土色差分析/检测方法及***
CN111294589B (zh) 摄像模组镜头表面检测方法
CN103402117B (zh) 基于Lab色度空间的视频图像偏色检测方法
CN102445239B (zh) 基于多点水位的明渠量水新方法
CN105445607A (zh) 一种基于等温线绘制的电力设备故障检测方法
CN106096603A (zh) 一种融合多特征的动态火焰检测方法及装置
CN105741328A (zh) 基于视觉感知的拍摄图像质量评价方法
CN107064169A (zh) 一种桥梁裂缝检测装置及检测方法
CN109035232A (zh) 一种基于深度学习和机器视觉融合的夹片缺陷检测方法
CN109523544B (zh) 一种建筑外墙质量缺陷检测***及其方法
CN104504722B (zh) 一种利用灰色点校正图像颜色的方法
CN112785579A (zh) 基于图像处理技术的混凝土裂缝识别方法
CN114627316B (zh) 基于人工智能的液压***漏油检测方法
CN106408526B (zh) 一种基于多层矢量图的能见度检测方法
CN107346429A (zh) 一种多晶电池片图像的颜色自动识别与分类方法
CN102800111A (zh) 基于颜色协调性的彩色融合图像颜色质量评价方法
CN106057700B (zh) 一种太阳能电池片的边红片的检测方法
CN104408714A (zh) 一种基于纹理抑制平滑滤波和分水岭算法的重织物图像分割方法
CN105719280A (zh) 一种p型涂胶视觉检测方法
CN105223208A (zh) 一种电路板检测模板及其制作方法、电路板检测方法
CN106815602B (zh) 一种基于多层次特征描述的跑道fod图像探测方法及装置
CN111311695B (zh) 一种基于卷积神经网络的清水混凝土表面色差分析方法
CN108830834B (zh) 一种爬索机器人视频缺陷信息自动提取方法
CN115761459A (zh) 一种面向桥隧表观病害识别的多场景自适应方法
CN110533698B (zh) 一种基于视觉检测的基坑施工桩检测控制方法

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 21822482

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 21822482

Country of ref document: EP

Kind code of ref document: A1