WO2021012735A1 - 屏幕显示缺陷的检测方法及系统 - Google Patents

屏幕显示缺陷的检测方法及系统 Download PDF

Info

Publication number
WO2021012735A1
WO2021012735A1 PCT/CN2020/086819 CN2020086819W WO2021012735A1 WO 2021012735 A1 WO2021012735 A1 WO 2021012735A1 CN 2020086819 W CN2020086819 W CN 2020086819W WO 2021012735 A1 WO2021012735 A1 WO 2021012735A1
Authority
WO
WIPO (PCT)
Prior art keywords
target image
pixel
screen display
module
frequency domain
Prior art date
Application number
PCT/CN2020/086819
Other languages
English (en)
French (fr)
Inventor
陈润康
邓远志
戴志威
陈志列
Original Assignee
研祥智能科技股份有限公司
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 研祥智能科技股份有限公司 filed Critical 研祥智能科技股份有限公司
Publication of WO2021012735A1 publication Critical patent/WO2021012735A1/zh

Links

Images

Classifications

    • 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
    • G01MEASURING; TESTING
    • G01MTESTING STATIC OR DYNAMIC BALANCE OF MACHINES OR STRUCTURES; TESTING OF STRUCTURES OR APPARATUS, NOT OTHERWISE PROVIDED FOR
    • G01M11/00Testing of optical apparatus; Testing structures by optical methods not otherwise provided for
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06TIMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T5/00Image enhancement or restoration
    • G06T5/70Denoising; Smoothing
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06VIMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00Arrangements for image or video recognition or understanding
    • G06V10/20Image preprocessing
    • G06V10/28Quantising the image, e.g. histogram thresholding for discrimination between background and foreground patterns
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06TIMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00Indexing scheme for image analysis or image enhancement
    • G06T2207/20Special algorithmic details
    • G06T2207/20048Transform domain processing
    • G06T2207/20056Discrete and fast Fourier transform, [DFT, FFT]

Definitions

  • the present invention relates to the technical field of image processing, in particular to a method and system for detecting screen display defects.
  • the existing method of detecting screen display defects through equipment usually adopts the contrast method for comparison to determine whether the screen has defects during display. Specifically, the measured image is compared with the normal image to obtain the difference information between the images, and then the defect points are obtained by the difference of the background gray threshold value.
  • the display defects of some screens are not uniform. For example, some dot-shaped display defects are fuzzy and irregular, and the brightness of the display defects of the screen is not uniform. This will reduce the display defect of the screen on the screen. Contrast, which in turn affects the test results.
  • the method and system for detecting screen display defects provided by the present invention can facilitate the observation and identification of display defects.
  • the present invention provides a method for detecting screen display defects, including:
  • the method before the step of converting the target image from the spatial domain to the frequency domain, the method further includes:
  • the method further includes:
  • the noise in the target image is reduced by means of filtering.
  • the judging whether the target image has defects through an image binarization algorithm includes:
  • the detection threshold and the pixel value of the pixel it is determined whether the pixel is a defect.
  • the step of determining the detection threshold includes:
  • the step of judging whether the pixel point is a defective point according to the detection threshold value and the pixel value of the pixel point includes:
  • the detection threshold According to the difference between the detection threshold and the pixel value of the pixel, and a preset local threshold, it is determined whether the pixel is a defect.
  • the judging whether the pixel point is a defective point according to the difference between the detection threshold value and the pixel value of the pixel point, and a preset local threshold value includes:
  • the difference is less than or equal to the local threshold, it is determined that the pixel point is not a defective point.
  • a detection system for screen display defects of the present invention includes:
  • the acquiring module is configured to acquire the target image displayed on the screen
  • the first conversion module is configured to convert the target image from the spatial domain to the frequency domain to highlight characteristic information of the target image, the characteristic information including high-frequency information;
  • a compression module configured to compress the high-frequency information
  • the second conversion module is configured to convert the target image from the frequency domain to the spatial domain
  • the judging module is configured to judge whether the target image has defects through an image binarization algorithm.
  • system further includes:
  • the enhancement module is configured to enhance the gray values of pixels with low gray values in the target image through logarithmic transformation before the first conversion module converts the target image from the spatial domain to the frequency domain.
  • system further includes:
  • the filtering module is configured to reduce noise in the target image through average filtering after the second conversion module converts the target image from the frequency domain to the spatial domain.
  • the judgment module includes:
  • the determining sub-module is configured to determine the detection threshold according to the pixel value of the neighboring block of the pixel in the target image
  • the judging sub-module is configured to judge whether the pixel is a defect according to the detection threshold and the pixel value of the pixel.
  • the screen display defect detection method and system provided by the embodiments of the present invention can facilitate the compression of high-frequency information in the target image by converting the target image from the spatial domain to the frequency domain, so that after passing the image binarization algorithm, The contrast between the display defect of the screen and the display background of the screen can be improved in the target image, thereby making the display defect of the screen more prominent in the target image, which facilitates the observation and recognition of the display defect.
  • FIG. 1 is a schematic flowchart of a method for detecting screen display defects according to an embodiment of the application
  • Figure 2 is a logarithmic transformation function diagram of an embodiment of the application
  • FIG. 3 is a schematic flowchart of a method for detecting screen display defects according to an embodiment of the application
  • Fig. 4 is a schematic structural diagram of a screen display defect detection system according to an embodiment of the application.
  • Convolution kernel the weight used in convolution, which is indicated by a matrix, which is a weight matrix.
  • Convolution operation the process of weighted summation. Each pixel in the used image area is multiplied by each element of the convolution kernel, that is, the weight, and the sum of all convolutions is used as the new value of the center pixel of the area .
  • Fourier transform (Discrete Fourier Transform, DFT), Fourier transform is the core of the Fourier analysis method, through which the signal is converted from the time domain or the space domain to the frequency domain, and then the spectral structure and changes of the signal can be studied law.
  • Inverse Fourier Transform Mainly convert the signal from the frequency domain to the time domain or the space domain.
  • Logarithmic transformation Adjust the grayscale value range of the input low-quality target image through the grayscale transformation function, that is, the logarithmic transformation formula, so as to expand the low gray value part of the target image and compress the high gray value part.
  • Frequency domain smoothing filter It is a method of smoothing the image in the frequency domain.
  • the frequency of information when the frequency of information is greater than or equal to 100KHz, it can be regarded as high frequency information. Signals with a frequency lower than 100KHz are regarded as low-frequency information.
  • the present invention provides a method for detecting screen display defects, especially applied to display defects of TV screens.
  • FIG. 1 shows a schematic flowchart of a method for detecting a screen display defect according to an embodiment of the present application, and the method includes:
  • Step S101 Obtain a target image displayed on the screen.
  • the target image is obtained by shooting with a camera when a preset image is displayed on the screen.
  • Step S102 Convert the target image from the spatial domain to the frequency domain to highlight characteristic information of the target image, the characteristic information including high frequency information.
  • Fourier transform is applied to the target image to transform the grayscale distribution function of the target image into the frequency distribution function of the target image.
  • Step S103 Compress the high-frequency information.
  • the target image that has undergone the Fourier transform is processed by a convolution operation, that is, the Gaussian difference algorithm, using Sigma, that is, ⁇ , and convolution is performed after the difference of two different Gaussian kernels.
  • the Fourier transformed target image is subtracted from the spatial information contained in the frequency band in another preset gray-scale image to suppress high frequency information and reduce the blur of the target image.
  • step S103 the method further includes:
  • Step S104 Convert the target image from the frequency domain to the spatial domain.
  • an inverse Fourier transform is used on the target image to transform the frequency distribution function of the target image into the grayscale distribution function of the target image.
  • Step S105 Determine whether the target image has defects through the image binarization algorithm.
  • the judging whether the target image has defects through an image binarization algorithm includes:
  • the detection threshold is determined according to the pixel value of the neighboring block of the pixel in the target image.
  • the detection threshold and the pixel value of the pixel it is determined whether the pixel is a defect.
  • the image binarization algorithm may use a local adaptive threshold method to make judgments.
  • the detection threshold is determined by the maximum and minimum pixel values in the neighborhood block; or the pixel values of the neighborhood block are arranged from large to small, and the pixel value at the center is taken as the detection threshold; or The average value of the pixel values of all neighboring blocks is used as the detection threshold.
  • the binarization threshold can be determined for the pixel distribution in the surrounding area.
  • the screen background grayscale will change and the defect is fuzzy.
  • the defective part can be accurately segmented, so as to facilitate observation and recognition Whether the screen is defective.
  • the step of determining the detection threshold includes:
  • the detection threshold is determined according to the average value of the pixel values of the neighboring blocks of the pixel points in the target image.
  • the step of judging whether the pixel point is a defective point according to the detection threshold value and the pixel value of the pixel point includes:
  • the detection threshold According to the difference between the detection threshold and the pixel value of the pixel, and a preset local threshold, it is determined whether the pixel is a defect.
  • the average value is used as the detection threshold, so that the binarization threshold of each pixel in the target image is not exactly the same, and the specific is determined by the setting of the pixels in the area around the pixel, so that the brightness is high.
  • the binarization threshold of pixel points is higher than the binarization threshold value of pixels with low brightness, and local image areas with different brightness, contrast, and texture will have corresponding local binarization thresholds.
  • the judging whether the pixel is a defect based on the difference between the detection threshold and the pixel value of the pixel and a preset local threshold includes:
  • the difference is greater than the local threshold, it is determined that the pixel point is a defective point.
  • the difference is less than or equal to the local threshold, it is determined that the pixel point is not a defective point.
  • a window with a preset pixel size of 3*3 is used to slide pixel by pixel in the target image until the entire target image is traversed; all pixels in the window are calculated in the image subregion corresponding to each window The average value of the pixel value; compare the average value with the pixel value of each pixel in the original image sub-region, if the subtractive floating range exceeds the preset local threshold, the pixel is determined to be a defective point.
  • the method before the step of converting the target image from the spatial domain to the frequency domain, the method further includes:
  • Equation 1 the formula for logarithmic transformation is Equation 1:
  • the output gray value after logarithmic transformation is 130, which is obviously larger than the gray value after the original gray value function transformation. , That is, through logarithmic transformation, the gray value of pixels with low gray values can be enhanced.
  • the output gray value after logarithmic transformation is 180, which is obviously smaller than the gray value after the original gray value function transformation, that is, the pixels with high gray value can be reduced through logarithmic transformation The gray value of the point.
  • the logarithmic transformation can not only map a narrow range of low gray values in the target image to a wider range of gray values, but also map a wider range of high gray values in the target image to a larger range
  • the narrow grayscale interval expands the grayscale value of low grayscale and compresses the grayscale value of high grayscale, which can enhance the low grayscale details in the target image.
  • the method further includes:
  • the noise in the target image is reduced by means of filtering.
  • the characteristic information includes brightness and texture.
  • the average filter is used to filter the target image.
  • the average gray value of the pixels in the average filter window is used to replace the original gray value of the pixel in the center of the window in the target image, thereby reducing the sharpness of the target image.
  • the degree of change can further reduce the noise of the target image and at the same time blur the edges of the target image.
  • Equation 2 the working principle of the average filter is shown in Equation 2.
  • s xy denote a filter window with a center point at (x, y) and a size of m ⁇ n.
  • g(s,t) represents the original image
  • f(x,y) represents the image obtained after mean filtering.
  • the arithmetic mean filter simply calculates the average value of the pixels in the window area, and then assigns the average value to the pixel at the center of the window.
  • the present invention can facilitate the compression of high-frequency information in the target image, so that after the image binarization algorithm is passed, the display defects and screen performance of the screen can be improved in the target image.
  • the contrast of the display background makes the display defect of the screen more prominent in the target image, which is convenient for the observation and recognition of the display defect.
  • the present invention provides a method for detecting screen display defects, especially applied to display defects of TV screens.
  • FIG. 3 shows a schematic flowchart of a method for detecting screen display defects according to an embodiment of the present application, and the method includes:
  • Step S301 Obtain the target image displayed on the screen.
  • the target image is obtained by shooting with a camera when a preset image is displayed on the screen.
  • Step S302 Through logarithmic transformation, the gray value of a pixel with a low gray value in the target image is enhanced.
  • Step S303 Convert the target image from the spatial domain to the frequency domain to highlight the characteristic information of the target image, the characteristic information including high frequency information.
  • Fourier transform is applied to the target image to transform the grayscale distribution function of the target image into the frequency distribution function of the target image.
  • Step S304 Compress the high frequency information.
  • the target image that has undergone the Fourier transform is processed by a convolution operation, that is, the Gaussian difference algorithm, using Sigma, that is, ⁇ , and convolution is performed after the difference of two different Gaussian kernels.
  • the Fourier transformed target image is subtracted from the spatial information contained in the frequency band in another preset gray-scale image to suppress high frequency information and reduce the blur of the target image.
  • Step S305 Convert the target image from the frequency domain to the spatial domain.
  • the inverse Fourier transform is applied to the target image to transform the frequency distribution function of the target image into the gray distribution function of the target image.
  • Step S306 Reduce noise in the target image through average filtering.
  • Step S307 Determine the detection threshold according to the average value of the pixel values of the neighboring blocks of the pixel points in the target image.
  • Step S308 According to the difference between the detection threshold and the pixel value of the pixel, and a preset local threshold, determine whether the pixel is a defect.
  • the difference is greater than the local threshold, it is determined that the pixel is a defective point; if the difference is less than or equal to the local threshold, it is determined that the pixel is not a defective point.
  • the present invention can facilitate the compression of high-frequency information in the target image, so that after the image binarization algorithm is passed, the display defects and screen performance of the screen can be improved in the target image.
  • the contrast of the display background makes the display defect of the screen more prominent in the target image, which is convenient for the observation and recognition of the display defect.
  • FIG. 4 shows a schematic structural diagram of a system for detecting a screen display defect according to an embodiment of the present application, including:
  • the obtaining module 401 is configured to obtain a target image displayed on the screen.
  • the first conversion module 402 is configured to convert the target image from the spatial domain to the frequency domain to highlight characteristic information of the target image, the characteristic information including high-frequency information.
  • the compression module 403 is configured to compress the high-frequency information.
  • the second conversion module 404 is configured to convert the target image from the frequency domain to the spatial domain.
  • the judgment module 405 is configured to judge whether the target image has a defect through an image binarization algorithm.
  • system further includes:
  • the enhancement module is configured to enhance the gray value of pixels with low gray values in the target image through logarithmic transformation before the first conversion module 402 converts the target image from the spatial domain to the frequency domain .
  • system further includes:
  • the filtering module is configured to reduce noise in the target image through average filtering after the second conversion module 404 converts the target image from the frequency domain to the spatial domain.
  • the judgment module 405 includes:
  • the determining sub-module is configured to determine the detection threshold according to the pixel value of the neighboring block of the pixel in the target image.
  • the judging sub-module is configured to judge whether the pixel is a defect according to the detection threshold and the pixel value of the pixel.
  • the determining submodule includes:
  • the determining unit is configured to determine the detection threshold according to the average value of the pixel values of the neighboring blocks of the pixel points in the target image.
  • the judgment sub-module includes:
  • the judging unit is configured to judge whether the pixel is a defect according to the difference between the detection threshold and the pixel value of the pixel, and a preset local threshold.
  • the determining unit is further configured to determine that the pixel point is a defective point when the difference value is greater than the local threshold value.
  • the difference is less than or equal to the local threshold, it is determined that the pixel point is not a defective point.
  • the present invention can facilitate the compression of high-frequency information in the target image, so that after the image binarization algorithm is passed, the display defects and screen performance of the screen can be improved in the target image.
  • the contrast of the display background makes the display defect of the screen more prominent in the target image, which is convenient for the observation and recognition of the display defect.

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Multimedia (AREA)
  • Quality & Reliability (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Chemical & Material Sciences (AREA)
  • Analytical Chemistry (AREA)
  • Image Processing (AREA)

Abstract

一种屏幕显示缺陷的检测方法及系统。所述方法包括:获取屏幕显示的目标图像(S101);将所述目标图像从空间域转换到频率域,以突出所述目标图像的特征信息,所述特征信息包括高频信息(S102);压缩所述高频信息(S103);将所述目标图像从频率域转换到空间域(S104);通过图像二值化算法,判断所述目标图像是否存在缺陷(S105)。通过将目标图像从空间域转换到频率域,能够便于对所述目标图像中高频信息进行压缩,从而在通过图像二值化算法后,能够在目标图像中提高屏幕的显示缺陷与屏幕的显示背景的对比度,进而使得屏幕的显示缺陷在目标图像中更加突出,便于显示缺陷的观察与识别。

Description

屏幕显示缺陷的检测方法及系统 技术领域
本发明涉及图像处理技术领域,尤其涉及一种屏幕显示缺陷的检测方法及系统。
背景技术
随着图像显示技术的发展,显示屏幕逐渐向大尺寸、高分辨率的方向发展。但屏幕的缺陷出现几率也随之增加,而对屏幕缺陷检测,已经成为一个重要的研究方向。
目前的屏幕生产厂家大都使用人工的方式检测屏幕在显示的时候是否存在缺陷,一方面极大的增加了人力成本;另一方面由于个人主观原因和视觉疲劳程度的影响,会出现大量的误测、漏测和过测,难以满足实际的工业需求。
现有通过设备检测屏幕显示缺陷的方法通常是采用差影法进行比对,来判断屏幕在显示的时候是否存在缺陷。具体是把被测图像与正常图像进行对比,得出图像间的差异信息,再通过背景灰度的阈值之差得出缺陷点。但是由于一些屏幕的显示缺陷表现并不均匀,比如一些点状的显示缺陷比较模糊且无规则,同时屏幕的显示缺陷本身的亮度也并不均匀,如此则会降低屏幕的显示缺陷在屏幕上的对比度,进而影响检测结果。
发明内容
为解决上述问题,本发明提供的屏幕显示缺陷的检测方法及系统,能够便于显示缺陷的观察与识别。
第一方面,本发明提供一种屏幕显示缺陷的检测方法,包括:
获取屏幕显示的目标图像;
将所述目标图像从空间域转换到频率域,以突出所述目标图像的特征信息,所述特征信息包括高频信息;
压缩所述高频信息;
将所述目标图像从频率域转换到空间域;
通过图像二值化算法,判断所述目标图像是否存在缺陷。
可选地,在所述将所述目标图像从空间域转换到频率域的步骤之前,所述方法还包括:
通过对数变换,增强所述目标图像中灰度值低的像素点的灰度值。
可选地,在所述将所述目标图像从频率域转换到空间域的步骤之后,所述方法还包括:
通过均值滤波降低所述目标图像中的噪声。
可选地,所述通过图像二值化算法,判断所述目标图像是否存在缺陷,包括:
根据所述目标图像中的像素点的邻域块的像素值,确定检测阈值;
根据所述检测阈值和所述像素点的像素值,判断所述像素点是否为缺陷点。
可选地,所述确定检测阈值的步骤,包括:
根据所述目标图像中的像素点的邻域块的像素值的平均值,确定检测阈值;
所述根据所述检测阈值和所述像素点的像素值,判断所述像素点是否为缺陷点的步骤,包括:
根据所述检测阈值与所述像素点的像素值之间的差值,以及预设的局部阈值,判断所述像素点是否为缺陷点。
可选地,所述根据所述检测阈值与所述像素点的像素值之间的差值,以及预设的局部阈值,判断所述像素点是否为缺陷点,包括:
在所述差值大于所述局部阈值的情况下,判定所述像素点为缺陷点;
在所述差值小于或等于所述局部阈值的情况下,判定所述像素点不为缺陷点。
第二方面,本发明一种屏幕显示缺陷的检测系统,包括:
获取模块,被配置为获取屏幕显示的目标图像;
第一转换模块,被配置为将所述目标图像从空间域转换到频率域,以突出所述目标图像的特征信息,所述特征信息包括高频信息;
压缩模块,被配置为压缩所述高频信息;
第二转换模块,被配置为将所述目标图像从频率域转换到空间域;
判断模块,被配置为通过图像二值化算法,判断所述目标图像是否存在缺陷。
可选地,所述系统还包括:
增强模块,被配置为在所述第一转换模块将所述目标图像从空间域转换到频率域之前,通过对数变换,增强所述目标图像中灰度值低的像素点的灰度值。
可选地,所述系统还包括:
滤波模块,被配置为在所述第二转换模块将所述目标图像从频率域转换到空间域之后,通过均值滤波降低所述目标图像中的噪声。
可选地,所述判断模块包括:
确定子模块,被配置为根据所述目标图像中的像素点的邻域块的像素值,确定检测阈值;
判断子模块,被配置为根据所述检测阈值和所述像素点的像素值,判断所 述像素点是否为缺陷点。
本发明实施例提供的屏幕显示缺陷的检测方法及系统,通过将目标图像从空间域转换到频率域,能够便于对所述目标图像中高频信息进行压缩,从而在通过图像二值化算法后,能够在目标图像中提高屏幕的显示缺陷与屏幕的显示背景的对比度,进而使得屏幕的显示缺陷在目标图像中更加突出,便于显示缺陷的观察与识别。
附图说明
图1为本申请实施例的屏幕显示缺陷的检测方法的示意性流程图;
图2为本申请实施例的对数变换函数图;
图3为本申请实施例的屏幕显示缺陷的检测方法的示意性流程图;
图4为本申请实施例的屏幕显示缺陷的检测系统的示意性结构图。
具体实施方式
为使本发明实施例的目的、技术方案和优点更加清楚,下面将结合本发明实施例中的附图,对本发明实施例中的技术方案进行清楚、完整地描述,显然,所描述的实施例仅仅是本发明一部分实施例,而不是全部的实施例。基于本发明中的实施例,本领域普通技术人员在没有做出创造性劳动前提下所获得的所有其他实施例,都属于本发明保护的范围。
首先,对本发明涉及的专有术语进行解释。
卷积核:卷积时使用到的权,用一个矩阵标示,该矩阵是一个权矩阵。
卷积运算:加权求和的过程,使用到的图像区域中的每个像素分别与卷积核的每个元素,即权值,对应相乘,所有卷积之和作为区域中心像素的新值。
傅里叶变换:(Discrete Fourier Transform,DFT),傅里叶变换是傅里叶分析方法的核心,通过它把信号从时间域或空间域转换到频率域,进而可以研究 信号的频谱结构和变化规律。
傅里叶逆变换:主要将信号从频率域转为时间域或空间域。
对数变换:通过灰度变换函数,即对数变换公式,调整输入低质的目标图像的灰度值范围,以将目标图像的低灰度值部分扩展,高灰度值部分压缩。
频域平滑滤波:是使图像在频率域中进行平滑处理的一种方法。
在本发明中,信息的频率大于或等于100KHz时,则可以将其视为高频信息。频率低于100KHz的信号视为低频信息。
第一方面,本发明提供一种屏幕显示缺陷的检测方法,特别是应用于电视机屏幕的显示缺陷。参见图1,图1示出了根据本申请一实施例的屏幕显示缺陷的检测方法的示意性流程图,所述方法包括:
步骤S101:获取屏幕显示的目标图像。
具体的,所述目标图像是在所述屏幕显示预设的图像的情况下,通过相机拍摄得到的。
步骤S102:将所述目标图像从空间域转换到频率域,以突出所述目标图像的特征信息,所述特征信息包括高频信息。
具体的,在本实施例中,对所述目标图像采用傅里叶变换,以将所述目标图像的灰度分布函数变换为所述目标图像的频率分布函数。
步骤S103:压缩所述高频信息。
具体的,对经过傅里叶变换的目标图像采用卷积运算进行处理,即高斯差分算法,采用Sigma,即σ,不同的两个高斯核做差后进行卷积。将经过傅里叶变换的目标图像通过减去另一幅预设的灰度图像中的频带所包含的空间信息以压制高频信息,降低目标图像的模糊度。
在一种可选的实施例中,在步骤S103之后,所述方法还包括:
对目标图像进行频域平滑滤波处理,以去除噪声改善目标图像的质量,从而达到平滑目标图像的目的。
步骤S104:将所述目标图像从频率域转换到空间域。
具体的,在本实施例中,对所述目标图像采用傅里叶逆变换,以将所述目标图像的频率分布函数变换为所述目标图像的灰度分布函数。
步骤S105:通过图像二值化算法,判断所述目标图像是否存在缺陷。
在一种可选的实施例中,所述通过图像二值化算法,判断所述目标图像是否存在缺陷,包括:
根据所述目标图像中的像素点的邻域块的像素值,确定检测阈值。
根据所述检测阈值和所述像素点的像素值,判断所述像素点是否为缺陷点。
具体的,所述图像二值化算法可采用局部自适应阈值的方法进行判断。例如,通过邻域块中最大和最小的像素值,确定所述检测阈值;或将邻域块的像素值从大到小排列,取位于中心位置的像素值作为所述检测阈值;再或者将所有邻域块的像素值的平均值作为所述检测阈值。
通过局部自适应阈值能够对周围领域像素分布决定二值化阈值,对于目标图像中屏幕背景灰度会产生变化并且缺陷较为模糊的图像能够精准的对有缺陷的部分进行分割,从而便于观察与识别屏幕是否存在缺陷。
在一种可选的实施例中,所述确定检测阈值的步骤,包括:
根据所述目标图像中的像素点的邻域块的像素值的平均值,确定检测阈值。
所述根据所述检测阈值和所述像素点的像素值,判断所述像素点是否为缺陷点的步骤,包括:
根据所述检测阈值与所述像素点的像素值之间的差值,以及预设的局部阈值,判断所述像素点是否为缺陷点。
采用平均值作为检测阈值的方法,使得所述目标图像中每个像素点的二值化阈值并不完全相同的,具体的是由像素点周围领域像素点的设置来确定的,从而使亮度高的像素点的二值化阈值高于亮度低的像素点的二值化阈值,并使不同亮度、不同对比度、不同纹理的局部图像区域将会拥有相对应的局部二值化阈值。
在一种可选的实施例中,所述根据所述检测阈值与所述像素点的像素值之间的差值,以及预设的局部阈值,判断所述像素点是否为缺陷点,包括:
在所述差值大于所述局部阈值的情况下,判定所述像素点为缺陷点。
在所述差值小于或等于所述局部阈值的情况下,判定所述像素点不为缺陷点。
在一种可选的实施例中,利用预设像素大小为3*3的窗口在目标图像中逐像素滑动,直到遍历整个目标图像;在每个窗口对应的图像子区域中计算窗口内所有像素的像素值的平均值;将平均值与原图像子区域每个像素点的像素值进行对比,如果相减浮动的范围超过预设的局部阈值,则判定该像素点为缺陷点。
在一种可选的实施例中,在所述将所述目标图像从空间域转换到频率域的步骤之前,所述方法还包括:
通过对数变换,增强所述目标图像中灰度值低的像素点的灰度值。
其中,对数变换的公式为式一:
s=c*log(1+r)   式一
参见图2,图2示出了根据本申请一实施例的对数变换函数图,其中,s= r为原灰度值函数;c是一个常数,r为输入灰度值,s为输出灰度值。
根据图2中的对数函数的曲线可以得出:当输入灰度值为50时,对数变换后的输出灰度值为130,明显比经过原灰度值函数变换后的灰度值大,即通过对数变换能够增强灰度值低的像素点的灰度的值。当输入灰度值为250时,对数变换后的输出灰度值为180,明显比经过原灰度值函数变换后的灰度值小,即通过对数变换能够降低灰度值高的像素点的灰度的值。
具体的,通过对数变换不但能够将目标图像中范围较窄的低灰度值映射到范围较宽的灰度区间,同时也能够将目标图像中范围较宽的高灰度值区间映射为较窄的灰度区间,从而扩展了低灰度的灰度值,压缩了高灰度的灰度值,进而能够对目标图像中低灰度细节进行增强。
在一种可选的实施例中,在所述将所述目标图像从频率域转换到空间域的步骤之后,所述方法还包括:
通过均值滤波降低所述目标图像中的噪声。
所述特征信息包括:亮度和纹理等。采用均值滤波器对目标图像进行滤波处理,主要是通过均值滤波器窗口内的像素的平均灰度值代替目标图像中位于窗口中心点处的像素原有的灰度值,从而降低目标图像的尖锐变化程度,进而在降低所述目标图像的噪声,同时还能够模糊所述目标图像的边缘。
具体的,所述均值滤波器的工作原理参见式二,
Figure PCTCN2020086819-appb-000001
令s xy表示中心点在(x,y)处,且大小为m×n的滤波器窗口。其中,g(s,t)表示原始图像,f(x,y)表示均值滤波后得到的图像。算术均值滤波器就是简单的计算窗口区域的像素均值,然后将均值赋值给窗口中心点处的像素。
本发明通过将目标图像从空间域转换到频率域,能够便于对所述目标图像中高频信息进行压缩,从而在通过图像二值化算法后,能够在目标图像中提高屏幕的显示缺陷与屏幕的显示背景的对比度,进而使得屏幕的显示缺陷在目标图像中更加突出,便于显示缺陷的观察与识别。
第二方面,本发明提供一种屏幕显示缺陷的检测方法,特别是应用于电视机屏幕的显示缺陷。参见图3,图3示出了根据本申请一实施例的屏幕显示缺陷的检测方法的示意性流程图,所述方法包括:
步骤S301:获取屏幕显示的目标图像。
具体的,所述目标图像是在所述屏幕显示预设的图像的情况下,通过相机拍摄得到的。
步骤S302:通过对数变换,增强所述目标图像中灰度值低的像素点的灰度值。
步骤S303:将所述目标图像从空间域转换到频率域,以突出所述目标图像的特征信息,所述特征信息包括高频信息。
具体的,在本实施例中,对所述目标图像采用傅里叶变换,以将所述目标图像的灰度分布函数变换为所述目标图像的频率分布函数。
步骤S304:压缩所述高频信息。
具体的,对经过傅里叶变换的目标图像采用卷积运算进行处理,即高斯差分算法,采用Sigma,即σ,不同的两个高斯核做差后进行卷积。将经过傅里叶变换的目标图像通过减去另一幅预设的灰度图像中的频带所包含的空间信息以压制高频信息,降低目标图像的模糊度。
步骤S305:将所述目标图像从频率域转换到空间域。
具体的,在本实施例中,对所述目标图像采用傅里叶逆变换,以将所述目 标图像的频率分布函数变换为所述目标图像的灰度分布函数。
步骤S306:通过均值滤波降低所述目标图像中的噪声。
步骤S307:根据所述目标图像中的像素点的邻域块的像素值的平均值,确定检测阈值。
步骤S308:根据所述检测阈值与所述像素点的像素值之间的差值,以及预设的局部阈值,判断所述像素点是否为缺陷点。
若所述差值大于所述局部阈值,则判定所述像素点为缺陷点;若所述差值小于或等于所述局部阈值,则判定所述像素点不为缺陷点。
本发明通过将目标图像从空间域转换到频率域,能够便于对所述目标图像中高频信息进行压缩,从而在通过图像二值化算法后,能够在目标图像中提高屏幕的显示缺陷与屏幕的显示背景的对比度,进而使得屏幕的显示缺陷在目标图像中更加突出,便于显示缺陷的观察与识别。
第三方面,本发明一种屏幕显示缺陷的检测系统400,参见图4,图4示出了根据本申请一实施例的屏幕显示缺陷的检测系统的示意性结构图,包括:
获取模块401,被配置为获取屏幕显示的目标图像。
第一转换模块402,被配置为将所述目标图像从空间域转换到频率域,以突出所述目标图像的特征信息,所述特征信息包括高频信息。
压缩模块403,被配置为压缩所述高频信息。
第二转换模块404,被配置为将所述目标图像从频率域转换到空间域。
判断模块405,被配置为通过图像二值化算法,判断所述目标图像是否存在缺陷。
在一种可选的实施例中,所述系统还包括:
增强模块,被配置为在所述第一转换模块402将所述目标图像从空间域转 换到频率域之前,通过对数变换,增强所述目标图像中灰度值低的像素点的灰度值。
在一种可选的实施例中,所述系统还包括:
滤波模块,被配置为在所述第二转换模块404将所述目标图像从频率域转换到空间域之后,通过均值滤波降低所述目标图像中的噪声。
在一种可选的实施例中,所述判断模块405包括:
确定子模块,被配置为根据所述目标图像中的像素点的邻域块的像素值,确定检测阈值。
判断子模块,被配置为根据所述检测阈值和所述像素点的像素值,判断所述像素点是否为缺陷点。
在一种可选的实施例中,所述确定子模块包括:
确定单元,被配置为根据所述目标图像中的像素点的邻域块的像素值的平均值,确定检测阈值。
所述判断子模块包括:
判断单元,被配置为根据所述检测阈值与所述像素点的像素值之间的差值,以及预设的局部阈值,判断所述像素点是否为缺陷点。
在一种可选的实施例中,所述判断单元,进一步被配置为在所述差值大于所述局部阈值的情况下,判定所述像素点为缺陷点。
在所述差值小于或等于所述局部阈值的情况下,判定所述像素点不为缺陷点。
本发明通过将目标图像从空间域转换到频率域,能够便于对所述目标图像中高频信息进行压缩,从而在通过图像二值化算法后,能够在目标图像中提高屏幕的显示缺陷与屏幕的显示背景的对比度,进而使得屏幕的显示缺陷在目标 图像中更加突出,便于显示缺陷的观察与识别。
以上所述,仅为本发明的具体实施方式,但本发明的保护范围并不局限于此,任何熟悉本技术领域的技术人员在本发明揭露的技术范围内,可轻易想到的变化或替换,都应涵盖在本发明的保护范围之内。因此,本发明的保护范围应该以权利要求的保护范围为准。

Claims (10)

  1. 一种屏幕显示缺陷的检测方法,其特征在于,包括:
    获取屏幕显示的目标图像;
    将所述目标图像从空间域转换到频率域,以突出所述目标图像的特征信息,所述特征信息包括高频信息;
    压缩所述高频信息;
    将所述目标图像从频率域转换到空间域;
    通过图像二值化算法,判断所述目标图像是否存在缺陷。
  2. 根据权利要求1所述的屏幕显示缺陷的检测方法,其特征在于,在所述将所述目标图像从空间域转换到频率域的步骤之前,所述方法还包括:
    通过对数变换,增强所述目标图像中灰度值低的像素点的灰度值。
  3. 根据权利要求1或2所述的屏幕显示缺陷的检测方法,其特征在于,在所述将所述目标图像从频率域转换到空间域的步骤之后,所述方法还包括:
    通过均值滤波降低所述目标图像中的噪声。
  4. 根据权利要求1所述的屏幕显示缺陷的检测方法,其特征在于,所述通过图像二值化算法,判断所述目标图像是否存在缺陷,包括:
    根据所述目标图像中的像素点的邻域块的像素值,确定检测阈值;
    根据所述检测阈值和所述像素点的像素值,判断所述像素点是否为缺陷点。
  5. 根据权利要求4所述的屏幕显示缺陷的检测方法,其特征在于,所述确定检测阈值的步骤,包括:
    根据所述目标图像中的像素点的邻域块的像素值的平均值,确定检测阈 值;
    所述根据所述检测阈值和所述像素点的像素值,判断所述像素点是否为缺陷点的步骤,包括:
    根据所述检测阈值与所述像素点的像素值之间的差值,以及预设的局部阈值,判断所述像素点是否为缺陷点。
  6. 根据权利要求5所述的屏幕显示缺陷的检测方法,其特征在于,所述根据所述检测阈值与所述像素点的像素值之间的差值,以及预设的局部阈值,判断所述像素点是否为缺陷点,包括:
    在所述差值大于所述局部阈值的情况下,判定所述像素点为缺陷点;
    在所述差值小于或等于所述局部阈值的情况下,判定所述像素点不为缺陷点。
  7. 一种屏幕显示缺陷的检测系统,其特征在于,包括:
    获取模块,被配置为获取屏幕显示的目标图像;
    第一转换模块,被配置为将所述目标图像从空间域转换到频率域,以突出所述目标图像的特征信息,所述特征信息包括高频信息;
    压缩模块,被配置为压缩所述高频信息;
    第二转换模块,被配置为将所述目标图像从频率域转换到空间域;
    判断模块,被配置为通过图像二值化算法,判断所述目标图像是否存在缺陷。
  8. 根据权利要求7所述的屏幕显示缺陷的检测系统,其特征在于,所述系统还包括:
    增强模块,被配置为在所述第一转换模块将所述目标图像从空间域转换到频率域之前,通过对数变换,增强所述目标图像中灰度值低的像素点的灰度值。
  9. 根据权利要求7或8所述的屏幕显示缺陷的检测系统,其特征在于,所述系统还包括:
    滤波模块,被配置为在所述第二转换模块将所述目标图像从频率域转换到空间域之后,通过均值滤波降低所述目标图像中的噪声。
  10. 根据权利要求7所述的屏幕显示缺陷的检测系统,其特征在于,所述判断模块包括:
    确定子模块,被配置为根据所述目标图像中的像素点的邻域块的像素值,确定检测阈值;
    判断子模块,被配置为根据所述检测阈值和所述像素点的像素值,判断所述像素点是否为缺陷点。
PCT/CN2020/086819 2019-07-25 2020-04-24 屏幕显示缺陷的检测方法及系统 WO2021012735A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201910675533.2 2019-07-25
CN201910675533.2A CN110895806A (zh) 2019-07-25 2019-07-25 屏幕显示缺陷的检测方法及系统

Publications (1)

Publication Number Publication Date
WO2021012735A1 true WO2021012735A1 (zh) 2021-01-28

Family

ID=69785443

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2020/086819 WO2021012735A1 (zh) 2019-07-25 2020-04-24 屏幕显示缺陷的检测方法及系统

Country Status (2)

Country Link
CN (1) CN110895806A (zh)
WO (1) WO2021012735A1 (zh)

Cited By (19)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113113128A (zh) * 2021-04-15 2021-07-13 王小娟 基于vr、算法和5g技术的医疗手术辅助系统以及方法
CN113393437A (zh) * 2021-06-15 2021-09-14 中航华东光电有限公司 基于oled显示模块的缺陷检测方法
CN114486903A (zh) * 2021-12-06 2022-05-13 浙江大学台州研究院 一种灰度自适应的卷材检测系统及装置、算法
CN114549441A (zh) * 2022-02-14 2022-05-27 中船重工鹏力(南京)智能装备系统有限公司 基于图像处理的吸管缺陷检测方法
CN114757949A (zh) * 2022-06-15 2022-07-15 济宁市海富电子科技有限公司 基于计算机视觉的电线电缆缺陷检测方法及系统
CN115393359A (zh) * 2022-10-28 2022-11-25 菲特(天津)检测技术有限公司 一种镜片表面缺陷检测用的图像处理方法及系统
CN115631198A (zh) * 2022-12-21 2023-01-20 深圳新视智科技术有限公司 玻璃显示屏的裂纹检测方法、装置及计算机设备
CN115661156A (zh) * 2022-12-28 2023-01-31 成都数联云算科技有限公司 图像生成方法、装置、存储介质、设备及计算机程序产品
CN116405661A (zh) * 2023-04-28 2023-07-07 可诺特软件(深圳)有限公司 一种智能电视开发性能测试方法和装置
CN116452580A (zh) * 2023-06-13 2023-07-18 山东古天电子科技有限公司 一种笔记本外观质量检测方法
CN116754566A (zh) * 2023-08-17 2023-09-15 绍兴旭源新材料科技有限公司 一种柔性折叠屏幕保护膜检测方法
CN116843688A (zh) * 2023-09-01 2023-10-03 山东虹纬纺织有限公司 一种纺织品质量视觉检测方法
CN116934752A (zh) * 2023-09-18 2023-10-24 山东国泰民安玻璃科技有限公司 一种基于人工智能的玻璃检测方法及系统
CN117115128A (zh) * 2023-09-11 2023-11-24 杭州深度视觉科技有限公司 图像的像素值计算方法、装置、电子设备及存储介质
CN117152180A (zh) * 2023-10-31 2023-12-01 山东克莱蒙特新材料科技有限公司 基于人工智能的矿物铸件缺陷检测方法
CN117274065A (zh) * 2023-11-21 2023-12-22 深圳市芯科云科技有限公司 基于双芯片交替处理的智能穿戴手表屏幕显示超分辨方法
CN117437600A (zh) * 2023-12-20 2024-01-23 山东海纳智能装备科技股份有限公司 基于图像识别技术的煤流监测系统
CN117745724A (zh) * 2024-02-20 2024-03-22 高唐县瑞景精密机械有限公司 基于视觉分析的石材打磨加工缺陷区域分割方法
CN117808809A (zh) * 2024-03-01 2024-04-02 深圳市志合云创科技有限公司 晶圆表面缺陷的视觉检测方法及系统

Families Citing this family (9)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110895806A (zh) * 2019-07-25 2020-03-20 研祥智能科技股份有限公司 屏幕显示缺陷的检测方法及系统
WO2021189259A1 (zh) * 2020-03-24 2021-09-30 京东方科技集团股份有限公司 Oled屏幕点缺陷判定方法、装置、存储介质及电子设备
CN111539954B (zh) * 2020-05-25 2024-01-23 国网湖南省电力有限公司 采用x射线数字影像特征识别电缆缓冲层缺陷的方法、系统及介质
CN111815630B (zh) * 2020-08-28 2020-12-15 歌尔股份有限公司 一种用于lcd屏幕的缺陷检测方法、装置
CN112184723B (zh) * 2020-09-16 2024-03-26 杭州三坛医疗科技有限公司 图像处理方法及装置、电子设备、存储介质
CN112235540A (zh) * 2020-10-09 2021-01-15 北京航空航天大学 一种用于屏幕显示故障识别报警的智能视频监控系统
CN112801891B (zh) * 2021-01-11 2023-10-03 Tcl华星光电技术有限公司 显示屏画面检测方法、显示屏画面检测系统
CN114820610B (zh) * 2022-06-29 2022-09-06 数聚(山东)医疗科技有限公司 基于图像处理的新材料医疗器械缺陷检测方法
CN116912233B (zh) * 2023-09-04 2024-02-13 深圳市明亚顺科技有限公司 基于液晶显示屏的缺陷检测方法、装置、设备及存储介质

Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN104978748A (zh) * 2015-07-06 2015-10-14 电子科技大学 一种基于局部像素值的液晶屏缺陷检测方法
CN106157310A (zh) * 2016-07-06 2016-11-23 南京汇川图像视觉技术有限公司 基于混合自适应水平集模型与多通道结合的TFT LCD mura缺陷检测方法
US20180184079A1 (en) * 2016-08-08 2018-06-28 Shenzhen China Star Optoelectronics Technology Co., Ltd. System and method for detecting image sticking on display
CN108280822A (zh) * 2017-12-20 2018-07-13 歌尔科技有限公司 屏幕划痕的检测方法及装置
CN109166111A (zh) * 2018-08-15 2019-01-08 苏州富鑫林光电科技有限公司 一种电子墨水屏缺陷检测方法及系统
CN109345528A (zh) * 2018-09-28 2019-02-15 凌云光技术集团有限责任公司 一种基于人眼视觉特性的显示屏缺陷检测方法及装置
CN110895806A (zh) * 2019-07-25 2020-03-20 研祥智能科技股份有限公司 屏幕显示缺陷的检测方法及系统

Family Cites Families (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US9105075B1 (en) * 2013-02-06 2015-08-11 Ihs Global Inc. Enhancing seismic features using an optical filter array
CN105320962B (zh) * 2015-10-21 2016-10-12 东南大学 一种基于分类器集成的路面破损类型识别方法
CN108169236A (zh) * 2016-12-07 2018-06-15 广州映博智能科技有限公司 一种基于视觉的金属表面缺陷检测方法
CN106846267A (zh) * 2016-12-30 2017-06-13 四川迪派锐科技有限公司 一种管道焊缝检测用多尺度图像对比度增强方法
CN106780486B (zh) * 2017-01-16 2019-12-24 大连海事大学 一种钢板表面缺陷图像提取方法
CN106815821B (zh) * 2017-01-23 2020-08-28 上海兴芯微电子科技有限公司 近红外图像的去噪方法和装置
CN107749268B (zh) * 2017-10-27 2021-06-29 歌尔光学科技有限公司 屏幕检测方法及设备
CN108426537B (zh) * 2018-01-30 2020-10-09 华侨大学 一种基于线阵相机的在位砂轮快速全场检测方法及系统

Patent Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN104978748A (zh) * 2015-07-06 2015-10-14 电子科技大学 一种基于局部像素值的液晶屏缺陷检测方法
CN106157310A (zh) * 2016-07-06 2016-11-23 南京汇川图像视觉技术有限公司 基于混合自适应水平集模型与多通道结合的TFT LCD mura缺陷检测方法
US20180184079A1 (en) * 2016-08-08 2018-06-28 Shenzhen China Star Optoelectronics Technology Co., Ltd. System and method for detecting image sticking on display
CN108280822A (zh) * 2017-12-20 2018-07-13 歌尔科技有限公司 屏幕划痕的检测方法及装置
CN109166111A (zh) * 2018-08-15 2019-01-08 苏州富鑫林光电科技有限公司 一种电子墨水屏缺陷检测方法及系统
CN109345528A (zh) * 2018-09-28 2019-02-15 凌云光技术集团有限责任公司 一种基于人眼视觉特性的显示屏缺陷检测方法及装置
CN110895806A (zh) * 2019-07-25 2020-03-20 研祥智能科技股份有限公司 屏幕显示缺陷的检测方法及系统

Cited By (32)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113113128A (zh) * 2021-04-15 2021-07-13 王小娟 基于vr、算法和5g技术的医疗手术辅助系统以及方法
CN113393437A (zh) * 2021-06-15 2021-09-14 中航华东光电有限公司 基于oled显示模块的缺陷检测方法
CN114486903B (zh) * 2021-12-06 2024-05-14 浙江大学台州研究院 一种灰度自适应的卷材检测系统及装置、算法
CN114486903A (zh) * 2021-12-06 2022-05-13 浙江大学台州研究院 一种灰度自适应的卷材检测系统及装置、算法
CN114549441A (zh) * 2022-02-14 2022-05-27 中船重工鹏力(南京)智能装备系统有限公司 基于图像处理的吸管缺陷检测方法
CN114757949A (zh) * 2022-06-15 2022-07-15 济宁市海富电子科技有限公司 基于计算机视觉的电线电缆缺陷检测方法及系统
CN115393359A (zh) * 2022-10-28 2022-11-25 菲特(天津)检测技术有限公司 一种镜片表面缺陷检测用的图像处理方法及系统
CN115631198A (zh) * 2022-12-21 2023-01-20 深圳新视智科技术有限公司 玻璃显示屏的裂纹检测方法、装置及计算机设备
CN115631198B (zh) * 2022-12-21 2023-08-08 深圳新视智科技术有限公司 玻璃显示屏的裂纹检测方法、装置及计算机设备
CN115661156B (zh) * 2022-12-28 2023-04-14 成都数联云算科技有限公司 图像生成方法、装置、存储介质、设备及计算机程序产品
CN115661156A (zh) * 2022-12-28 2023-01-31 成都数联云算科技有限公司 图像生成方法、装置、存储介质、设备及计算机程序产品
CN116405661A (zh) * 2023-04-28 2023-07-07 可诺特软件(深圳)有限公司 一种智能电视开发性能测试方法和装置
CN116405661B (zh) * 2023-04-28 2023-09-29 可诺特软件(深圳)有限公司 一种智能电视开发性能测试方法和装置
CN116452580A (zh) * 2023-06-13 2023-07-18 山东古天电子科技有限公司 一种笔记本外观质量检测方法
CN116452580B (zh) * 2023-06-13 2023-09-01 山东古天电子科技有限公司 一种笔记本外观质量检测方法
CN116754566A (zh) * 2023-08-17 2023-09-15 绍兴旭源新材料科技有限公司 一种柔性折叠屏幕保护膜检测方法
CN116754566B (zh) * 2023-08-17 2023-10-31 绍兴旭源新材料科技有限公司 一种柔性折叠屏幕保护膜检测方法
CN116843688A (zh) * 2023-09-01 2023-10-03 山东虹纬纺织有限公司 一种纺织品质量视觉检测方法
CN116843688B (zh) * 2023-09-01 2023-11-24 山东虹纬纺织有限公司 一种纺织品质量视觉检测方法
CN117115128A (zh) * 2023-09-11 2023-11-24 杭州深度视觉科技有限公司 图像的像素值计算方法、装置、电子设备及存储介质
CN116934752A (zh) * 2023-09-18 2023-10-24 山东国泰民安玻璃科技有限公司 一种基于人工智能的玻璃检测方法及系统
CN116934752B (zh) * 2023-09-18 2023-12-08 山东国泰民安玻璃科技有限公司 一种基于人工智能的玻璃检测方法及系统
CN117152180B (zh) * 2023-10-31 2024-01-26 山东克莱蒙特新材料科技有限公司 基于人工智能的矿物铸件缺陷检测方法
CN117152180A (zh) * 2023-10-31 2023-12-01 山东克莱蒙特新材料科技有限公司 基于人工智能的矿物铸件缺陷检测方法
CN117274065A (zh) * 2023-11-21 2023-12-22 深圳市芯科云科技有限公司 基于双芯片交替处理的智能穿戴手表屏幕显示超分辨方法
CN117274065B (zh) * 2023-11-21 2024-02-27 深圳市芯科云科技有限公司 基于双芯片交替处理的智能穿戴手表屏幕显示超分辨方法
CN117437600A (zh) * 2023-12-20 2024-01-23 山东海纳智能装备科技股份有限公司 基于图像识别技术的煤流监测系统
CN117437600B (zh) * 2023-12-20 2024-03-26 山东海纳智能装备科技股份有限公司 基于图像识别技术的煤流监测系统
CN117745724A (zh) * 2024-02-20 2024-03-22 高唐县瑞景精密机械有限公司 基于视觉分析的石材打磨加工缺陷区域分割方法
CN117745724B (zh) * 2024-02-20 2024-04-26 高唐县瑞景精密机械有限公司 基于视觉分析的石材打磨加工缺陷区域分割方法
CN117808809A (zh) * 2024-03-01 2024-04-02 深圳市志合云创科技有限公司 晶圆表面缺陷的视觉检测方法及系统
CN117808809B (zh) * 2024-03-01 2024-05-14 深圳市志合云创科技有限公司 晶圆表面缺陷的视觉检测方法及系统

Also Published As

Publication number Publication date
CN110895806A (zh) 2020-03-20

Similar Documents

Publication Publication Date Title
WO2021012735A1 (zh) 屏幕显示缺陷的检测方法及系统
Golestaneh et al. No-reference quality assessment of JPEG images via a quality relevance map
US10521885B2 (en) Image processing device and image processing method
Jaya et al. IEM: a new image enhancement metric for contrast and sharpness measurements
US8532428B2 (en) Noise reducing apparatus, noise reducing method, and noise reducing program for improving image quality
CN109472788B (zh) 一种飞机铆钉表面的伤痕检测方法
Balanov et al. Image quality assessment based on DCT subband similarity
CN111932532A (zh) 胶囊内窥镜无参考图像评价方法、电子设备及介质
Xiong et al. Application of histogram equalization for image enhancement in corrosion areas
Zhang et al. Reduced reference image quality assessment based on statistics of edge
Tang et al. A reduced-reference quality assessment metric for super-resolution reconstructed images with information gain and texture similarity
CN117351019B (zh) 一种焊接缺陷检测方法
Nevriyanto et al. Image enhancement using the image sharpening, contrast enhancement, and Standard Median Filter (Noise Removal) with pixel-based and human visual system-based measurements
Vu et al. S3: a spectral and spatial sharpness measure
CN114332081A (zh) 基于图像处理的纺织品表面异常判定方法
CN116740098B (zh) 一种铝合金氩弧焊焊接图像分割方法及系统
CN114897885B (zh) 一种红外图像质量综合评价系统和方法
Asatryan et al. Quality assessment measure based on image structural properties
JP5321939B2 (ja) 画質検査装置および画質検査方法
KR101432227B1 (ko) 전자기기에서 비트 해상도 증가 및 에지 강화 방법 및 장치
Qin et al. Traffic sign image enhancement in low light environment
Asadi Amiri et al. No-reference image quality assessment based on localized discrete cosine transform for JPEG compressed images
Isono et al. Exponentiated combination of two scores for image quality assessment
Sharma et al. A comparative analysis of various image enhancement techniques for facial images
Kumar et al. Visual information fidelity in evaluating retinex enhancement algorithms

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: 20843804

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: 20843804

Country of ref document: EP

Kind code of ref document: A1

122 Ep: pct application non-entry in european phase

Ref document number: 20843804

Country of ref document: EP

Kind code of ref document: A1