WO2021147383A1 - 相机模糊检测方法及装置 - Google Patents

相机模糊检测方法及装置 Download PDF

Info

Publication number
WO2021147383A1
WO2021147383A1 PCT/CN2020/120886 CN2020120886W WO2021147383A1 WO 2021147383 A1 WO2021147383 A1 WO 2021147383A1 CN 2020120886 W CN2020120886 W CN 2020120886W WO 2021147383 A1 WO2021147383 A1 WO 2021147383A1
Authority
WO
WIPO (PCT)
Prior art keywords
photos
detected
blurred
camera
template
Prior art date
Application number
PCT/CN2020/120886
Other languages
English (en)
French (fr)
Inventor
刘尧
常树林
陈敏
Original Assignee
上海万物新生环保科技集团有限公司
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 上海万物新生环保科技集团有限公司 filed Critical 上海万物新生环保科技集团有限公司
Priority to JP2022506070A priority Critical patent/JP2022542948A/ja
Publication of WO2021147383A1 publication Critical patent/WO2021147383A1/zh

Links

Images

Classifications

    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04NPICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N17/00Diagnosis, testing or measuring for television systems or their details
    • H04N17/002Diagnosis, testing or measuring for television systems or their details for television cameras
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06NCOMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00Computing arrangements based on biological models
    • G06N3/02Neural networks
    • G06N3/04Architecture, e.g. interconnection topology
    • G06N3/045Combinations of networks
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06NCOMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00Computing arrangements based on biological models
    • G06N3/02Neural networks
    • G06N3/08Learning methods

Definitions

  • the invention relates to the field of computers, in particular to a camera blur detection method and device.
  • An object of the present invention is to provide a camera blur detection method and device.
  • a camera blur detection method including:
  • the camera to be detected is judged to be a problematic camera with blurred shooting.
  • training a camera blur detection model based on the preset number of clear template photos and blur template photos includes:
  • each clear template photo is divided into multiple small clear photos in sequence.
  • the two adjacent small clear photos in the segmentation sequence have some overlapping areas;
  • the step size divides each blurred template photo into multiple small blurred photos in sequence, among which two adjacent small blurred photos in the segmentation sequence have a partial overlap area;
  • judging whether the photo to be detected is blurred based on the camera blur detection model includes:
  • Each of the photos to be detected is sequentially divided into multiple small pieces of photos to be detected according to the preset size and step length, wherein the two adjacent small pieces of photos to be detected in the segmentation sequence have parts Overlapping area
  • the probability of a small piece of photo to be detected relative to the clear category is greater than the preset probability threshold, it is determined that the small piece of photo to be detected is blurry
  • the template image includes black stripes arranged at preset intervals on a white background, and the black stripes in the template image are in multiple rows.
  • the odd-numbered rows of black stripes in the template image form a first preset angle with the positive direction of the horizontal direction
  • the even-numbered rows of black stripes in the template image form a second predetermined angle with the positive direction of the horizontal direction.
  • a camera blur detection device wherein the device includes:
  • the template acquisition module is used to acquire a preset number of clear template photos and fuzzy template photos obtained by shooting the template image;
  • the shooting module is used to control the camera to be detected to shoot the template image to obtain the photos to be detected;
  • the determination module is configured to determine whether the photo to be detected is blurred based on the camera blur detection model; if the photo to be detected is determined to be blurred, then the camera to be detected is determined to be a camera with blurred shooting.
  • the training module is used to sequentially divide each clear template photo into a plurality of small clear photos according to a preset size and step length, wherein the two adjacent ones in the segmentation order
  • the small clear photos have some overlapping areas; each blurred template photo is divided into multiple small blurred photos according to the preset size and step length.
  • the two adjacent small blurred photos in the segmentation sequence have Partially overlapping area; obtain multiple small background photos of the preset size; multiple small clear photos representing the clear category, multiple small blurred photos representing the fuzzy category, and multiple small blocks representing the background category
  • the background photos are put into the convolutional neural network model for training to obtain the camera blur detection model.
  • the determination module is configured to divide each of the photos to be detected into a plurality of small pieces of photos to be detected in sequence according to a preset size and step length, wherein the segmentation Two adjacent small pieces of photos to be detected in sequence have a partial overlap area; each small piece of photos to be detected is input into the camera blur detection model to obtain each small piece of photos to be detected relative to the clear category and the fuzzy category.
  • the probability of the background category if the probability of a small piece of photo to be detected relative to the clear category is greater than the preset probability threshold, the small piece of photo to be detected is judged to be blurred; if it is judged to be a blurry piece of the photo to be detected If the number exceeds the preset number threshold, it is determined that the photo to be detected is blurred.
  • the template image includes black stripes arranged at preset intervals on a white background, and the black stripes in the template image are in multiple rows.
  • the odd rows of black stripes in the template image and the horizontal direction are at a first preset angle
  • the even rows of black stripes in the template image are at a second angle to the horizontal direction.
  • a computing-based device which includes:
  • a memory arranged to store computer-executable instructions which, when executed, cause the processor to:
  • the camera to be detected is judged to be a problematic camera with blurred shooting.
  • a computer-readable storage medium having computer-executable instructions stored thereon, wherein, when the computer-executable instructions are executed by a processor, the processor:
  • the camera to be detected is judged to be a problematic camera with blurred shooting.
  • the present invention trains a camera blur detection model based on the preset number of clear template photos and blur template photos, and based on the camera blur detection model, determines whether the photo to be detected is blurred, and then based on the As a result of whether the photo to be detected is blurred, it is determined whether the camera to be detected is a blurry camera. If the photo to be detected is judged to be blurry, the camera to be detected is judged to be a problematic camera with blurry shooting; otherwise, it is judged as a clear camera. The problem camera can accurately and efficiently determine the shooting performance of the camera to be tested.
  • Fig. 1 shows a flowchart of a camera blur detection method according to an embodiment of the present invention
  • Fig. 2 shows a schematic diagram of a template image according to an embodiment of the present invention.
  • the terminal, the equipment of the service network, and the trusted party all include one or more processors (CPU), input/output interfaces, network interfaces, and memory.
  • processors CPU
  • input/output interfaces network interfaces
  • memory volatile and non-volatile memory
  • the memory may include non-permanent memory in computer readable media, random access memory (RAM) and/or non-volatile memory, such as read-only memory (ROM) or flash memory (flash RAM). Memory is an example of computer readable media.
  • RAM random access memory
  • ROM read-only memory
  • flash RAM flash memory
  • Computer-readable media include permanent and non-permanent, removable and non-removable media, and information storage can be realized by any method or technology.
  • the information can be computer-readable instructions, data structures, program modules, or other data.
  • Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, CD-ROM, digital versatile disc (DVD) or other optical storage, Magnetic cassettes, magnetic tape disk storage or other magnetic storage devices or any other non-transmission media can be used to store information that can be accessed by computing devices.
  • computer-readable media does not include non-transitory computer-readable media (transitory media), such as modulated data signals and carrier waves.
  • the present invention provides a camera blur detection method, the method includes:
  • Step S1 obtaining a preset number of clear template photos and blurred template photos obtained by shooting the template image
  • the template image captured in the clear template photo is completely clear, and the template image captured in the blurred template photo is all or partly blurred;
  • the camera may be a camera on a smart terminal such as a mobile phone or a PAD;
  • Step S2 training a camera blur detection model based on the preset number of clear template photos and blur template photos;
  • Step S3 controlling the camera to be detected to shoot the template image to obtain the photo to be detected
  • the camera of takes the template image to obtain a photo to be tested;
  • Step S4 judging whether the photo to be detected is blurred based on the camera blur detection model
  • step S5 if the photo to be detected is judged to be blurred, then the camera to be detected is judged to be a problematic camera with blurred shooting.
  • the present invention trains a camera blur detection model based on the preset number of clear template photos and blur template photos, and based on the camera blur detection model, determines whether the photo to be detected is blurred, and then based on whether the photo to be detected is blurred As a result of the blur, it is judged whether the camera to be inspected is a blurry camera. If the photograph to be inspected is judged to be blurry, the camera to be inspected is judged to be the camera with blurry shooting, otherwise it is judged to be a camera with no problem with clear shooting. The shooting performance of the camera to be tested can be judged accurately and efficiently.
  • step S2 training a camera blur detection model based on the preset number of clear template photos and blur template photos, includes:
  • each clear template photo is sequentially divided into a plurality of small clear photos according to the preset size and step length, wherein two adjacent small clear photos in the segmentation sequence have a partial overlap area;
  • Set the size and step length to divide each blurred template photo into multiple small blurred photos in sequence.
  • two adjacent small blurred photos in the segmentation sequence have a partial overlap area;
  • the preset size and step length it can be divided to two adjacent small blocks in the division order, and the clear picture has a partial overlap area, and at the same time, it can be divided into two adjacent small blocks in the division sequence. There are some overlapping areas in the photos to facilitate accurate training of subsequent models;
  • Step S22 acquiring a plurality of small background photos of the preset size
  • Step S23 Put multiple small clear photos representing the clear category, multiple small blurred photos representing the fuzzy category, and multiple small background photos representing the background category into the convolutional neural network model for training to obtain all Describes the camera blur detection model.
  • three types of photos representing multiple small clear photos representing a clear category, multiple small blurred photos representing a fuzzy category, and multiple small background photos representing a background category are put into the convolutional neural network.
  • the network model is trained, and the camera blur detection model can be obtained efficiently and reliably.
  • step S4 judging whether the photo to be detected is blur based on the camera blur detection model, includes:
  • each of the photos to be detected is sequentially divided into a plurality of small pieces of photos to be detected according to a preset size and step length, wherein two adjacent small pieces in the order of segmentation are to be detected
  • the photo has some overlapping areas;
  • the segmentation method of the photos to be detected is the same as the segmentation method of the template photos in the previous model training;
  • Step S42 input each small piece of photo to be detected into the camera blur detection model to obtain the probability of each small piece of photo to be detected relative to the clear category, the fuzzy category and the background category;
  • Step S43 if the probability of a certain small block of photos to be detected relative to the clear category is greater than the preset probability threshold, it is determined that the small block of photos to be detected is blurred;
  • step S44 if the number of small pieces of photos to be detected that are determined to be blurred exceeds the preset number threshold, then it is determined that the photos to be detected are blurred.
  • the template image includes black stripes arranged at preset intervals on a white background, and the black stripes in the template image are in multiple rows.
  • the present invention displays the template image as black stripes arranged at preset intervals on a white background, so that it is easier to identify whether each black stripe is in the process of training the model and identifying the photos to be detected by the model. Clear or fuzzy.
  • the odd-numbered rows of black stripes in the template image and the horizontal direction are at a first preset angle, and the even-numbered rows of black stripes in the template image are aligned with the horizontal direction.
  • the black stripes are arranged obliquely and longitudinally with a preset included angle with the horizontal direction, which facilitates subsequent image recognition of the black and white stripe area, and improves the recognition efficiency and reliability.
  • the black stripes in the first row and the third row in the template image form an angle of 60 degrees with the positive horizontal direction
  • the black stripes in the second row and the fourth row in the template image are at an angle of 60 degrees with the horizontal direction.
  • the positive direction of the direction is at an angle of 120 degrees, which facilitates the subsequent identification of the fuzzy position of the fringe more efficiently.
  • the power line 1 of the backlight source can be connected to provide an illumination light source for the template image.
  • the spacing distance of the black stripes in each row may be 10 mm, the spacing distance of the black stripes in the same row is 8 mm, and the width of each black stripe is 8 mm.
  • the present invention provides a camera blur detection device, the device includes:
  • the template acquisition module is used to acquire a preset number of clear template photos and fuzzy template photos obtained by shooting the template image;
  • the template image captured in the clear template photo is completely clear, and the template image captured in the blurred template photo is all or partly blurred;
  • the camera may be a camera on a smart terminal such as a mobile phone or a PAD;
  • a training module for training a camera blur detection model based on the preset number of clear template photos and blur template photos;
  • the photographing module is used to control the camera to be inspected to photograph the template image to obtain the photograph to be inspected;
  • the camera of takes the template image to obtain a photo to be tested;
  • the determination module is configured to determine whether the photo to be detected is blurred based on the camera blur detection model; if the photo to be detected is determined to be blurred, then the camera to be detected is determined to be a camera with blurred shooting.
  • the present invention trains a camera blur detection model based on the preset number of clear template photos and blur template photos, and based on the camera blur detection model, determines whether the photo to be detected is blurred, and then based on whether the photo to be detected is blurred As a result of the blur, it is judged whether the camera to be inspected is a blurry camera. If the photograph to be inspected is judged to be blurry, the camera to be inspected is judged to be the camera with blurry shooting, otherwise it is judged as the camera with no problem with clear shooting, The shooting performance of the camera to be tested can be judged accurately and efficiently.
  • the training module is used to sequentially divide each clear template photo into a plurality of small clear photos according to a preset size and step length.
  • Two adjacent small clear photos have a partial overlap area; each blurred template photo is divided into multiple small blurred photos according to the preset size and step length.
  • the two adjacent ones are divided in order Small fuzzy photos have a partial overlap area; multiple small background photos of the preset size are obtained; multiple small clear photos representing the clear category, multiple small blurred photos representing the blurred category, and background categories are obtained respectively
  • a plurality of small background photos of is put into a convolutional neural network model for training, so as to obtain the camera blur detection model.
  • the preset size and step length it can be divided to two adjacent small blocks in the division order, and the clear picture has a partial overlap area, and at the same time, it can be divided into two adjacent small blocks in the division sequence. There are some overlapping areas in the photos to facilitate accurate training of subsequent models;
  • three types of photos representing multiple small clear photos representing a clear category, multiple small blurred photos representing a fuzzy category, and multiple small background photos representing a background category are put into the convolutional neural network model. Training can efficiently and reliably train the camera blur detection model.
  • the determination module is configured to divide the photos to be detected into multiple small pieces of photos to be detected in sequence according to a preset size and step length. , Where two adjacent small pieces of photos to be detected in the segmentation sequence have a partial overlap area; each small piece of photos to be detected is input into the camera blur detection model to obtain each small piece of photos to be detected relative to The probability of clear category, fuzzy category and background category; if the probability of a small piece of photo to be detected relative to the clear category is greater than the preset probability threshold, then the small piece of photo to be detected is judged to be blurry; if judged to be a small piece of blur If the number of photos to be detected exceeds the preset number threshold, it is determined that the photos to be detected are blurred.
  • the segmentation method of the photos to be detected is the same as the segmentation method of the template photos in the previous model training;
  • the template image includes black stripes arranged at preset intervals on a white background, and the black stripes in the template image are in multiple rows.
  • the present invention displays the template image as black stripes arranged at preset intervals on a white background, so that it is easier to identify whether each black stripe is in the process of training the model and identifying the photos to be detected by the model. Clear or fuzzy.
  • the odd-numbered rows of black stripes in the template image and the horizontal direction are at a first preset angle, and the even-numbered rows of black stripes in the template image are aligned with the horizontal direction.
  • the black stripes are arranged obliquely and longitudinally with a preset included angle with the horizontal direction, which facilitates subsequent image recognition of the black and white stripe area, and improves the recognition efficiency and reliability.
  • the black stripes in the first row and the third row in the template image form an angle of 60 degrees with the positive horizontal direction
  • the black stripes in the second row and the fourth row in the template image are at an angle of 60 degrees with the horizontal direction.
  • the positive direction of the direction is at an angle of 120 degrees, which facilitates the subsequent identification of the fuzzy position of the fringe more efficiently.
  • the power line 1 of the backlight source can be connected to provide an illumination light source for the template image.
  • the spacing distance of the black stripes in each row may be 10 mm, the spacing distance of the black stripes in the same row is 8 mm, and the width of each black stripe is 8 mm.
  • a computing-based device which includes:
  • a memory arranged to store computer-executable instructions which, when executed, cause the processor to:
  • the camera to be detected is judged to be a problematic camera with blurred shooting.
  • a computer-readable storage medium having computer-executable instructions stored thereon, wherein, when the computer-executable instructions are executed by a processor, the processor:
  • the camera to be detected is judged to be a problematic camera with blurred shooting.
  • the present invention can be implemented in software and/or a combination of software and hardware.
  • it can be implemented by an application specific integrated circuit (ASIC), a general purpose computer or any other similar hardware device.
  • the software program of the present invention may be executed by a processor to realize the above-mentioned steps or functions.
  • the software program (including related data structure) of the present invention can be stored in a computer-readable recording medium, such as a RAM memory, a magnetic or optical drive or a floppy disk and similar devices.
  • some steps or functions of the present invention may be implemented by hardware, for example, as a circuit that cooperates with a processor to execute each step or function.
  • a part of the present invention can be applied as a computer program product, such as a computer program instruction, when it is executed by a computer, through the operation of the computer, the method and/or technical solution according to the present invention can be invoked or provided.
  • the program instructions for invoking the method of the present invention may be stored in a fixed or removable recording medium, and/or transmitted through a data stream in a broadcast or other signal-bearing medium, and/or stored in accordance with the Said program instructions run in the working memory of the computer equipment.
  • an embodiment according to the present invention includes a device including a memory for storing computer program instructions and a processor for executing the program instructions, wherein, when the computer program instructions are executed by the processor, trigger
  • the operation of the device is based on the aforementioned methods and/or technical solutions according to multiple embodiments of the present invention.

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Health & Medical Sciences (AREA)
  • Health & Medical Sciences (AREA)
  • Biomedical Technology (AREA)
  • Computing Systems (AREA)
  • General Engineering & Computer Science (AREA)
  • Data Mining & Analysis (AREA)
  • Evolutionary Computation (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Molecular Biology (AREA)
  • Artificial Intelligence (AREA)
  • Computational Linguistics (AREA)
  • General Physics & Mathematics (AREA)
  • Mathematical Physics (AREA)
  • Software Systems (AREA)
  • Biophysics (AREA)
  • Multimedia (AREA)
  • Signal Processing (AREA)
  • Image Analysis (AREA)
  • Studio Devices (AREA)

Abstract

本发明的目的是提供一种相机模糊检测方法及装置,本发明基于所述预设数量的清晰模版照片和模糊模版照片训练相机模糊检测模型,基于所述相机模糊检测模型,判断所述待检测照片是否模糊,进而基于所述待检测照片是否模糊的结果,判断待检测的相机是否为拍摄模糊相机,若所述待检测照片判断为模糊,则判定实施待检测的相机为拍摄模糊的问题相机,否则判定为拍摄清楚的无问题相机,从而可以准确、高效的判断待检测的相机拍摄性能

Description

相机模糊检测方法及装置 技术领域
本发明涉及计算机领域,尤其涉及一种相机模糊检测方法及装置。
背景技术
现有的手机等智能终端上的相机,多是通过人工进行模糊检测,检测过程费时、费力,并且有不准确的问题。
发明内容
本发明的一个目的是提供一种相机模糊检测方法及装置。
根据本发明的一个方面,提供了一种相机模糊检测方法,该方法包括:
获取对模板图像进行拍摄得到的预设数量的清晰模版照片和模糊模版照片;
基于所述预设数量的清晰模版照片和模糊模版照片训练相机模糊检测模型;
控制待检测的相机对模板图像进行拍摄得到待检测照片;
基于所述相机模糊检测模型,判断所述待检测照片是否模糊;
若所述待检测照片判断为模糊,则判定所述待检测的相机为拍摄模糊的问题相机。
进一步的,上述方法中,基于所述预设数量的清晰模版照片和模糊模版照片训练相机模糊检测模型,包括:
按预设大小和步长将每张清晰模版照片依序切分为多张小块清晰照片,其中,切分顺序上相邻的两张小块清晰照片有部分重叠区域;按预设大小和步长将每张模糊模版照片依序切分为多张小块模糊照片,其中,切分顺序上相邻的两张小块模糊照片有部分重叠区域;
获取所述预设大小的多张小块背景照片;
分别将代表清晰类别的多张小块清晰照片、代表模糊类别的多张小块模糊照片及代表背景类别的多张小块背景照片放入卷积神经网络模型进行训练,以得到所述相机模糊检测模型。
进一步的,上述方法中,基于所述相机模糊检测模型,判断所述待检测照片是否模糊,包括:
将所述待检测照片按预设大小和步长将每张待检测照片依序切分为多张小块待检测照片,其中,切分顺序上相邻的两张小块待检测照片有部分重叠区域;
将每张小块待检测照片输入所述相机模糊检测模型,以得到每张小块待检测照片分别相对于清晰类别、模糊类别和背景类别的概率;
若某张小块待检测照片相对于清晰类别的概率大于预设概率阈值,则判定该张小块待检测照片为模糊;
若判定为模糊的小块待检测照片的张数超过预设张数阈值,则判定所述待检测照片为模糊。
进一步的,上述方法中,所述模板图像包括在白底上按预设间隔布置的黑条纹,所述模板图像中的黑条纹为多排。
进一步的,上述方法中,所述模板图像中奇数排的黑条纹与水平方向的正向呈第一预设夹角,所述模板图像中偶数排的黑条纹与水平方向的正向呈第二预设夹角,其中,第二预设夹角=180度-第一预设夹角。
根据本发明的另一方面,还提供一种相机模糊检测装置,其中,该装置包括:
模版获取模块,用于获取对模板图像进行拍摄得到的预设数量的清晰模版照片和模糊模版照片;
训练模块,用于基于所述预设数量的清晰模版照片和模糊模版照片训 练相机模糊检测模型;
拍摄模块,用于控制待检测的相机对模板图像进行拍摄得到待检测照片;
判定模块,用于基于所述相机模糊检测模型,判断所述待检测照片是否模糊;若所述待检测照片判断为模糊,则判定所述待检测的相机为拍摄模糊的问题相机。
进一步的,上述装置中,所述训练模块,用于按预设大小和步长将每张清晰模版照片依序切分为多张小块清晰照片,其中,切分顺序上相邻的两张小块清晰照片有部分重叠区域;按预设大小和步长将每张模糊模版照片依序切分为多张小块模糊照片,其中,切分顺序上相邻的两张小块模糊照片有部分重叠区域;获取所述预设大小的多张小块背景照片;分别将代表清晰类别的多张小块清晰照片、代表模糊类别的多张小块模糊照片及代表背景类别的多张小块背景照片放入卷积神经网络模型进行训练,以得到所述相机模糊检测模型。
进一步的,上述装置中,所述判定模块,用于将所述待检测照片按预设大小和步长将每张待检测照片依序切分为多张小块待检测照片,其中,切分顺序上相邻的两张小块待检测照片有部分重叠区域;将每张小块待检测照片输入所述相机模糊检测模型,以得到每张小块待检测照片分别相对于清晰类别、模糊类别和背景类别的概率;若某张小块待检测照片相对于清晰类别的概率大于预设概率阈值,则判定该张小块待检测照片为模糊;若判定为模糊的小块待检测照片的张数超过预设张数阈值,则判定所述待检测照片为模糊。
进一步的,上述装置中,所述模板图像包括在白底上按预设间隔布置的黑条纹,所述模板图像中的黑条纹为多排。
进一步的,上述装置中,所述模板图像中奇数排的黑条纹与水平方向的正向呈第一预设夹角,所述模板图像中偶数排的黑条纹与水平方向的正 向呈第二预设夹角,其中,第二预设夹角=180度-第一预设夹角。
根据本发明的另一方面,还提供一种基于计算的设备,其中,包括:
处理器;以及
被安排成存储计算机可执行指令的存储器,所述可执行指令在被执行时使所述处理器:
获取对模板图像进行拍摄得到的预设数量的清晰模版照片和模糊模版照片;
基于所述预设数量的清晰模版照片和模糊模版照片训练相机模糊检测模型;
控制待检测的相机对模板图像进行拍摄得到待检测照片;
基于所述相机模糊检测模型,判断所述待检测照片是否模糊;
若所述待检测照片判断为模糊,则判定所述待检测的相机为拍摄模糊的问题相机。
根据本发明的另一方面,还提供一种计算机可读存储介质,其上存储有计算机可执行指令,其中,该计算机可执行指令被处理器执行时使得该处理器:
获取对模板图像进行拍摄得到的预设数量的清晰模版照片和模糊模版照片;
基于所述预设数量的清晰模版照片和模糊模版照片训练相机模糊检测模型;
控制待检测的相机对模板图像进行拍摄得到待检测照片;
基于所述相机模糊检测模型,判断所述待检测照片是否模糊;
若所述待检测照片判断为模糊,则判定所述待检测的相机为拍摄模糊的问题相机。
与现有技术相比,本发明基于所述预设数量的清晰模版照片和模糊模版照片训练相机模糊检测模型,基于所述相机模糊检测模型,判断所述待检测照片是否模糊,进而基于所述待检测照片是否模糊的结果,判断待检测的相机是否为拍摄模糊相机,若所述待检测照片判断为模糊,则判定实施待检测的相机为拍摄模糊的问题相机,否则判定为拍摄清楚的无问题相机,从而可以准确、高效的判断待检测的相机拍摄性能。
附图说明
通过阅读参照以下附图所作的对非限制性实施例所作的详细描述,本发明的其它特征、目的和优点将会变得更明显:
图1示出本发明一实施例的相机模糊检测方法的流程图;
图2示出本发明一实施例的模板图像示意图。
附图中相同或相似的附图标记代表相同或相似的部件。
具体实施方式
下面结合附图对本发明作进一步详细描述。
在本申请一个典型的配置中,终端、服务网络的设备和可信方均包括一个或多个处理器(CPU)、输入/输出接口、网络接口和内存。
内存可能包括计算机可读介质中的非永久性存储器,随机存取存储器(RAM)和/或非易失性内存等形式,如只读存储器(ROM)或闪存(flash RAM)。内存是计算机可读介质的示例。
计算机可读介质包括永久性和非永久性、可移动和非可移动媒体可以由任何方法或技术来实现信息存储。信息可以是计算机可读指令、数据结构、程序的模块或其他数据。计算机的存储介质的例子包括,但不限于相变内存(PRAM)、静态随机存取存储器(SRAM)、动态随机存取存储器(DRAM)、其他类型的随机存取存储器(RAM)、只读存储器(ROM)、电可擦 除可编程只读存储器(EEPROM)、快闪记忆体或其他内存技术、只读光盘只读存储器(CD-ROM)、数字多功能光盘(DVD)或其他光学存储、磁盒式磁带,磁带磁盘存储或其他磁性存储设备或任何其他非传输介质,可用于存储可以被计算设备访问的信息。按照本文中的界定,计算机可读介质不包括非暂存电脑可读媒体(transitory media),如调制的数据信号和载波。
如图1所示,本发明提供一种相机模糊检测方法,所述方法包括:
步骤S1,获取对模板图像进行拍摄得到的预设数量的清晰模版照片和模糊模版照片;
在此,清晰模版照片中所拍摄到的模板图像是完全清晰的,所述模糊模版照片中拍摄到的模板图像是全部或部分是模糊的;
可以在一背光源上设置静止显示的模板图像;
所述相机可以是手机、PAD等智能终端上的相机;
步骤S2,基于所述预设数量的清晰模版照片和模糊模版照片训练相机模糊检测模型;
步骤S3,控制待检测的相机对模板图像进行拍摄得到待检测照片;
在此,可以控制移动设备带动手机等终端移动到拍摄位置后,控制手机等终端保持静止,等待10秒后,控制手机等终端上的待检测的相机自动调焦和调光后,控制待检测的相机对模板图像进行拍摄得到一张待检测照片;
步骤S4,基于所述相机模糊检测模型,判断所述待检测照片是否模糊;
步骤S5,若所述待检测照片判断为模糊,则判定所述待检测的相机为拍摄模糊的问题相机。
在此,本发明基于所述预设数量的清晰模版照片和模糊模版照片训练相机模糊检测模型,基于所述相机模糊检测模型,判断所述待检测照片是否模糊,进而基于所述待检测照片是否模糊的结果,判断待检测的相机是 否为拍摄模糊相机,若所述待检测照片判断为模糊,则判定实施待检测的相机为拍摄模糊的问题相机,否则判定为拍摄清楚的无问题相机,从而可以准确、高效的判断待检测的相机拍摄性能。
本发明的相机模糊检测方法一实施例中,步骤S2,基于所述预设数量的清晰模版照片和模糊模版照片训练相机模糊检测模型,包括:
步骤S21,按预设大小和步长将每张清晰模版照片依序切分为多张小块清晰照片,其中,切分顺序上相邻的两张小块清晰照片有部分重叠区域;按预设大小和步长将每张模糊模版照片依序切分为多张小块模糊照片,其中,切分顺序上相邻的两张小块模糊照片有部分重叠区域;
在此,按预设大小和步长,可以切分到切分顺序上相邻的两张小块清晰照片有部分重叠区域,同时可以切分到切分顺序上相邻的两张小块模糊照片有部分重叠区域,便于后续模型的准确训练;
步骤S22,获取所述预设大小的多张小块背景照片;
步骤S23,分别将代表清晰类别的多张小块清晰照片、代表模糊类别的多张小块模糊照片及代表背景类别的多张小块背景照片放入卷积神经网络模型进行训练,以得到所述相机模糊检测模型。
在此,本实施例通过分别将代表清晰类别的多张小块清晰照片、代表模糊类别的多张小块模糊照片及代表背景类别的多张小块背景照片的三类照片放入卷积神经网络模型进行训练,可以高效、可靠的训练得到相机模糊检测模型。
本发明的相机模糊检测方法一实施例中,步骤S4,基于所述相机模糊检测模型,判断所述待检测照片是否模糊,包括:
步骤S41,将所述待检测照片按预设大小和步长将每张待检测照片依序切分为多张小块待检测照片,其中,切分顺序上相邻的两张小块待检测照片有部分重叠区域;
在此,所述待检测照片的切分方式与前面模型训练时的模版照片的切 分方式相同;
步骤S42,将每张小块待检测照片输入所述相机模糊检测模型,以得到每张小块待检测照片分别相对于清晰类别、模糊类别和背景类别的概率;
步骤S43,若某张小块待检测照片相对于清晰类别的概率大于预设概率阈值,则判定该张小块待检测照片为模糊;
步骤S44,若判定为模糊的小块待检测照片的张数超过预设张数阈值,则判定所述待检测照片为模糊。
在此,通过判定各张小块待检测照片相对于清晰类别的概率是否大于预设概率阈值,同时判定模糊的小块待检测照片的张数是否超过预设张数阈值,能够可靠、高效的判定所述待检测照片是否模糊。
本发明的相机模糊检测方法一实施例中,如图2所示,所述模板图像包括在白底上按预设间隔布置的黑条纹,所述模板图像中的黑条纹为多排。
在此,本发明通过将所述模板图像显示为在白底上按预设间隔布置的黑条纹,便于后续在训练模型和通过模型识别待检测照片的过程中,更易于识别各条黑条纹是否清晰或模糊。
如图2所示,通过在白底上设置多排的间隔分布的黑条纹,可以实现在拍摄照片时,让间隔分布的黑条纹尽量布满整个视场,以提高各广角相机拍摄的照片的效果。
本发明的相机模糊检测方法一实施例中,所述模板图像中奇数排的黑条纹与水平方向的正向呈第一预设夹角,所述模板图像中偶数排的黑条纹与水平方向的正向呈第二预设夹角,其中,第二预设夹角=180度-第一预设夹角。
在此,如图2所示,黑条纹与水平方向成预设夹角的斜纵向布置,便于后续对黑白条纹区域进行图像识别,提高识别效率和可靠度。
如图2所示,所述模板图像中第一排和第三排的黑条纹与水平方向的正向呈60度夹角,所述模板图像中第二排和第四排的黑条纹与水平方向的正向呈120度夹角,便于后续更高效的识别出条纹模糊的位置,可接通背光源的电源线1,为模板图像提供照明光源。
如图2所示,各排的黑条纹的间隔距离可以是10毫米,同一排上的黑条纹的间隔距离为8毫米,每个黑条纹的宽度为8毫米。
本发明提供一种相机模糊检测装置,所述装置包括:
模版获取模块,用于获取对模板图像进行拍摄得到的预设数量的清晰模版照片和模糊模版照片;
在此,清晰模版照片中所拍摄到的模板图像是完全清晰的,所述模糊模版照片中拍摄到的模板图像是全部或部分是模糊的;
可以在一背光源上设置静止显示的模板图像;
所述相机可以是手机、PAD等智能终端上的相机;
训练模块,用于基于所述预设数量的清晰模版照片和模糊模版照片训练相机模糊检测模型;
拍摄模块,用于控制待检测的相机对模板图像进行拍摄得到待检测照片;
在此,可以控制移动设备带动手机等终端移动到拍摄位置后,控制手机等终端保持静止,等待10秒后,控制手机等终端上的待检测的相机自动调焦和调光后,控制待检测的相机对模板图像进行拍摄得到一张待检测照片;
判定模块,用于基于所述相机模糊检测模型,判断所述待检测照片是否模糊;若所述待检测照片判断为模糊,则判定所述待检测的相机为拍摄模糊的问题相机。
在此,本发明基于所述预设数量的清晰模版照片和模糊模版照片训练 相机模糊检测模型,基于所述相机模糊检测模型,判断所述待检测照片是否模糊,进而基于所述待检测照片是否模糊的结果,判断待检测的相机是否为拍摄模糊相机,若所述待检测照片判断为模糊,则判定实施待检测的相机为拍摄模糊的问题相机,否则判定为拍摄清楚的无问题相机,从而可以准确、高效的判断待检测的相机拍摄性能。
本发明的相机模糊检测装置一实施例中,所述训练模块,用于按预设大小和步长将每张清晰模版照片依序切分为多张小块清晰照片,其中,切分顺序上相邻的两张小块清晰照片有部分重叠区域;按预设大小和步长将每张模糊模版照片依序切分为多张小块模糊照片,其中,切分顺序上相邻的两张小块模糊照片有部分重叠区域;获取所述预设大小的多张小块背景照片;分别将代表清晰类别的多张小块清晰照片、代表模糊类别的多张小块模糊照片及代表背景类别的多张小块背景照片放入卷积神经网络模型进行训练,以得到所述相机模糊检测模型。
在此,按预设大小和步长,可以切分到切分顺序上相邻的两张小块清晰照片有部分重叠区域,同时可以切分到切分顺序上相邻的两张小块模糊照片有部分重叠区域,便于后续模型的准确训练;
本实施例通过分别将代表清晰类别的多张小块清晰照片、代表模糊类别的多张小块模糊照片及代表背景类别的多张小块背景照片的三类照片放入卷积神经网络模型进行训练,可以高效、可靠的训练得到相机模糊检测模型。
本发明的相机模糊检测方法一实施例中,所述判定模块,用于将所述待检测照片按预设大小和步长将每张待检测照片依序切分为多张小块待检测照片,其中,切分顺序上相邻的两张小块待检测照片有部分重叠区域;将每张小块待检测照片输入所述相机模糊检测模型,以得到每张小块待检测照片分别相对于清晰类别、模糊类别和背景类别的概率;若某张小块待检测照片相对于清晰类别的概率大于预设概率阈值,则判定该张小块待检 测照片为模糊;若判定为模糊的小块待检测照片的张数超过预设张数阈值,则判定所述待检测照片为模糊。
在此,所述待检测照片的切分方式与前面模型训练时的模版照片的切分方式相同;
在此,通过判定各张小块待检测照片相对于清晰类别的概率是否大于预设概率阈值,同时判定模糊的小块待检测照片的张数是否超过预设张数阈值,能够可靠、高效的判定所述待检测照片是否模糊。
本发明的相机模糊检测方法一实施例中,如图2所示,所述模板图像包括在白底上按预设间隔布置的黑条纹,所述模板图像中的黑条纹为多排。
在此,本发明通过将所述模板图像显示为在白底上按预设间隔布置的黑条纹,便于后续在训练模型和通过模型识别待检测照片的过程中,更易于识别各条黑条纹是否清晰或模糊。
如图2所示,通过在白底上设置多排的间隔分布的黑条纹,可以实现在拍摄照片时,让间隔分布的黑条纹尽量布满整个视场,以提高各广角相机拍摄的照片的效果。
本发明的相机模糊检测方法一实施例中,所述模板图像中奇数排的黑条纹与水平方向的正向呈第一预设夹角,所述模板图像中偶数排的黑条纹与水平方向的正向呈第二预设夹角,其中,第二预设夹角=180度-第一预设夹角。
在此,如图2所示,黑条纹与水平方向成预设夹角的斜纵向布置,便于后续对黑白条纹区域进行图像识别,提高识别效率和可靠度。
如图2所示,所述模板图像中第一排和第三排的黑条纹与水平方向的正向呈60度夹角,所述模板图像中第二排和第四排的黑条纹与水平方向的正向呈120度夹角,便于后续更高效的识别出条纹模糊的位置,可接通背光源的电源线1,为模板图像提供照明光源。
如图2所示,各排的黑条纹的间隔距离可以是10毫米,同一排上的黑条纹的间隔距离为8毫米,每个黑条纹的宽度为8毫米。
根据本发明的另一方面,还提供一种基于计算的设备,其中,包括:
处理器;以及
被安排成存储计算机可执行指令的存储器,所述可执行指令在被执行时使所述处理器:
获取对模板图像进行拍摄得到的预设数量的清晰模版照片和模糊模版照片;
基于所述预设数量的清晰模版照片和模糊模版照片训练相机模糊检测模型;
控制待检测的相机对模板图像进行拍摄得到待检测照片;
基于所述相机模糊检测模型,判断所述待检测照片是否模糊;
若所述待检测照片判断为模糊,则判定所述待检测的相机为拍摄模糊的问题相机。
根据本发明的另一方面,还提供一种计算机可读存储介质,其上存储有计算机可执行指令,其中,该计算机可执行指令被处理器执行时使得该处理器:
获取对模板图像进行拍摄得到的预设数量的清晰模版照片和模糊模版照片;
基于所述预设数量的清晰模版照片和模糊模版照片训练相机模糊检测模型;
控制待检测的相机对模板图像进行拍摄得到待检测照片;
基于所述相机模糊检测模型,判断所述待检测照片是否模糊;
若所述待检测照片判断为模糊,则判定所述待检测的相机为拍摄模糊 的问题相机。
本发明的各设备和存储介质实施例的详细内容,具体可参见各方法实施例的对应部分,在此,不再赘述。
显然,本领域的技术人员可以对本申请进行各种改动和变型而不脱离本申请的精神和范围。这样,倘若本申请的这些修改和变型属于本申请权利要求及其等同技术的范围之内,则本申请也意图包含这些改动和变型在内。
需要注意的是,本发明可在软件和/或软件与硬件的组合体中被实施,例如,可采用专用集成电路(ASIC)、通用目的计算机或任何其他类似硬件设备来实现。在一个实施例中,本发明的软件程序可以通过处理器执行以实现上文所述步骤或功能。同样地,本发明的软件程序(包括相关的数据结构)可以被存储到计算机可读记录介质中,例如,RAM存储器,磁或光驱动器或软磁盘及类似设备。另外,本发明的一些步骤或功能可采用硬件来实现,例如,作为与处理器配合从而执行各个步骤或功能的电路。
另外,本发明的一部分可被应用为计算机程序产品,例如计算机程序指令,当其被计算机执行时,通过该计算机的操作,可以调用或提供根据本发明的方法和/或技术方案。而调用本发明的方法的程序指令,可能被存储在固定的或可移动的记录介质中,和/或通过广播或其他信号承载媒体中的数据流而被传输,和/或被存储在根据所述程序指令运行的计算机设备的工作存储器中。在此,根据本发明的一个实施例包括一个装置,该装置包括用于存储计算机程序指令的存储器和用于执行程序指令的处理器,其中,当该计算机程序指令被该处理器执行时,触发该装置运行基于前述根据本发明的多个实施例的方法和/或技术方案。
对于本领域技术人员而言,显然本发明不限于上述示范性实施例的细节,而且在不背离本发明的精神或基本特征的情况下,能够以其他的具体形式实现本发明。因此,无论从哪一点来看,均应将实施例看作是示范性 的,而且是非限制性的,本发明的范围由所附权利要求而不是上述说明限定,因此旨在将落在权利要求的等同要件的含义和范围内的所有变化涵括在本发明内。不应将权利要求中的任何附图标记视为限制所涉及的权利要求。此外,显然“包括”一词不排除其他单元或步骤,单数不排除复数。装置权利要求中陈述的多个单元或装置也可以由一个单元或装置通过软件或者硬件来实现。第一,第二等词语用来表示名称,而并不表示任何特定的顺序。

Claims (12)

  1. 一种相机模糊检测方法,其中,该方法包括:
    获取对模板图像进行拍摄得到的预设数量的清晰模版照片和模糊模版照片;
    基于所述预设数量的清晰模版照片和模糊模版照片训练相机模糊检测模型;
    控制待检测的相机对模板图像进行拍摄得到待检测照片;
    基于所述相机模糊检测模型,判断所述待检测照片是否模糊;
    若所述待检测照片判断为模糊,则判定所述待检测的相机为拍摄模糊的问题相机。
  2. 根据权利要求1所述的方法,其中,基于所述预设数量的清晰模版照片和模糊模版照片训练相机模糊检测模型,包括:
    按预设大小和步长将每张清晰模版照片依序切分为多张小块清晰照片,其中,切分顺序上相邻的两张小块清晰照片有部分重叠区域;按预设大小和步长将每张模糊模版照片依序切分为多张小块模糊照片,其中,切分顺序上相邻的两张小块模糊照片有部分重叠区域;
    获取所述预设大小的多张小块背景照片;
    分别将代表清晰类别的多张小块清晰照片、代表模糊类别的多张小块模糊照片及代表背景类别的多张小块背景照片放入卷积神经网络模型进行训练,以得到所述相机模糊检测模型。
  3. 根据权利要求2所述的方法,其中,基于所述相机模糊检测模型,判断所述待检测照片是否模糊,包括:
    将所述待检测照片按预设大小和步长将每张待检测照片依序切分为 多张小块待检测照片,其中,切分顺序上相邻的两张小块待检测照片有部分重叠区域;
    将每张小块待检测照片输入所述相机模糊检测模型,以得到每张小块待检测照片分别相对于清晰类别、模糊类别和背景类别的概率;
    若某张小块待检测照片相对于清晰类别的概率大于预设概率阈值,则判定该张小块待检测照片为模糊;
    若判定为模糊的小块待检测照片的张数超过预设张数阈值,则判定所述待检测照片为模糊。
  4. 根据权利要求1所述的方法,其中,所述模板图像包括在白底上按预设间隔布置的黑条纹,所述模板图像中的黑条纹为多排。
  5. 根据权利要求4所述的方法,其中,所述模板图像中奇数排的黑条纹与水平方向的正向呈第一预设夹角,所述模板图像中偶数排的黑条纹与水平方向的正向呈第二预设夹角,其中,第二预设夹角=180度-第一预设夹角。
  6. 一种相机模糊检测装置,其中,该装置包括:
    模版获取模块,用于获取对模板图像进行拍摄得到的预设数量的清晰模版照片和模糊模版照片;
    训练模块,用于基于所述预设数量的清晰模版照片和模糊模版照片训练相机模糊检测模型;
    拍摄模块,用于控制待检测的相机对模板图像进行拍摄得到待检测照片;
    判定模块,用于基于所述相机模糊检测模型,判断所述待检测照片是 否模糊;若所述待检测照片判断为模糊,则判定所述待检测的相机为拍摄模糊的问题相机。
  7. 根据权利要求6所述的装置,其中,所述训练模块,用于按预设大小和步长将每张清晰模版照片依序切分为多张小块清晰照片,其中,切分顺序上相邻的两张小块清晰照片有部分重叠区域;按预设大小和步长将每张模糊模版照片依序切分为多张小块模糊照片,其中,切分顺序上相邻的两张小块模糊照片有部分重叠区域;获取所述预设大小的多张小块背景照片;分别将代表清晰类别的多张小块清晰照片、代表模糊类别的多张小块模糊照片及代表背景类别的多张小块背景照片放入卷积神经网络模型进行训练,以得到所述相机模糊检测模型。
  8. 根据权利要求7所述的装置,其中,所述判定模块,用于将所述待检测照片按预设大小和步长将每张待检测照片依序切分为多张小块待检测照片,其中,切分顺序上相邻的两张小块待检测照片有部分重叠区域;将每张小块待检测照片输入所述相机模糊检测模型,以得到每张小块待检测照片分别相对于清晰类别、模糊类别和背景类别的概率;若某张小块待检测照片相对于清晰类别的概率大于预设概率阈值,则判定该张小块待检测照片为模糊;若判定为模糊的小块待检测照片的张数超过预设张数阈值,则判定所述待检测照片为模糊。
  9. 根据权利要求6所述的装置,其中,所述模板图像包括在白底上按预设间隔布置的黑条纹,所述模板图像中的黑条纹为多排。
  10. 根据权利要求9所述的装置,其中,所述模板图像中奇数排的黑条纹与水平方向的正向呈第一预设夹角,所述模板图像中偶数排的黑条纹 与水平方向的正向呈第二预设夹角,其中,第二预设夹角=180度-第一预设夹角。
  11. 一种基于计算的设备,其中,包括:
    处理器;以及
    被安排成存储计算机可执行指令的存储器,所述可执行指令在被执行时使所述处理器:
    获取对模板图像进行拍摄得到的预设数量的清晰模版照片和模糊模版照片;
    基于所述预设数量的清晰模版照片和模糊模版照片训练相机模糊检测模型;
    控制待检测的相机对模板图像进行拍摄得到待检测照片;
    基于所述相机模糊检测模型,判断所述待检测照片是否模糊;
    若所述待检测照片判断为模糊,则判定所述待检测的相机为拍摄模糊的问题相机。
  12. 一种计算机可读存储介质,其上存储有计算机可执行指令,其中,该计算机可执行指令被处理器执行时使得该处理器:
    获取对模板图像进行拍摄得到的预设数量的清晰模版照片和模糊模版照片;
    基于所述预设数量的清晰模版照片和模糊模版照片训练相机模糊检测模型;
    控制待检测的相机对模板图像进行拍摄得到待检测照片;
    基于所述相机模糊检测模型,判断所述待检测照片是否模糊;
    若所述待检测照片判断为模糊,则判定所述待检测的相机为拍摄模糊的问题相机。
PCT/CN2020/120886 2020-01-21 2020-10-14 相机模糊检测方法及装置 WO2021147383A1 (zh)

Priority Applications (1)

Application Number Priority Date Filing Date Title
JP2022506070A JP2022542948A (ja) 2020-01-21 2020-10-14 カメラぼやけ検査方法及び装置

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN202010072735.0 2020-01-21
CN202010072735.0A CN111246203A (zh) 2020-01-21 2020-01-21 相机模糊检测方法及装置

Publications (1)

Publication Number Publication Date
WO2021147383A1 true WO2021147383A1 (zh) 2021-07-29

Family

ID=70872961

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2020/120886 WO2021147383A1 (zh) 2020-01-21 2020-10-14 相机模糊检测方法及装置

Country Status (3)

Country Link
JP (1) JP2022542948A (zh)
CN (1) CN111246203A (zh)
WO (1) WO2021147383A1 (zh)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN116028258A (zh) * 2023-03-22 2023-04-28 国网智联电商有限公司 基于深度学习的设备故障诊断方法及系统

Families Citing this family (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111246203A (zh) * 2020-01-21 2020-06-05 上海悦易网络信息技术有限公司 相机模糊检测方法及装置

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US7457458B1 (en) * 1999-11-26 2008-11-25 Inb Vision Ag. Method and apparatus for defining and correcting image data
CN106920229A (zh) * 2017-01-22 2017-07-04 北京奇艺世纪科技有限公司 图像模糊区域自动检测方法及系统
CN109167998A (zh) * 2018-11-19 2019-01-08 深兰科技(上海)有限公司 检测相机状态的方法及装置、电子设备、存储介质
CN109886359A (zh) * 2019-03-25 2019-06-14 西安电子科技大学 基于卷积神经网络的小目标检测方法及检测模型
CN111246203A (zh) * 2020-01-21 2020-06-05 上海悦易网络信息技术有限公司 相机模糊检测方法及装置

Family Cites Families (9)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN100525475C (zh) * 2005-12-28 2009-08-05 华晶科技股份有限公司 影像清晰检测方法
JP2012247743A (ja) * 2011-05-31 2012-12-13 Konica Minolta Advanced Layers Inc 解像度検査用チャート及び解像度の検査方法
CN104780310A (zh) * 2015-01-27 2015-07-15 深圳市中瀛鑫科技股份有限公司 一种图像模糊的检测方法、系统及摄像机
CN104899255B (zh) * 2015-05-15 2018-06-26 浙江大学 适用于训练深度卷积神经网络的图像数据库的构建方法
CN105160678A (zh) * 2015-09-02 2015-12-16 山东大学 基于卷积神经网络的无参考立体图像质量评估方法
CN106096605B (zh) * 2016-06-02 2019-03-19 史方 一种基于深度学习的图像模糊区域检测方法及装置
JP6849824B2 (ja) * 2017-05-31 2021-03-31 ザ プロクター アンド ギャンブル カンパニーThe Procter & Gamble Company セルフィーを撮影するためにユーザをガイドするためのシステム及び方法
CN107644209A (zh) * 2017-09-21 2018-01-30 百度在线网络技术(北京)有限公司 人脸检测方法和装置
CN107948510B (zh) * 2017-11-27 2020-04-07 北京小米移动软件有限公司 焦距调整的方法、装置及存储介质

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US7457458B1 (en) * 1999-11-26 2008-11-25 Inb Vision Ag. Method and apparatus for defining and correcting image data
CN106920229A (zh) * 2017-01-22 2017-07-04 北京奇艺世纪科技有限公司 图像模糊区域自动检测方法及系统
CN109167998A (zh) * 2018-11-19 2019-01-08 深兰科技(上海)有限公司 检测相机状态的方法及装置、电子设备、存储介质
CN109886359A (zh) * 2019-03-25 2019-06-14 西安电子科技大学 基于卷积神经网络的小目标检测方法及检测模型
CN111246203A (zh) * 2020-01-21 2020-06-05 上海悦易网络信息技术有限公司 相机模糊检测方法及装置

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN116028258A (zh) * 2023-03-22 2023-04-28 国网智联电商有限公司 基于深度学习的设备故障诊断方法及系统
CN116028258B (zh) * 2023-03-22 2023-08-18 国网智联电商有限公司 基于深度学习的设备故障诊断方法及系统

Also Published As

Publication number Publication date
JP2022542948A (ja) 2022-10-07
CN111246203A (zh) 2020-06-05

Similar Documents

Publication Publication Date Title
WO2021147385A1 (zh) 屏幕检测方法及装置
US10984293B2 (en) Image processing method and apparatus
US11138478B2 (en) Method and apparatus for training, classification model, mobile terminal, and readable storage medium
US20200160040A1 (en) Three-dimensional living-body face detection method, face authentication recognition method, and apparatuses
WO2021147387A1 (zh) 屏幕划痕碎裂检测方法及设备
WO2021147386A1 (zh) 屏幕划痕碎裂检测方法及设备
CN108229480A (zh) 一种数字表读数的识别方法、装置及设备
WO2021147383A1 (zh) 相机模糊检测方法及装置
WO2021147388A1 (zh) 相机色斑检测方法及设备
WO2021147384A1 (zh) 相机抖动检测方法及装置
WO2021082922A1 (zh) 一种检测屏幕显示断线的方法及设备
US8488004B2 (en) System and method for identifying discrepancy of image of object
JP5141245B2 (ja) 画像処理装置、補正情報生成方法、および撮像装置
TWI482468B (zh) 物體偵測裝置、方法及其電腦可讀取紀錄媒體
CN110969045B (zh) 一种行为检测方法、装置、电子设备和存储介质
US20220270266A1 (en) Foreground image acquisition method, foreground image acquisition apparatus, and electronic device
CN110705532A (zh) 一种识别翻拍图像的方法、装置及设备
CN111325717A (zh) 手机缺陷位置识别方法及设备
CN109068060A (zh) 图像处理方法和装置、终端设备、计算机可读存储介质
US8547430B2 (en) System and method for marking discrepancies in image of object
CN109102026A (zh) 一种车辆图像检测方法、装置及系统
CN111272388A (zh) 一种相机闪光灯的检测方法及设备
CN112287905A (zh) 车辆损伤识别方法、装置、设备及存储介质
US11900589B2 (en) Detection device of display panel and detection method thereof, electronic device and readable medium
CN115147422A (zh) 一种晶圆中心晶粒生成方法、装置、设备及存储介质

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: 20914988

Country of ref document: EP

Kind code of ref document: A1

ENP Entry into the national phase

Ref document number: 2022506070

Country of ref document: JP

Kind code of ref document: A

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 20914988

Country of ref document: EP

Kind code of ref document: A1