WO2019114027A1 - 一种图像处理方法、存储介质及智能终端 - Google Patents

一种图像处理方法、存储介质及智能终端 Download PDF

Info

Publication number
WO2019114027A1
WO2019114027A1 PCT/CN2017/118026 CN2017118026W WO2019114027A1 WO 2019114027 A1 WO2019114027 A1 WO 2019114027A1 CN 2017118026 W CN2017118026 W CN 2017118026W WO 2019114027 A1 WO2019114027 A1 WO 2019114027A1
Authority
WO
WIPO (PCT)
Prior art keywords
image
mapping model
optimized
standard
processing method
Prior art date
Application number
PCT/CN2017/118026
Other languages
English (en)
French (fr)
Inventor
周永进
金英健
Original Assignee
深圳大学
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 深圳大学 filed Critical 深圳大学
Publication of WO2019114027A1 publication Critical patent/WO2019114027A1/zh

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T5/00—Image enhancement or restoration
    • G06T5/73—Deblurring; Sharpening
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/20—Special algorithmic details
    • G06T2207/20081—Training; Learning
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/30—Subject of image; Context of image processing
    • G06T2207/30004—Biomedical image processing
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/30—Subject of image; Context of image processing
    • G06T2207/30004—Biomedical image processing
    • G06T2207/30008—Bone

Definitions

  • the present invention relates to the field of communications technologies, and in particular, to an image processing method, a storage medium, and an intelligent terminal.
  • smart terminals With the rapid spread of smart terminals, smart terminals have been widely used in various industries, bringing too much convenience to people's work and life.
  • smart terminals basically have imaging functions, ranging from mobile phones or cameras that are available to everyone, to instruments and equipment used in special fields, such as X-ray equipment in medical fields, ultrasound imaging equipment, and the like.
  • the quality of imaging is also affected by many factors. For example, if the user uses a mobile phone to take pictures at night, if the exposure rate is low, the captured image is too dark to recognize the target information; if the exposure rate is high, the brightness of the image reaches a certain standard, not only the autofocus time is prolonged, but the edge of the scene is blurred. The details of the information are greatly reduced, which seriously affects the image quality and brings inconvenience to the user.
  • the above problems can be improved by replacing the smart terminal, for the smart terminal with high image quality, there is generally a defect that the device cost is high, the volume is large, and the imaging speed is slow.
  • the technical problem to be solved by the present invention is to provide an image processing method, a storage medium, and an intelligent terminal, aiming at solving the problem of poor image quality of some smart terminals in the prior art.
  • An image processing method wherein the method comprises:
  • the intelligent terminal collects image data of a certain thing and obtains an original image of the object
  • the intelligent terminal displays the output optimized image.
  • the image processing method wherein the original image represents an image to be optimized
  • the optimized image represents an image whose image quality is higher than the original image and satisfies the user's use requirements.
  • the image processing method wherein the smart terminal performs image data collection on a certain object, and before obtaining the original image of the object, the method further includes:
  • a mapping model for reconstructing and optimizing an image is created or input in advance in the smart terminal.
  • mapping model for creating and inputting an image for reconstruction and optimization in advance in the smart terminal specifically includes:
  • the non-standard image represents an image with obvious defect in image quality
  • the standard image represents the standard
  • the image represents an image with a significantly higher image quality than a non-standard image
  • mapping model having a one-to-one mapping relationship between the non-standard image and the standard image content is established.
  • mapping model has a parameter initialization function
  • the parameter initialization function is started or the image is re-learned and trained offline, and automatically acquired. Mapping the relationship parameters, obtaining a new mapping model, or imaging different requirements according to different application fields, artificially adding some parameters in the network.
  • the image processing method wherein the calling a preset mapping model for optimizing an image, optimizing an original image according to a mapping relationship; obtaining an optimized image of the thing, and saving the optimized image Or the output specifically includes:
  • the image optimization function is automatically activated
  • mapping model Calling a preset mapping model to input the original image to a mapping model
  • the original image is processed into an optimized image according to the mapping relationship in the mapping model, and the optimized image is stored or output to the display module of the smart terminal.
  • a storage medium having stored thereon a plurality of instructions, wherein the instructions are adapted to be loaded and executed by a processor to implement the image processing method of any of the above.
  • An intelligent terminal comprising: a processor, a storage medium communicatively coupled to the processor, the storage medium being adapted to store a plurality of instructions; the processor being adapted to invoke an instruction in the storage medium to perform an implementation
  • the image processing method according to any of the above.
  • the present invention can automatically optimize the original image by automatically calling the mapping model after the smart terminal obtains the original image of the object by creating a mapping model for optimizing the image, especially for some image quality.
  • Poor low-end equipment without changing the hardware, can only achieve the user's use requirements by changing the image quality algorithm, effectively improving the image quality, and the whole optimization process is faster and the time required is shorter.
  • the degree of imaging speed is increased and the life of the device is extended.
  • FIG. 1 is a flow chart of a preferred embodiment of the image processing method of the present invention.
  • FIG. 2 is a schematic diagram of an original image to be optimized in the present invention.
  • FIG 3 is a schematic diagram of a standard image obtained after the original image of the present invention passes through a mapping model.
  • FIG. 4 is a functional block diagram of a preferred embodiment of the intelligent terminal of the present invention.
  • FIG. 1 is a flowchart of a preferred embodiment of the image processing method of the present invention.
  • the image processing method includes the following steps:
  • Step S100 The intelligent terminal performs image data collection on a certain object to obtain an original image of the object.
  • the present invention implements a mapping model for optimizing an image, and optimizes an original image that has not undergone any optimization processing according to a mapping relationship to meet user usage requirements. Therefore, before the above step S100, the present invention needs to create or input a mapping model for reconstructing and optimizing an image in advance in the smart terminal.
  • the present invention can use an intelligent terminal with an image optimization function to first perform image data collection on a certain thing or scene in a state where the image optimization function is turned off, obtain a non-standard image of the thing or scene, and then turn on.
  • image data is collected for the same thing or scene, and a standard image of the thing or scene is obtained, and non-standard images and standard images of the same thing or scene are obtained through the above operations.
  • the method for acquiring non-standard images and standard images of the same thing or scene can also perform image data collection on the same thing or the same scene by using two intelligent terminals with significant differences in performance, so that the performance is high.
  • the smart terminal obtains a standard image
  • the low-performance smart terminal obtains a non-standard image. Its main purpose is to create the mapping model.
  • the present invention can selectively perform bicubic interpolation preprocessing on the non-standard image, enlarge the non-standard image to the target object size, and then Standard images are tagged with standard images and entered into a network for learning and training for learning and training.
  • learning and training the detailed features of standard images and non-standard images are extracted, and the one-to-one mapping relationship between non-standard images and standard image content is established, and then the mapping model with the mapping relationship is created. It is also possible to add some parameters to the network for human intervention in response to imaging requirements in different fields.
  • the acquisition mode and the program of the mapping model can be embedded in any intelligent terminal having an imaging system, so that some intelligent terminals with poor imaging effects can automatically optimize the image using the mapping model without replacing the smart terminal, saving The cost has brought convenience to the user.
  • the non-standard image obtained in the process of creating the mapping model represents an image with obvious image quality defects, that is, the user can obviously perceive the image with unclear blur
  • the standard image indicates The standard image represents an image with a significantly higher image quality than a non-standard image. Whether it is through the user's visual observation of the sensory effect, or the resolution, signal-to-noise ratio and other parameters analyzed by the instrument, there are significant improvements. Therefore, the mapping model created by the present invention is essentially transforming a low quality image into a high quality image.
  • the mapping model for optimizing the image is created, the acquiring manner and the program of the mapping model or directly embedding the mapping model into the smart terminal, so that the smart terminal has an image optimization function .
  • the object of the present invention is to optimize the original image to be optimized by the smart terminal into a standard image when the mapping model is created. Optimized images with the same or similar imaging effects to improve the imaging quality of smart terminals.
  • Step S200 Calling a preset mapping model for optimizing an image, and performing optimization processing on the original image according to the mapping relationship; obtaining an optimized image of the thing, and outputting the optimized image.
  • the step S200 specifically includes:
  • Step S201 After the imaging system of the smart terminal obtains the original image of the object, the image optimization function is automatically started;
  • Step S202 calling a preset mapping model, and inputting the original image into the mapping model;
  • Step S203 Processing the original image into an optimized image according to the mapping relationship in the mapping model, and storing or outputting the optimized image to the display module of the smart terminal.
  • the smart terminal when the smart terminal obtains the original image to be optimized, the smart terminal automatically starts the image optimization function, and then automatically invokes the preset mapping model.
  • the mapping model preprocesses the original image, inputs the original image to the mapping model, and automatically optimizes the original image into high quality according to the mapping relationship established in the mapping model to convert the low quality image into a high quality image.
  • the optimized image does not require manual operation by the user, and the entire optimization process is fast and efficient.
  • the smart terminal stores or outputs the optimized image to the display module of the smart terminal to display the optimized image. According to the image optimization disclosed above, the present invention can automatically call the mapping model to optimize the original image, thereby satisfying the user's use requirements and performing image processing with the conventional using algorithm.
  • the present invention is applicable to some low-end devices with poor imaging quality. It can have obvious improvement of imaging quality, and the whole optimization process is faster and the time required is shorter, which improves the imaging speed of the intelligent terminal to some extent, and reduces the subjective and erroneous image enhancement mode.
  • the parameters in the mapping model created by the present invention can be autonomously set and adjusted according to the usage requirements, that is, have a parameter initialization function.
  • the parameter initialization function is started, the image is re-learned and trained, the mapping relationship parameters are automatically acquired, and a new mapping model is established.
  • the present invention can also input non-standard images and standard image lines to re-establish the mapping model. For example, if a smart terminal with the mapping model is applied to the medical field, the user needs a higher imaging effect when using the smart terminal to observe the bone tissue or other organizational structure, and only needs to acquire the bone tissue or other tissue.
  • the non-standard image of the structure and the input of the standard image line can automatically establish a mapping model applicable to the medical field to meet the user's use requirements, and then input it into the intelligent terminal, thereby increasing the application range of the mapping model of the present invention.
  • Step S300 The smart terminal displays the output standard image.
  • FIG. 2 is a schematic diagram of an original image to be optimized in the present invention.
  • 3 is a schematic diagram of an optimized image obtained after the original image of the present invention passes through a mapping model.
  • the original image in Fig. 2 is not subjected to any optimization processing, and is an image to be processed.
  • the line stripe path in Fig. 2 has obvious blur, and the peak of the original image obtained by analysis
  • the sensory effect directly observed from the naked eye is significantly clearer in FIG. 3 than in FIG.
  • the peak signal to noise ratio (PSNR) of the standard image obtained by the analysis is 26.3647.
  • PSNR peak signal to noise ratio
  • the present invention also discloses an intelligent terminal.
  • the processor 10 includes a storage medium (memory) 20 connected to the processor 10, wherein the processor 10 is configured to invoke program instructions in the storage medium 20 to execute
  • the method provided by the above embodiment for example, performs:
  • Step S100 The smart terminal performs image data collection on a certain object to obtain an original image of the object;
  • Step S200 Calling a preset mapping model for optimizing an image to perform optimization processing on the original image; obtaining an optimized image of the thing, and saving or outputting the optimized image;
  • Step S300 The smart terminal displays the output optimized image.
  • the intelligent terminal in the present invention includes an instrument device of any imaging principle, and is applicable to any technical field and scene requiring imaging.
  • Embodiments of the present invention also provide a storage medium on which computer instructions are stored, the computer instructions causing a computer to perform the methods provided by the above embodiments.
  • the present invention provides an image processing method, a storage medium, and an intelligent terminal.
  • the method includes: the intelligent terminal performs image data collection on a certain object to obtain an original image of the object; and invokes a preset mapping for optimizing the image.
  • the model optimizes the original image according to the mapping relationship; obtains the optimized image of the thing, and saves or outputs the optimized image; the intelligent terminal displays the output optimized image.
  • the invention creates a mapping model for optimizing an image, and can automatically call the mapping model to optimize the original image after the smart terminal obtains the original image of the object, thereby improving the image quality, the imaging rate, and prolonging the service life of the device.
  • the user's use requirements can be achieved only by changing the imaging quality algorithm without changing the hardware, which not only improves the imaging quality, but also improves the imaging to some extent. speed.

Landscapes

  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Image Processing (AREA)

Abstract

本发明公开了一种图像处理方法、存储介质及智能终端，方法包括：智能终端对某事物进行影像数据采集，获得事物的原始图像；调用预设的用于优化图像的映射模型，根据映射关系对原始图像进行优化处理；获得事物的优化后图像，并将所述优化后图像保存或输出；智能终端将输出的优化图像进行显示。本发明通过创建用于对图像进行优化处理的映射模型，能够在智能终端获得某事物的原始图像之后自动调用映射模型对原始图像进行优化处理，提高图像质量、成像速率，延长设备的使用寿命，特别是对一些成像质量较差的低端设备，在不更改硬件的前提下，仅通过更改成像质量的算法，就可达到用户的使用需求，不但改善了成像质量，在一定程度上提高了成像速度。

Description

一种图像处理方法、存储介质及智能终端 技术领域
本发明涉及通信技术领域，具体涉及一种图像处理方法、存储介质及智能终端。
背景技术
随着智能终端的快速普及，智能终端已经广泛使用在各行各业，给人们的工作、生活带来了太多的便利。目前的智能终端基本都具有成像功能，小到人人都具有的手机或者相机，大到运用在特殊领域的仪器设备，例如在医学领域的X光设备、超声成像设备等等。
但是现有技术中的很多智能终端成像效果上不是很理想，尤其是一些低端设备，成像效果很不理想。例如日常使用的一些智能终端，成像质量的高低也是受到多方面因素的影响。例如用户使用手机在夜间拍照，如果曝光率较低，拍出的图像过暗，无法辨认目标信息；如果曝光率较高，图像亮度达到一定标准，不仅自动对焦时间延长，其中景物的边缘存在模糊，其细节信息大量减少，严重影响了成像质量，给用户的使用带来了不便。虽然可以通过更换智能终端来改善上述存在的问题，但是对于成像质量高的智能终端来说，普遍具有设备成本高，体积较大，且成像速度较慢的缺陷。
因此，现有技术还有待于改进和发展。
发明内容
本发明要解决的技术问题在于，针对现有技术的上述缺陷，提供一种图像处理方法、存储介质及智能终端，旨在解决现有技术中的一些智能终端的成像质量差的问题。
本发明解决技术问题所采用的技术方案如下：
一种图像处理方法，其中，所述方法包括：
智能终端对某事物进行影像数据采集，获得事物的原始图像；
调用预设的用于优化图像的映射模型，根据映射关系对原始图像进行优化处理；获得所述事物的优化后图像，并将所述优化后图像保存或输出；
智能终端将输出的优化后图像进行显示。
所述的图像处理方法，其中，所述原始图像表示的是待优化处理的图像；
所述优化后图像表示的是图像质量高于所述原始图像，且满足用户使用需求的图像。
所述的图像处理方法，其中，所述智能终端对某事物进行影像数据采集，获得事物的原始图像之前还包括：
预先在智能终端中创建或输入用于对图像进行重建与优化的映射模型。
所述的图像处理方法，其中，所述预先在智能终端中创建或输入用于对图像进行重建与优化的映射模型具体包括：
对同一事物或场景进行影像数据采集，分别获得该事物或场景的非标准图像与标准图像；所述非标准图像表示的是图像质量存在明显缺陷的图像，所述标准图像表示的是所述标准图像表示的是图像质量明显高于非标准图像的图像；
分别对所述非标准图像与标准图像进行配对标签标注，并输入至用于学习和训练的网络模型中；
通过学习与训练，建立具有所述非标准图像与标准图像内容的一一映射关系的映射模型。
所述的图像处理方法，其中，所述映射模型具有参数初始化功能，当需要改变所述映射模型中的映射关系时，启动所述参数初始化功能或线下对图像重新进行学习和训练，自动获取映射关系参数，得到新的映射模型，或者根据不同的应用领域，对图像成像不同的要求，在网络中人为加入一些参数。
所述的图像处理方法，其中，所述调用预设的用于优化图像的映射模型，根据映射关系对原始图像进行优化处理；获得所述事物的优化后图像，并将所述优化后图像保存或输出具体包括：
当智能终端的成像系统得到某事物的原始图像之后，自动启动图像优化功能；
调用预设的映射模型，将所述原始图像输入至映射模型；
根据映射模型中的映射关系将原始图像处理成优化后图像，并将优化后图像存储或输出至智能终端的显示模块。
一种存储介质，其上存储有多条指令，其中，所述指令适于由处理器加载并执行，以实现上述任一项所述的图像处理方法。
一种智能终端，其中，包括：处理器、与处理器通信连接的存储介质，所述存储介质适于存储多条指令；所述处理器适于调用所述存储介质中的指令，以执行实现上述任一项所述的图像处理方法。
本发明的有益效果：本发明通过创建用于对图像进行优化处理的映射模型，能够在智能终端获得某事物的原始图像之后自动调用映射模型对原始图像进行优化处理，特别是对一些成像质量较差的低端设备，在不更改硬件的前提下，仅通过更改成像质量的算法，就可达到用户的使用需求，有效改善了成像质量，并且整个优化过程更加快速，所需时间更短，一定程度上提高了成像速度，延长设备的使用寿命。
附图说明
图1是本发明的图像处理方法的较佳实施例的流程图。
图2是本发明中待优化处理的原始图像的示意图。
图3是本发明中的原始图像经过映射模型之后得到的标准图像的示意图。
图4本发明的智能终端的较佳实施例的功能原理框图。
具体实施方式
为使本发明的目的、技术方案及优点更加清楚、明确，以下参照附图并举实施例对本发明进一步详细说明。应当理解，此处所描述的具体实施例仅仅用以解释本发明，并不用于限定本发明。
由于目前的很多具有成像功能的智能终端的成像效果不理想，尤其是一些对成像质量要求较高的领域，普通的低端智能终端难以满足需求。例如医疗领域，成像质量的高低会严重影响诊断结果以及诊断效率。而且大多数的医疗设备价格都比较昂贵，更换更高性能的设备也是不现实的。此外，即便是更换更高性能的设备，虽然成像质量提高了，但是成像速度相对来说会减缓，反而影响用户的使用。因此，如何对现有智能终端进行改进，使之能够获得更高质量的图像才是现今亟待解决的问题。
为了解决现有智能终端的缺陷，本发明提供了一种图像处理方法，如图1所示，图1是本发明的图像处理方法的较佳实施例的流程图。所述图像处理方法包括以下步骤：
步骤S100、智能终端对某事物进行影像数据采集，获得事物的原始图像。
具体实施时，本发明所实现的是通过创建用于对图像进行优化处理的映射模型，将未经过任何优化处理的原始图像根据映射关系优化以满足用户使用需求。因此，在上述步骤S100之前，本发明需要预先在智能终端中创建或输入用于对图像进行重建与优化的映射模型。
具体地，首先，本发明可以使用自带图像优化功能的智能终端，首先在关闭图像优化功 能的状态下对某事物或场景进行影像数据采集，得到该事物或场景的非标准图像，然后在开启图像优化功能的状态下对同一事物或场景进行影像数据采集，得到该事物或场景的标准图像，通过上述操作获得了同一事物或场景的非标准图像以及标准图像。
当然，本发明对于获取同一事物或者场景的非标准图像以及标准图像的方法，还可以线下分别采用两个性能具有明显差异的智能终端对同一事物或者同一场景进行影像数据采集，这样，性能高的智能终端获得的是标准图像，性能低的智能终端获得的是非标准图像。其主要目的是为了创建所述映射模型。
上述这两种获取非标准图像以及标准图像的方法仅仅只是在创建映射模型过程中的实施例，并不用于限定本发明，其他方式获得非标准图像以及标准图像的方法仍属于本发明保护的范围。
进一步地，当获取到同一事物或者同一场景的非标准图像和标准图像之后，本发明可以选择性的对非标准图像进行双三次插值预处理，将非标准图像放大到目标对象大小，然后对非标准图像与标准图像进行配对标签标注，并输入至用于学习和训练的网络中进行学习与训练。在学习与训练的过程中，提取标准图像与非标准图像的细节特征，建立起非标准图像与标准图像内容的一一映射关系，进而创建出具有该映射关系的映射模型。还可针对不同领域的成像要求，在网络中加入一些参数，进行人为的干预。所述映射模型的获取方式以及程序可以嵌入至任意具有成像系统的智能终端中，以使一些成像效果较差的智能终端可以使用该映射模型自动对图像进行优化处理，而无需更换智能终端，节约了成本，给用户的使用也带来了方便。
值得说明的是，上述在创建映射模型的过程中所获得的非标准图像表示的是具有明显图像质量缺陷的图像，也就是说用户可以明显感受到模糊不清晰的图像，而标准图像表示的是所述标准图像表示的是图像质量明显高于非标准图像的图像。无论是通过用户肉眼观察的感官效果，还是通过仪器分析出的分辨率、信噪比以及其他参数，都是有着明显的改进。因此，本发明所创建的映射模型其实质就是将低质量图像转变为高质量图像。
进一步地，在创建好所述用于对图像进行优化处理的映射模型之后，将所述映射模型的获取方式以及程序或将所述映射模型直接嵌入至智能终端中，使智能终端具有图像优化功能。当智能终端对某事物进行影像数据采集，获得待优化处理的原始图像，本发明所需要达到的 目的就是将智能终端所获得的待优化处理的原始图像优化处理成和创建映射模型时的标准图像具有相同或相近成像效果的优化后图像，从而提高智能终端成像质量。
步骤S200、调用预设的用于优化图像的映射模型，根据映射关系对原始图像进行优化处理；获得所述事物的优化后图像，并将所述优化后图像输出。
较佳地，所述步骤S200具体包括：
步骤S201、当智能终端的成像系统得到某事物的原始图像之后，自动启动图像优化功能；
步骤S202、调用预设的映射模型，将所述原始图像输入至映射模型；
步骤S203、根据映射模型中的映射关系将原始图像处理成优化后图像，并将优化后图像存储或输出至智能终端的显示模块。
具体实施时，当智能终端得到待优化处理的原始图像时，智能终端自动启动图像优化功能，进而自动调用预设的映射模型。所述映射模型会对原始图像进行预处理，将所述原始图像输入至映射模型，根据映射模型中建立好的将低质量图像转化成高质量图像的映射关系，将原始图像自动优化成高质量的优化后图像，无需用户手动操作，整个优化过程快速、高效。此外，智能终端在优化处理完成之后，还将优化后图像存储或输出至智能终端的显示模块，以便将优化后图像进行显示。从上述公开的图像优化，本发明能够自动调用映射模型对原始图像进行优化处理，从而满足用户使用需求，与传统的使用算法进行图像处理的方法，本发明对于一些成像质量较差的低端设备能够具有明显的成像质量改善效果，并且整个优化过程更加快速，所需时间更短，一定程度上提高了智能终端的成像速度，并减少了人为的主观臆断的图像增强方式。
进一步较佳地，本发明所创建的映射模型中的参数是可以根据使用需求进行自主设置与调整的，即具有参数初始化功能。当用户需要改变所述映射模型中的映射关系时，启动所述参数初始化功能，对图像重新进行学习和训练，自动获取映射关系参数，建立新的映射模型。当然，本发明也可以将非标准图像以及标准图像线下输入，从而重新建立映射模型。例如，如果将带有所述映射模型的智能终端应用至医学领域，则用户在使用智能终端观测骨骼组织或者其他组织结构时是需要更加高的成像效果，只需将获取的骨骼组织或者其他组织结构的非标准图像以及标准图像线下输入，便可自动建立可应用至医学领域的映射模型从而满足用户的使用需求，再将其输入至智能终端中，增加本发明的映射模型的应用范围。
步骤S300、智能终端将输出的标准图像进行显示。
智能终端在优化处理过程完成之后，自动将优化后图像进行显示，以供用户使用。具体地，本发明所达到的效果如图2与图3所示，图2是本发明中待优化处理的原始图像的示意图。图3是本发明中的原始图像经过映射模型之后得到的优化后图像的示意图。图2中的原始图像没有经过任何的优化处理，是待处理的图像，首先从肉眼直接观察的感官效果来看，图2中线条纹路都存在明显的模糊，并且通过分析得到的原始图像的峰值信噪比PSNR＝20.3804dB。而图3是将图2中的原始图像经过本发明的映射模型之后，所得到的优化后图像，同样地，首先从肉眼直接观察的感官效果来看，图3中明显比图2更加清晰，并且通过分析得到的标准图像的峰值信噪比PSNR＝26.3647。从这些具体的参数数据可以看出，图3中的标准图像的成像质量明显比图2中的原始图像高。
基于上述实施例，本发明还公开了一种智能终端。如图4所示，包括：处理器(processor)10、与处理器10连接的存储介质(memory)20；其中，所述处理器10用于调用所述存储介质20中的程序指令，以执行上述实施例所提供的方法，例如执行：
步骤S100、智能终端对某事物进行影像数据采集，获得事物的原始图像；
步骤S200、调用预设的用于优化图像的映射模型对原始图像进行优化处理；获得所述事物的优化后图像，并将所述优化后图像保存或输出；
步骤S300、智能终端将输出的优化后图像进行显示。
需要说明的是，本发明中的智能终端包括任何成像原理的仪器设备，并且适用于任何需要成像的技术领域以及场景。
本发明实施例还提供一种存储介质，所述存储介质上存储计算机指令，所述计算机指令使计算机执行上述各实施例所提供的方法。
综上所述，本发明提供的一种图像处理方法、存储介质及智能终端，方法包括：智能终端对某事物进行影像数据采集，获得事物的原始图像；调用预设的用于优化图像的映射模型，根据映射关系对原始图像进行优化处理；获得所述事物的优化后图像，并将所述优化后图像保存或输出；智能终端将输出的优化后图像进行显示。本发明通过创建用于对图像进行优化处理的映射模型，能够在智能终端获得某事物的原始图像之后自动调用映射模型对原始图像进行优化处理，提高图像质量、成像速率，延长设备的使用寿命，特别是对一些成像质量较 差的低端设备，在不更改硬件的前提下，仅通过更改成像质量的算法，就可达到用户的使用需求，不但改善了成像质量，在一定程度上提高了成像速度。
应当理解的是，本发明的应用不限于上述的举例，对本领域普通技术人员来说，可以根据上述说明加以改进或变换，所有这些改进和变换都应属于本发明所附权利要求的保护范围。

Claims (8)

  1. 一种图像处理方法，其特征在于，所述方法包括：
    智能终端对某事物进行影像数据采集，获得事物的原始图像；
    调用预设的用于优化图像的映射模型，根据映射关系对原始图像进行优化处理；
    获得所述事物的优化后图像，并将所述优化后图像保存或输出；
    智能终端将输出的优化后图像进行显示。
  2. 根据权利要求1中所述的图像处理方法，其特征在于，所述原始图像表示的是待优化处理的图像；
    所述优化后图像表示的是图像质量高于所述原始图像，且满足用户使用需求的图像。
  3. 根据权利要求1中所述的图像处理方法，其特征在于，所述智能终端对某事物进行影像数据采集，获得事物的原始图像之前还包括：
    预先在智能终端中创建或输入用于对图像进行重建与优化的映射模型。
  4. 根据权利要求3中所述的图像处理方法，其特征在于，所述预先在智能终端中创建或输入用于对图像进行重建与优化的映射模型具体包括：
    对同一事物或场景进行影像数据采集，分别获得该事物或场景的非标准图像与标准图像；所述非标准图像表示的是图像质量存在明显缺陷的图像，所述标准图像表示的是所述标准图像表示的是图像质量明显高于非标准图像的图像；
    分别对所述非标准图像与标准图像进行配对标签标注，并输入至用于学习和训练的网络模型中；
    通过学习与训练，建立具有所述非标准图像与标准图像内容的一一映射关系的映射模型。
  5. 根据权利要求3中所述的图像处理方法，其特征在于，所述映射模型具有参数初始化功能，当需要改变所述映射模型中的映射关系时，启动所述参数初始化功能或线下对图像重新进行学习和训练，自动获取映射关系参数，得到新的映射模型，或者根据不同的应用领域，对图像成像不同的要求，在网络中人为加入一些参数。
  6. 根据权利要求1中所述的图像处理方法，其特征在于，所述调用预设的用于 优化图像的映射模型，根据映射关系对原始图像进行优化处理；获得所述事物的优化后图像，并将所述优化后图像保存或输出具体包括：
    当智能终端的成像系统得到某事物的原始图像之后，自动启动图像优化功能；调用预设的映射模型，将所述原始图像输入至映射模型；
    根据映射模型中的映射关系将原始图像处理成优化后图像，并将优化后图像存储或输出至智能终端的显示模块。
  7. 一种存储介质，其上存储有多条指令，其特征在于，所述指令适于由处理器加载并执行，以实现上述权利要求1-6任一项所述的图像处理方法。
  8. 一种智能终端，其特征在于，包括：处理器、与处理器通信连接的存储介质，所述存储介质适于存储多条指令；所述处理器适于调用所述存储介质中的指令，以执行实现上述权利要求1-6任一项所述的图像处理方法。
PCT/CN2017/118026 2017-12-15 2017-12-22 一种图像处理方法、存储介质及智能终端 WO2019114027A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN2017113490553 2017-12-15
CN201711349055.3A CN107945139B (zh) 2017-12-15 2017-12-15 一种图像处理方法、存储介质及智能终端

Publications (1)

Publication Number Publication Date
WO2019114027A1 true WO2019114027A1 (zh) 2019-06-20

Family

ID=61944305

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2017/118026 WO2019114027A1 (zh) 2017-12-15 2017-12-22 一种图像处理方法、存储介质及智能终端

Country Status (2)

Country Link
CN (1) CN107945139B (zh)
WO (1) WO2019114027A1 (zh)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113689342A (zh) * 2020-05-18 2021-11-23 上海联影医疗科技股份有限公司 一种图像质量优化的方法及系统

Families Citing this family (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109284719A (zh) * 2018-09-28 2019-01-29 成都臻识科技发展有限公司 一种基于机器学习的初始数据处理方法和系统
CN112950337A (zh) * 2021-04-27 2021-06-11 拉扎斯网络科技（上海）有限公司 一种图像处理方法、装置、电子设备以及存储介质

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101452575A (zh) * 2008-12-12 2009-06-10 北京航空航天大学 一种基于神经网络的图像自适应增强方法
WO2011081646A1 (en) * 2009-12-15 2011-07-07 Thomson Licensing Stereo-image quality and disparity/depth indications
CN103871037A (zh) * 2012-12-07 2014-06-18 汤姆逊许可公司 图像之间颜色转换的方法和装置
CN107240066A (zh) * 2017-04-28 2017-10-10 天津大学 基于浅层和深层卷积神经网络的图像超分辨率重建算法

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101452575A (zh) * 2008-12-12 2009-06-10 北京航空航天大学 一种基于神经网络的图像自适应增强方法
WO2011081646A1 (en) * 2009-12-15 2011-07-07 Thomson Licensing Stereo-image quality and disparity/depth indications
CN103871037A (zh) * 2012-12-07 2014-06-18 汤姆逊许可公司 图像之间颜色转换的方法和装置
CN107240066A (zh) * 2017-04-28 2017-10-10 天津大学 基于浅层和深层卷积神经网络的图像超分辨率重建算法

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113689342A (zh) * 2020-05-18 2021-11-23 上海联影医疗科技股份有限公司 一种图像质量优化的方法及系统
CN113689342B (zh) * 2020-05-18 2024-04-02 上海联影医疗科技股份有限公司 一种图像质量优化的方法及系统

Also Published As

Publication number Publication date
CN107945139A (zh) 2018-04-20
CN107945139B (zh) 2020-01-07

Similar Documents

Publication Publication Date Title
CN111641778B (zh) 一种拍摄方法、装置与设备
CN105100632B (zh) 成像设备自动曝光的调整方法及装置、成像设备
KR101662846B1 (ko) 아웃 포커싱 촬영에서 빛망울 효과를 생성하기 위한 장치 및 방법
CN105745914B (zh) 用于逆色调映射的方法和系统
WO2019114027A1 (zh) 一种图像处理方法、存储介质及智能终端
WO2014169579A1 (zh) 一种色彩增强方法及装置
WO2015095529A1 (en) Image adjustment using texture mask
CN110445986B (zh) 图像处理方法、装置、存储介质及电子设备
CN109587558B (zh) 视频处理方法、装置、电子设备以及存储介质
CN116113976A (zh) 图像处理方法及装置、计算机可读介质和电子设备
CN109587559A (zh) 视频处理方法、装置、电子设备以及存储介质
CN116109483A (zh) 一种基于gpu的超分辨率和高动态范围处理方法及存储介质
CN108961179B (zh) 一种医学图像后处理系统及其使用方法
CN113658065A (zh) 图像降噪方法及装置、计算机可读介质和电子设备
JP2024037722A (ja) コンテンツに基づく画像処理
WO2019095485A1 (zh) 视频信号处理方法、装置及可读存储介质
CN109636749B (zh) 图像处理方法
CN112104812B (zh) 一种采集图片的方法和装置
CN112887597A (zh) 图像处理方法及装置、计算机可读介质和电子设备
WO2018112979A1 (zh) 一种图像处理方法、装置以及终端设备
CN116668773B (zh) 增强视频画质的方法与电子设备
WO2023272506A1 (zh) 图像处理方法及装置、可移动平台及存储介质
CN113766142B (zh) 图像处理装置、图像信号前处理模块、设备和处理方法
KS Enhancing Visual Realism with Convolutional Neural Networks
CN118134786A (zh) 红外图像的多路isp处理方法、装置、设备及介质

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: 17934858

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

32PN Ep: public notification in the ep bulletin as address of the adressee cannot be established

Free format text: NOTING OF LOSS OF RIGHTS PURSUANT TO RULE 112(1) EPC (EPO FORM 1205A DATED 22.09.2020)

122 Ep: pct application non-entry in european phase

Ref document number: 17934858

Country of ref document: EP

Kind code of ref document: A1