CN117475121A - Image processing system and image processing method - Google Patents

Image processing system and image processing method Download PDF

Info

Publication number
CN117475121A
CN117475121A CN202311651983.0A CN202311651983A CN117475121A CN 117475121 A CN117475121 A CN 117475121A CN 202311651983 A CN202311651983 A CN 202311651983A CN 117475121 A CN117475121 A CN 117475121A
Authority
CN
China
Prior art keywords
image
module
adopting
sub
generate
Prior art date
Legal status (The legal status 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 status listed.)
Withdrawn
Application number
CN202311651983.0A
Other languages
Chinese (zh)
Inventor
项泽宇
董永生
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Henan University of Science and Technology
Original Assignee
Henan University of Science and Technology
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 Henan University of Science and Technology filed Critical Henan University of Science and Technology
Priority to CN202311651983.0A priority Critical patent/CN117475121A/en
Publication of CN117475121A publication Critical patent/CN117475121A/en
Withdrawn legal-status Critical Current

Links

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/10—Image acquisition
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/50—Information retrieval; Database structures therefor; File system structures therefor of still image data
    • G06F16/51—Indexing; Data structures therefor; Storage structures
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/50—Information retrieval; Database structures therefor; File system structures therefor of still image data
    • G06F16/58—Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually
    • G06F16/583—Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually using metadata automatically derived from the content
    • G06F16/5838—Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually using metadata automatically derived from the content using colour
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/50—Information retrieval; Database structures therefor; File system structures therefor of still image data
    • G06F16/58—Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually
    • G06F16/583—Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually using metadata automatically derived from the content
    • G06F16/5854—Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually using metadata automatically derived from the content using shape and object relationship
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/50—Information retrieval; Database structures therefor; File system structures therefor of still image data
    • G06F16/58—Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually
    • G06F16/583—Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually using metadata automatically derived from the content
    • G06F16/5862—Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually using metadata automatically derived from the content using texture
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/0464—Convolutional networks [CNN, ConvNet]
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/20—Image preprocessing
    • G06V10/24—Aligning, centring, orientation detection or correction of the image
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/20—Image preprocessing
    • G06V10/26—Segmentation of patterns in the image field; Cutting or merging of image elements to establish the pattern region, e.g. clustering-based techniques; Detection of occlusion
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/20—Image preprocessing
    • G06V10/30—Noise filtering
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/40—Extraction of image or video features
    • G06V10/44—Local feature extraction by analysis of parts of the pattern, e.g. by detecting edges, contours, loops, corners, strokes or intersections; Connectivity analysis, e.g. of connected components
    • G06V10/443—Local feature extraction by analysis of parts of the pattern, e.g. by detecting edges, contours, loops, corners, strokes or intersections; Connectivity analysis, e.g. of connected components by matching or filtering
    • G06V10/449—Biologically inspired filters, e.g. difference of Gaussians [DoG] or Gabor filters
    • G06V10/451—Biologically inspired filters, e.g. difference of Gaussians [DoG] or Gabor filters with interaction between the filter responses, e.g. cortical complex cells
    • G06V10/454—Integrating the filters into a hierarchical structure, e.g. convolutional neural networks [CNN]
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/40—Extraction of image or video features
    • G06V10/54—Extraction of image or video features relating to texture
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
    • G06V10/764—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using classification, e.g. of video objects

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Multimedia (AREA)
  • Library & Information Science (AREA)
  • Databases & Information Systems (AREA)
  • General Engineering & Computer Science (AREA)
  • Data Mining & Analysis (AREA)
  • Evolutionary Computation (AREA)
  • Software Systems (AREA)
  • General Health & Medical Sciences (AREA)
  • Artificial Intelligence (AREA)
  • Health & Medical Sciences (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Computing Systems (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Biomedical Technology (AREA)
  • Molecular Biology (AREA)
  • Biodiversity & Conservation Biology (AREA)
  • Medical Informatics (AREA)
  • Biophysics (AREA)
  • Computational Linguistics (AREA)
  • Mathematical Physics (AREA)
  • Image Analysis (AREA)

Abstract

The invention relates to the technical field of image processing, in particular to an image processing system and an image processing method, wherein the system comprises an image acquisition module, an image preprocessing module, a feature extraction module, an image quality evaluation module, an intelligent screening module, an image labeling module and an image storage and retrieval module; the image acquisition module acquires an original image by adopting a multispectral imaging technology and a real-time data stream processing algorithm based on the multisource input to generate an original image set; in the invention, the original image quality obtained from multi-source input is obviously improved through a multi-spectrum imaging technology and a real-time data stream processing algorithm, the images are efficiently processed by utilizing a wavelet transformation denoising and local self-adaptive histogram equalization technology, the defect of processing different source image data is overcome, the high-precision image feature extraction is realized by utilizing a ResNet model and an image analysis technology, the peak signal-to-noise ratio algorithm is used for comprehensively evaluating the image quality, and the evaluation accuracy is improved.

Description

Image processing system and image processing method
Technical Field
The present invention relates to the field of image processing technologies, and in particular, to an image processing system and an image processing method.
Background
Image processing technology, in which researchers and engineers develop algorithms and systems for image enhancement, restoration, reconstruction, pattern recognition, image compression, and image content understanding tasks, focuses on enabling a computer to acquire high-level information from an image or multidimensional data, involves acquisition, analysis, processing, and understanding of an image, and generally focuses on conversion and interpretation of image data.
Among them, an image processing system is a computing system designed for analyzing and manipulating image data, and the main purpose is to improve image quality or extract valuable information from images through processing and analyzing image data, and applications of the image processing system are very wide, including fields of medical imaging, satellite data processing, machine vision, photo editing, and the like.
Conventional image processing techniques lack efficient methods for synthesizing and optimizing image data from different sources when processing multi-source input data, particularly in the application of multi-spectral imaging techniques, while algorithms such as image enhancement and restoration have been widely used, the efficiency and accuracy of the algorithms still remain to be improved when processing large-scale or complex image data, such as high-precision image feature extraction and comprehensive quality assessment are prone to uncertainty, and for storage and retrieval of large-scale image data, particularly in cloud storage and distributed systems, there are limitations in data indexing and retrieval efficiency, and it is difficult to effectively manage and access a large amount of data.
Disclosure of Invention
The invention aims to solve the defects in the prior art and provides an image processing system and an image processing method.
In order to achieve the above purpose, the present invention adopts the following technical scheme: the image processing system comprises an image acquisition module, an image preprocessing module, a feature extraction module, an image quality evaluation module, an intelligent screening module, an image labeling module and an image storage and retrieval module;
the image acquisition module acquires an original image by adopting a multispectral imaging technology and a real-time data stream processing algorithm based on the multisource input to generate an original image set;
the image preprocessing module optimizes the image set based on the original image set by adopting wavelet transformation denoising and local self-adaptive histogram equalization technology to generate an optimized image set;
the feature extraction module is used for extracting image features by adopting a ResNet model and an image analysis technology based on the optimized image set to generate an image feature data set;
the image quality evaluation module is used for carrying out image quality analysis by adopting a peak signal-to-noise ratio algorithm based on the image characteristic data set to generate an image quality evaluation report;
The intelligent screening module performs image identification screening by adopting an intelligent classification algorithm and a pattern recognition technology based on the image quality evaluation report to generate a screened image set;
the image annotation module automatically generates a description tag on an image based on the screened image set by adopting a text generation model based on a transducer and a computer vision technology to generate an image annotation data set;
the image storage and retrieval module is used for storing images and labels thereof and providing retrieval service by adopting a distributed NoSQL database and an image retrieval algorithm based on an image label data set to generate an image database;
the original image set comprises image data from a camera, a cloud image library and a real-time monitoring video stream, the optimized image set specifically refers to an image set subjected to multistage denoising, dynamic contrast adjustment, color balance correction, intelligent clipping adjustment and self-adaptive rotation correction, the image characteristic data set comprises multi-layer characteristic extraction for image edge detection, detail texture analysis, geometric shape recognition and full color spectrum analysis, the image quality evaluation report comprises evaluation for image definition, exposure, noise level, color saturation and contrast, the screened image set specifically refers to an image set subjected to fuzzy image elimination, overexposure correction, multistage noise filtration and color recovery treatment, and the image annotation data set comprises object recognition, scene content analysis, emotion and context analysis and label distribution of the image.
As a further scheme of the invention, the image acquisition module comprises a camera acquisition sub-module, a cloud image library synchronization sub-module, a real-time monitoring video stream processing sub-module and a format compatible conversion sub-module;
the camera acquisition submodule is used for carrying out image acquisition by adopting a CMOS image sensor and an image capturing algorithm based on multi-source input to generate a captured image data set;
the cloud image library synchronization submodule generates cloud synchronous image data by adopting a cloud storage technology and a network transmission protocol based on a captured image data set;
the real-time monitoring video stream processing sub-module is used for optimizing the video stream by adopting a real-time video processing technology and an H.264 video compression standard based on cloud synchronous image data to generate a video stream data set.
The format compatible conversion submodule carries out image format conversion and optimization by adopting an image format conversion tool and a JPEG encoding and decoding standard based on the video stream data set to generate an original image set.
As a further scheme of the invention, the image preprocessing module comprises a multi-stage denoising processing sub-module, a dynamic contrast adjusting sub-module, a color balance correcting sub-module and an intelligent clipping adjusting sub-module;
The multistage denoising processing submodule adopts a wavelet transformation denoising technology to decompose and reconstruct signals based on an original image set, and generates denoised image data;
the dynamic contrast adjustment submodule carries out dynamic contrast adjustment on the image by adopting a local self-adaptive histogram equalization technology based on the denoised image data to generate contrast optimized image data;
the color balance correction submodule optimizes the color of the image by adopting a white balance algorithm based on the image data with optimized contrast, and generates image data with balanced color;
the intelligent clipping and adjusting submodule adopts an intelligent clipping technology based on the image data of color balance, and automatically optimizes the size and proportion of the image according to the content of the image to generate an optimized image set.
As a further scheme of the invention, the feature extraction module comprises an edge detection sub-module, a texture analysis sub-module, a shape recognition sub-module and a full color spectrum analysis sub-module;
the edge detection submodule carries out edge detection by adopting a Canny operator based on the optimized image set to generate edge-enhanced image data;
the texture analysis submodule carries out texture feature analysis by adopting a gray level co-occurrence matrix based on the edge-enhanced image data to generate texture feature analysis data;
The shape recognition submodule carries out geometric shape recognition and classification by adopting Hough transformation based on texture feature analysis data to generate shape recognition data;
the full-color spectrum analysis submodule is used for carrying out color distribution and contrast analysis by adopting an RGB color model based on the shape identification data to generate an image characteristic data set.
As a further scheme of the invention, the image quality evaluation module comprises a definition analysis sub-module, an intelligent exposure evaluation sub-module, a noise level detection sub-module and a color saturation sub-module;
the definition analysis submodule carries out definition analysis by using a Laplacian operator based on the image characteristic data set to generate definition analysis data;
the intelligent exposure evaluation submodule evaluates the exposure level of the image by adopting a histogram analysis method based on the definition analysis data to generate exposure evaluation data;
the noise level detection submodule is used for identifying image noise points and granularity by adopting a frequency domain analysis algorithm based on exposure evaluation data to generate noise level detection data;
the color saturation measurement submodule adopts HSV color space analysis to measure the color saturation of the image based on noise level detection data and generates an image quality assessment report.
As a further scheme of the invention, the intelligent screening module comprises a blurred image elimination sub-module, an overexposure correction sub-module, a multi-stage noise filtering sub-module and a color recovery processing sub-module;
the blurred image elimination sub-module is used for identifying and eliminating blurred images by adopting a Laplace transformation algorithm based on an image quality evaluation report, and a refined image data set is generated;
the overexposure correction submodule carries out overexposure correction by adopting a dynamic range compression technology based on the thinned image data set to generate an exposure correction image set;
the multistage noise filtering submodule removes image noise points by adopting a Gaussian filtering algorithm and a bilateral filtering algorithm based on the exposure correction image set to generate a processed image set;
the color recovery processing sub-module optimizes the colors of the images by adopting a color recovery technology based on the processed image set, and generates a filtered image set.
As a further scheme of the invention, the image annotation module comprises an intelligent object recognition sub-module, a scene content analysis sub-module, a mood and context analysis sub-module and an automatic keyword marking sub-module;
the intelligent object identification sub-module is used for identifying main objects in the image by adopting an SSD algorithm based on the screened image set, and generating an object identification data set;
The scene content analysis submodule analyzes the scene content of the image by using a scene analysis algorithm based on the object identification data set to generate a scene analysis data set;
the emotion and context analysis submodule evaluates the emotion and the context of the image based on the scene analysis data set by applying an emotion recognition technology and a context understanding algorithm to generate an emotion context data set;
the automatic keyword marking submodule automatically extracts and marks keywords based on the emotion context data set by adopting a TF-IDF analysis technology to generate an image marking data set.
As a further scheme of the invention, the image storage and retrieval module comprises an image storage sub-module, a labeling information management sub-module, an intelligent retrieval engine sub-module and a query optimization sub-module;
the image storage submodule stores image data by adopting a distributed storage technology based on an image annotation data set to generate a distributed image storage library;
the annotation information management submodule is used for maintaining and managing image annotation information by adopting MongoDB based on the distributed image storage library to generate an annotation information management library;
the intelligent search engine submodule develops an intelligent search engine by adopting an image search algorithm based on the annotation information management library to generate a search system;
The query optimization submodule adopts a query optimization technology to improve the retrieval performance based on the retrieval system and generates an image database.
An image processing method, which is executed based on the above image processing system, comprising the steps of:
s1: based on multi-source input, acquiring an original image by adopting a CMOS image sensor and an image capturing algorithm to generate an original image set;
s2: based on the original image set, performing image preprocessing by adopting a wavelet transformation denoising technology and local self-adaptive histogram equalization to generate an optimized image set;
s3: based on the optimized image set, feature extraction is carried out by adopting a ResNet model and an image analysis technology, and an image feature data set is generated;
s4: based on the image characteristic data set, performing image quality evaluation by adopting a peak signal-to-noise ratio algorithm, and generating an image quality evaluation report;
s5: based on the image quality evaluation report, intelligent screening is carried out by adopting a support vector machine intelligent classification algorithm and a pattern recognition technology, and a screened image set is generated;
s6: based on the screened image set, carrying out image annotation by adopting a natural language processing technology based on a transducer model, and generating an image annotation data set;
S7: based on the image annotation data set, performing image storage and retrieval by adopting a distributed NoSQL database and an image retrieval algorithm to generate an image database;
s8: based on the image database, adopting MongoDB technology to manage and optimize the annotation information, and generating an annotation information management library;
s9: and based on the annotation information management library, searching by adopting a query optimization technology and an intelligent search engine to generate a comprehensive query optimization and search system.
As a further scheme of the invention, the optimized image set specifically refers to an image subjected to denoising and contrast adjustment, the image data set comprises edge, texture, shape and color information of the image, the image quality evaluation report comprises evaluation of definition, exposure, noise, color and contrast, the screened image set specifically refers to an image without blurring and moderate exposure, and the image annotation data set comprises image description text and keyword labels.
Compared with the prior art, the invention has the advantages and positive effects that:
according to the invention, the acquisition quality of an original image is effectively improved by combining a multispectral imaging technology and a real-time data stream processing algorithm aiming at multisource input, the image is processed by using wavelet transformation denoising and local self-adaptive histogram equalization technology, the defects in the aspect of processing image data from different sources in the traditional technology are effectively overcome, the feature extraction is carried out on an optimized image set by adopting a ResNet model and an image analysis technology, the high-precision image feature analysis is realized, the accuracy of feature extraction is improved, the data processing efficiency of the whole system is optimized, the limitation of the traditional algorithm in the aspect of processing complex image data is solved, the image quality evaluation is realized, the image quality is comprehensively evaluated by adopting a peak signal-to-noise ratio algorithm, the accuracy and the reliability of evaluation are improved, the storage structure of the data is optimized by an image storage and retrieval module, the efficiency of data indexing and retrieval is improved, and the efficiency problem of the traditional system in the process and management of a large-scale image library is solved.
Drawings
FIG. 1 is a system flow diagram of the present invention;
FIG. 2 is a schematic diagram of a system framework of the present invention;
FIG. 3 is a flow chart of an image acquisition module of the present invention;
FIG. 4 is a flow chart of an image preprocessing module according to the present invention;
FIG. 5 is a flow chart of the feature extraction module of the present invention;
FIG. 6 is a flow chart of the image quality assessment module of the present invention;
FIG. 7 is a flow chart of the intelligent screening module of the present invention;
FIG. 8 is a flow chart of the image annotation module of the present invention;
FIG. 9 is a flow chart of the image storage and retrieval module of the present invention;
FIG. 10 is a schematic diagram of the method steps of the present invention.
Detailed Description
The present invention will be described in further detail with reference to the drawings and examples, in order to make the objects, technical solutions and advantages of the present invention more apparent. It should be understood that the specific embodiments described herein are for purposes of illustration only and are not intended to limit the scope of the invention.
In the description of the present invention, it should be understood that the terms "length," "width," "upper," "lower," "front," "rear," "left," "right," "vertical," "horizontal," "top," "bottom," "inner," "outer," and the like indicate orientations or positional relationships based on the orientation or positional relationships shown in the drawings, merely to facilitate describing the present invention and simplify the description, and do not indicate or imply that the devices or elements referred to must have a specific orientation, be configured and operated in a specific orientation, and therefore should not be construed as limiting the present invention. Furthermore, in the description of the present invention, the meaning of "a plurality" is two or more, unless explicitly defined otherwise.
Example 1
Referring to fig. 1-2, the present invention provides a technical solution: the image processing system comprises an image acquisition module, an image preprocessing module, a feature extraction module, an image quality evaluation module, an intelligent screening module, an image labeling module and an image storage and retrieval module;
the image acquisition module acquires an original image by adopting a multispectral imaging technology and a real-time data stream processing algorithm based on the multisource input to generate an original image set;
the image preprocessing module optimizes the image set based on the original image set by adopting wavelet transformation denoising and local self-adaptive histogram equalization technology to generate an optimized image set;
the feature extraction module is used for extracting image features by adopting a ResNet model and an image analysis technology based on the optimized image set to generate an image feature data set;
the image quality evaluation module is used for carrying out image quality analysis by adopting a peak signal-to-noise ratio algorithm based on the image characteristic data set to generate an image quality evaluation report;
the intelligent screening module carries out image identification screening by adopting an intelligent classification algorithm and a pattern recognition technology based on the image quality evaluation report to generate a screened image set;
the image annotation module automatically generates a description tag on the image based on the screened image set by adopting a text generation model based on a transducer and a computer vision technology to generate an image annotation data set;
The image storage and retrieval module is used for storing images and labels thereof and providing retrieval service by adopting a distributed NoSQL database and an image retrieval algorithm based on an image label data set to generate an image database;
the original image set comprises image data from a camera, a cloud image library and a real-time monitoring video stream, the optimized image set specifically refers to an image subjected to multistage denoising, dynamic contrast adjustment, color balance correction, intelligent clipping adjustment and self-adaptive rotation correction, the image characteristic data set comprises multi-layer characteristic extraction for image edge detection, detail texture analysis, geometric shape recognition and full color spectrum analysis, the image quality evaluation report comprises evaluation for image definition, exposure, noise level, color saturation and contrast, the screened image set specifically refers to an image set subjected to fuzzy image elimination, overexposure correction, multistage noise filtration and color recovery, and the image annotation data set comprises object recognition, scene content analysis, emotion and context analysis and label distribution of the image.
Based on a transform text generation model, a computer vision technology, a distributed NoSQL database and an image retrieval algorithm, remarkable efficiency and quality improvement are brought to image processing, the system can efficiently acquire original images from various sources and optimize the images through an advanced preprocessing technology so as to facilitate more accurate subsequent processing, the application of a deep learning model enables feature extraction and quality evaluation to become more accurate, reliable data support is provided for intelligent screening and automatic labeling, and the high-efficiency storage and quick retrieval functions of the system remarkably improve the capability of processing a large-scale image library, so that the system is superior to the traditional image processing method in terms of automation degree, processing speed and accuracy, and is particularly suitable for scenes needing to rapidly process and manage a large amount of image data.
Referring to fig. 3, the image acquisition module includes a camera acquisition sub-module, a cloud image library synchronization sub-module, a real-time monitoring video stream processing sub-module, and a format compatible conversion sub-module;
the camera acquisition submodule adopts a CMOS image sensor and an image capturing algorithm to acquire images based on multi-source input, and a captured image data set is generated;
the cloud image library synchronization submodule generates cloud synchronous image data by adopting a cloud storage technology and a network transmission protocol based on the captured image data set;
the real-time monitoring video stream processing sub-module is used for optimizing the video stream by adopting a real-time video processing technology and an H.264 video compression standard based on cloud synchronous image data to generate a video stream data set.
The format compatible conversion submodule carries out image format conversion and optimization by adopting an image format conversion tool and a JPEG encoding and decoding standard based on the video stream data set to generate an original image set.
In the image acquisition module, the camera acquisition submodule captures images by using a CMOS image sensor, the quality and definition of the captured image data are ensured, the cloud image library synchronization submodule uploads the captured image data sets to cloud storage, the speed and the safety of data transmission are ensured by means of a rapid network transmission protocol, the real-time monitoring video stream processing submodule uses the image data synchronized to the cloud for generating and optimizing a real-time video stream, the quality of the video stream is improved by using a real-time video processing technology and an H.264 video compression standard, meanwhile, smooth and coherent video playing is ensured, the video stream data are converted into various formats by the format compatibility conversion submodule, different use scenes and equipment requirements are met, and the image format conversion tool and the JPEG encoding and decoding standard are adopted, so that the wide compatibility and the flexibility of the image data are realized, the high-quality acquisition and the processing of the image data are ensured, and the accessibility and the applicability of the image data are improved by cloud synchronization and format conversion.
Referring to fig. 4, the image preprocessing module includes a multi-stage denoising processing sub-module, a dynamic contrast adjustment sub-module, a color balance correction sub-module, and an intelligent clipping adjustment sub-module;
the multistage denoising processing submodule adopts a wavelet transformation denoising technology to decompose and reconstruct signals based on an original image set, and generates denoised image data;
the dynamic contrast adjustment submodule carries out dynamic contrast adjustment on the image by adopting a local self-adaptive histogram equalization technology based on the denoised image data to generate contrast optimized image data;
the color balance correction submodule optimizes the colors of the image by adopting a white balance algorithm based on the image data with optimized contrast, and generates image data with balanced colors;
the intelligent clipping and adjusting submodule adopts an intelligent clipping technology based on the image data of color balance, and automatically optimizes the size and proportion of the image according to the content of the image to generate an optimized image set.
In the image preprocessing module, a multi-stage denoising processing sub-module adopts a wavelet transformation denoising technology to process an original image set from an image acquisition module, the method comprises the steps of decomposing and reconstructing an image signal, effectively removing noise in an image, generating denoised image data with higher definition, keeping original details of the image, reducing interference information, receiving a dynamic contrast adjusting sub-module, working based on the denoised image data, enabling details of the image to be more obvious, generating contrast optimized image data, further processing the contrast optimized image data by a color balance correcting sub-module, adjusting color distribution in the image by a white balance algorithm, enabling colors to be more natural and balanced, generating color balanced image data, and automatically optimizing the size and proportion of the image according to the image content by an intelligent clipping adjusting sub-module according to the color balanced image data, ensuring optimal visual performance and applicability of the image, and generating an optimized image set.
Referring to fig. 5, the feature extraction module includes an edge detection sub-module, a texture analysis sub-module, a shape recognition sub-module, and a full color spectrum analysis sub-module;
the edge detection submodule carries out edge detection by adopting a Canny operator based on the optimized image set to generate edge-enhanced image data;
the texture analysis submodule carries out texture feature analysis by adopting a gray level co-occurrence matrix based on the edge-enhanced image data to generate texture feature analysis data;
the shape recognition submodule carries out geometric shape recognition and classification by adopting Hough transformation based on texture feature analysis data to generate shape recognition data;
the full-color spectrum analysis submodule adopts an RGB color model to carry out color distribution and contrast analysis based on the shape identification data, and generates an image characteristic data set.
The edge detection sub-module takes over the adjusted image data output by the intelligent clipping and adjusting sub-module, carries out edge detection by adopting a Canny operator, can effectively identify the outline and other important edges of an object in an image, generates image data with enhanced edges, the texture analysis sub-module carries out deep analysis on the image data with enhanced edges by a gray level co-occurrence matrix technology, provides important information about the surface characteristics of the image, and further generates texture feature analysis data, the shape identification sub-module identifies and classifies geometric shapes in the image, such as straight lines, circles and the like, by adopting a Hough transformation algorithm based on the texture feature analysis data, through the step, the system can generate accurate shape identification data, the full color spectrum analysis sub-module carries out analysis by adopting an RGB color model based on the shape identification data, and the stage carries out key analysis on the color distribution and comparison of the image, so as to generate an image feature data set, the data not only provides detailed information about the colors of the image, but also helps better understand and express the overall visual effect of the image.
Referring to fig. 6, the image quality evaluation module includes a sharpness analysis sub-module, an intelligent exposure evaluation sub-module, a noise level detection sub-module, and a color saturation sub-module;
the definition analysis submodule carries out definition analysis by using a Laplacian operator based on the image characteristic data set to generate definition analysis data;
the intelligent exposure evaluation submodule evaluates the exposure level of the image by adopting a histogram analysis method based on the definition analysis data to generate exposure evaluation data;
the noise level detection submodule is used for identifying image noise points and granularity by adopting a frequency domain analysis algorithm based on exposure evaluation data to generate noise level detection data;
the color saturation measurement submodule adopts HSV color space analysis to measure the color saturation of the image based on the noise level detection data and generates an image quality assessment report.
The definition analysis sub-module receives the image characteristic data set, applies the Laplacian to carry out definition analysis, highlights the details and the outline of the image, thereby accurately evaluating the definition of the image, the intelligent exposure evaluation sub-module evaluates the exposure level of the image by using a histogram analysis method based on the definition analysis data, thereby effectively judging whether the image is overexposed or underexposed, generating exposure evaluation data, the noise level detection sub-module recognizes the noise point and granularity in the image by using a frequency domain analysis algorithm based on the exposure evaluation data, thereby effectively detecting and quantifying the noise in the image, generating noise level detection data, the color saturation measurement sub-module carries out color saturation measurement by using an HSV color space according to accurately quantify the color saturation of the image, generating an image quality evaluation report, and ensuring that the quality of the finally output image accords with the standard.
Referring to fig. 7, the intelligent screening module includes a blurred image removing sub-module, an overexposure correcting sub-module, a multi-stage noise filtering sub-module, and a color recovery processing sub-module;
the fuzzy image elimination sub-module is used for identifying and eliminating fuzzy images by adopting a Laplace transformation algorithm based on the image quality evaluation report, and generating a thinned image data set;
the overexposure correction submodule carries out overexposure correction by adopting a dynamic range compression technology based on the thinned image data set to generate an exposure correction image set;
the multistage noise filtering submodule removes image noise points by adopting a Gaussian filtering algorithm and a bilateral filtering algorithm based on the exposure correction image set to generate a processed image set;
the color recovery processing sub-module optimizes the colors of the images by adopting a color recovery technology based on the processed image set, and generates a filtered image set.
The fuzzy image elimination sub-module receives an image quality evaluation report, analyzes the image by adopting a Laplace transformation algorithm, identifies and eliminates a fuzzy image, thereby effectively identifying a fuzzy and defocused image area, the generated thinned image data set comprises an image with higher definition, the overexposure correction sub-module adjusts the exposure level of the image by adopting a dynamic range compression technology based on the thinned image data set, ensures the brightness and contrast balance of the image, generates an exposure correction image set, the multistage noise filtering sub-module further processes the exposure correction image set by adopting a Gaussian filter and a bilateral filter algorithm, the Gaussian filter can effectively smooth the image, small noise points are removed, the bilateral filter reduces noise while maintaining the edge definition, the processed image set is generated, and the color recovery processing sub-module performs color optimization on the processed image set, so that the visual effect of the image is improved, and the screened image set is generated.
Referring to fig. 8, the image labeling module includes an intelligent object recognition sub-module, a scene content analysis sub-module, a mood and context analysis sub-module, and an automatic keyword labeling sub-module;
the intelligent object recognition sub-module recognizes main objects in the image by adopting an SSD algorithm based on the screened image set, and generates an object recognition data set;
the scene content analysis submodule analyzes the scene content of the image by using a scene analysis algorithm based on the object identification data set to generate a scene analysis data set;
the emotion and context analysis submodule evaluates the emotion and the context of the image based on the scene analysis data set by applying an emotion recognition technology and a context understanding algorithm to generate an emotion context data set;
the automatic keyword marking submodule automatically extracts and marks keywords based on the emotion context data set by adopting a TF-IDF analysis technology to generate an image marking data set.
The intelligent object recognition submodule receives the image set after color optimization, performs object recognition by using a single multi-frame detection algorithm, recognizes different types of objects in the image, generates an object recognition data set, the scene content analysis submodule deeply analyzes the context information of the image based on the object recognition data set by using a scene analysis algorithm, understands the whole scene and background of the image, such as natural scenery, urban environment and the like, generates a scene analysis data set, improves the accuracy and richness of image annotation, further deepens and analyzes emotion and context in the image by using an emotion recognition technology and a context understanding algorithm, evaluates emotion atmosphere and context in the image, thereby recognizing emotion elements such as happiness, sadness and the like in the image, simultaneously analyzes the context and implicit meaning of the image, generates an emotion context data set, and the automatic keyword marking submodule extracts and marks keywords of the image content by using a word frequency-inverse document frequency analysis technology based on the emotion context data set, so that keywords in the image content can be effectively recognized, and the image annotation data set can be generated.
Referring to fig. 9, the image storage and retrieval module includes an image storage sub-module, a labeling information management sub-module, an intelligent retrieval engine sub-module, and a query optimization sub-module;
the image storage sub-module stores image data by adopting a distributed storage technology based on the image annotation data set to generate a distributed image storage library;
the annotation information management submodule is used for maintaining and managing image annotation information by adopting a MongoDB based on the distributed image storage library to generate an annotation information management library;
the intelligent search engine submodule develops an intelligent search engine by adopting an image search algorithm based on the annotation information management library to generate a search system;
the query optimization submodule adopts a query optimization technology to improve the retrieval performance based on the retrieval system and generates an image database.
The image storage submodule adopts an advanced distributed storage technology, can store a large amount of image data on a plurality of server nodes to form a highly reliable and extensible distributed image storage library, the storage efficiency is improved, the safety and stability of the data are enhanced, the annotation information management submodule is used for maintaining and managing image annotation information by using a MongoDB (unified database) which is a powerful non-relational database based on the storage library, a large amount of image annotation data can be effectively processed to generate a structured annotation information management library, the intelligent search engine submodule adopts an advanced image search algorithm to develop an intelligent search engine, related images can be quickly searched according to a query request of a user to provide accurate search results, so that the efficiency and experience of the user for searching the images are improved, the query optimization submodule further optimizes the intelligent search engine, the intelligent search engine is applied to various query optimization technologies such as query caching, index optimization and the like, the performance and response speed of a search system are improved, and an optimized high-efficiency query system is generated.
Referring to fig. 10, an image processing method, which is executed based on the image processing system, includes the following steps:
s1: based on multi-source input, acquiring an original image by adopting a CMOS image sensor and an image capturing algorithm to generate an original image set;
s2: based on the original image set, performing image preprocessing by adopting a wavelet transformation denoising technology and local self-adaptive histogram equalization to generate an optimized image set;
s3: based on the optimized image set, adopting a ResNet model and an image analysis technology to perform feature extraction to generate an image feature data set;
s4: based on the image characteristic data set, performing image quality evaluation by adopting a peak signal-to-noise ratio algorithm, and generating an image quality evaluation report;
s5: based on the image quality evaluation report, intelligent screening is carried out by adopting a support vector machine intelligent classification algorithm and a pattern recognition technology, and a screened image set is generated;
s6: based on the screened image set, carrying out image annotation by adopting a natural language processing technology based on a transducer model, and generating an image annotation data set;
s7: based on the image annotation data set, performing image storage and retrieval by adopting a distributed NoSQL database and an image retrieval algorithm to generate an image database;
S8: based on the image database, adopting MongoDB technology to manage and optimize the annotation information, and generating an annotation information management library;
s9: based on the label information management library, searching is carried out by adopting a query optimization technology and an intelligent search engine, and a comprehensive query optimization and search system is generated.
The optimized image set specifically refers to an image subjected to denoising and contrast adjustment, the image data set comprises edge, texture, shape and color information of the image, the image quality evaluation report comprises evaluation on definition, exposure, noise, color and contrast, the screened image set specifically refers to an image without blurring and with moderate exposure, and the image annotation data set comprises an image description text and a keyword label.
The CMOS image sensor and the image capturing algorithm can be used for efficiently acquiring original image data from various sources, so that the richness and diversity of the image data are guaranteed, the application of wavelet transform denoising technology and local self-adaptive histogram equalization greatly improves the quality of images in an image preprocessing stage, the visual effect of the images is enhanced, the ResNet model and the image analysis technology adopted in a feature extraction stage can accurately extract key features from the optimized images in a concentrated mode, the image data can be accurately interpreted and utilized, the evaluation of the image quality is performed through a peak signal-to-noise ratio algorithm, the quality of a plurality of aspects such as the definition, exposure and noise of the images can be comprehensively evaluated, reliable judgment basis is provided for intelligent screening, the application of a vector machine intelligent classification algorithm and a pattern recognition technology enables the method to automatically select high-quality images in an intelligent screening stage, the requirements of manual screening are reduced, the processing efficiency is improved, the image annotation link is processed through natural language of a tranmer model, the image data can be accurately interpreted and utilized, the image quality is evaluated through a peak signal-to-noise ratio algorithm, the quality of the image can be comprehensively evaluated, the quality of the image is reliably judged, the quality is reliably judged basis is provided for intelligent screening, the application of a large-quality database is improved, the query is more convenient to search system is provided, the retrieval system is more convenient to store, and has the advantages of the retrieval and has high retrieval and has the advantages, and high retrieval quality, and is more convenient and convenient to use.
The present invention is not limited to the above embodiments, and any equivalent embodiments which can be changed or modified by the technical disclosure described above can be applied to other fields, but any simple modification, equivalent changes and modification made to the above embodiments according to the technical matter of the present invention will still fall within the scope of the technical disclosure.

Claims (10)

1. An image processing system, characterized by: the system comprises an image acquisition module, an image preprocessing module, a feature extraction module, an image quality evaluation module, an intelligent screening module, an image labeling module and an image storage and retrieval module;
the image acquisition module acquires an original image by adopting a multispectral imaging technology and a real-time data stream processing algorithm based on the multisource input to generate an original image set;
the image preprocessing module optimizes the image set based on the original image set by adopting wavelet transformation denoising and local self-adaptive histogram equalization technology to generate an optimized image set;
the feature extraction module is used for extracting image features by adopting a ResNet model and an image analysis technology based on the optimized image set to generate an image feature data set;
The image quality evaluation module is used for carrying out image quality analysis by adopting a peak signal-to-noise ratio algorithm based on the image characteristic data set to generate an image quality evaluation report;
the intelligent screening module performs image identification screening by adopting an intelligent classification algorithm and a pattern recognition technology based on the image quality evaluation report to generate a screened image set;
the image annotation module automatically generates a description tag on an image based on the screened image set by adopting a text generation model based on a transducer and a computer vision technology to generate an image annotation data set;
the image storage and retrieval module is used for storing images and labels thereof and providing retrieval service by adopting a distributed NoSQL database and an image retrieval algorithm based on an image label data set to generate an image database;
the original image set comprises image data from a camera, a cloud image library and a real-time monitoring video stream, the optimized image set specifically refers to an image set subjected to multistage denoising, dynamic contrast adjustment, color balance correction, intelligent clipping adjustment and self-adaptive rotation correction, the image characteristic data set comprises multi-layer characteristic extraction for image edge detection, detail texture analysis, geometric shape recognition and full color spectrum analysis, the image quality evaluation report comprises evaluation for image definition, exposure, noise level, color saturation and contrast, the screened image set specifically refers to an image set subjected to fuzzy image elimination, overexposure correction, multistage noise filtration and color recovery treatment, and the image annotation data set comprises object recognition, scene content analysis, emotion and context analysis and label distribution of the image.
2. The image processing system according to claim 1, wherein: the image acquisition module comprises a camera acquisition sub-module, a cloud image library synchronization sub-module, a real-time monitoring video stream processing sub-module and a format compatible conversion sub-module;
the camera acquisition submodule is used for carrying out image acquisition by adopting a CMOS image sensor and an image capturing algorithm based on multi-source input to generate a captured image data set;
the cloud image library synchronization submodule generates cloud synchronous image data by adopting a cloud storage technology and a network transmission protocol based on a captured image data set;
the real-time monitoring video stream processing sub-module is used for optimizing the video stream by adopting a real-time video processing technology and an H.264 video compression standard based on cloud synchronous image data to generate a video stream data set.
The format compatible conversion submodule carries out image format conversion and optimization by adopting an image format conversion tool and a JPEG encoding and decoding standard based on the video stream data set to generate an original image set.
3. The image processing system according to claim 2, wherein: the image preprocessing module comprises a multistage denoising processing sub-module, a dynamic contrast adjusting sub-module, a color balance correcting sub-module and an intelligent clipping adjusting sub-module;
The multistage denoising processing submodule adopts a wavelet transformation denoising technology to decompose and reconstruct signals based on an original image set, and generates denoised image data;
the dynamic contrast adjustment submodule carries out dynamic contrast adjustment on the image by adopting a local self-adaptive histogram equalization technology based on the denoised image data to generate contrast optimized image data;
the color balance correction submodule optimizes the color of the image by adopting a white balance algorithm based on the image data with optimized contrast, and generates image data with balanced color;
the intelligent clipping and adjusting submodule adopts an intelligent clipping technology based on the image data of color balance, and automatically optimizes the size and proportion of the image according to the content of the image to generate an optimized image set.
4. An image processing system according to claim 3, wherein: the feature extraction module comprises an edge detection sub-module, a texture analysis sub-module, a shape recognition sub-module and a full color spectrum analysis sub-module;
the edge detection submodule carries out edge detection by adopting a Canny operator based on the optimized image set to generate edge-enhanced image data;
the texture analysis submodule carries out texture feature analysis by adopting a gray level co-occurrence matrix based on the edge-enhanced image data to generate texture feature analysis data;
The shape recognition submodule carries out geometric shape recognition and classification by adopting Hough transformation based on texture feature analysis data to generate shape recognition data;
the full-color spectrum analysis submodule is used for carrying out color distribution and contrast analysis by adopting an RGB color model based on the shape identification data to generate an image characteristic data set.
5. The image processing system according to claim 4, wherein: the image quality evaluation module comprises a definition analysis sub-module, an intelligent exposure evaluation sub-module, a noise level detection sub-module and a color saturation quantum module;
the definition analysis submodule carries out definition analysis by using a Laplacian operator based on the image characteristic data set to generate definition analysis data;
the intelligent exposure evaluation submodule evaluates the exposure level of the image by adopting a histogram analysis method based on the definition analysis data to generate exposure evaluation data;
the noise level detection submodule is used for identifying image noise points and granularity by adopting a frequency domain analysis algorithm based on exposure evaluation data to generate noise level detection data;
the color saturation measurement submodule adopts HSV color space analysis to measure the color saturation of the image based on noise level detection data and generates an image quality assessment report.
6. The image processing system according to claim 5, wherein: the intelligent screening module comprises a blurred image elimination sub-module, an overexposure correction sub-module, a multi-stage noise filtering sub-module and a color recovery processing sub-module;
the blurred image elimination sub-module is used for identifying and eliminating blurred images by adopting a Laplace transformation algorithm based on an image quality evaluation report, and a refined image data set is generated;
the overexposure correction submodule carries out overexposure correction by adopting a dynamic range compression technology based on the thinned image data set to generate an exposure correction image set;
the multistage noise filtering submodule removes image noise points by adopting a Gaussian filtering algorithm and a bilateral filtering algorithm based on the exposure correction image set to generate a processed image set;
the color recovery processing sub-module optimizes the colors of the images by adopting a color recovery technology based on the processed image set, and generates a filtered image set.
7. The image processing system according to claim 6, wherein: the image annotation module comprises an intelligent object identification sub-module, a scene content analysis sub-module, a mood and context analysis sub-module and an automatic keyword marking sub-module;
The intelligent object identification sub-module is used for identifying main objects in the image by adopting an SSD algorithm based on the screened image set, and generating an object identification data set;
the scene content analysis submodule analyzes the scene content of the image by using a scene analysis algorithm based on the object identification data set to generate a scene analysis data set;
the emotion and context analysis submodule evaluates the emotion and the context of the image based on the scene analysis data set by applying an emotion recognition technology and a context understanding algorithm to generate an emotion context data set;
the automatic keyword marking submodule automatically extracts and marks keywords based on the emotion context data set by adopting a TF-IDF analysis technology to generate an image marking data set.
8. The image processing system according to claim 7, wherein: the image storage and retrieval module comprises an image storage sub-module, a labeling information management sub-module, an intelligent retrieval engine sub-module and a query optimization sub-module;
the image storage submodule stores image data by adopting a distributed storage technology based on an image annotation data set to generate a distributed image storage library;
the annotation information management submodule is used for maintaining and managing image annotation information by adopting MongoDB based on the distributed image storage library to generate an annotation information management library;
The intelligent search engine submodule develops an intelligent search engine by adopting an image search algorithm based on the annotation information management library to generate a search system;
the query optimization submodule adopts a query optimization technology to improve the retrieval performance based on the retrieval system and generates an image database.
9. An image processing method, characterized in that the image processing system according to any of claims 1-8 is executed, comprising the steps of:
based on multi-source input, acquiring an original image by adopting a CMOS image sensor and an image capturing algorithm to generate an original image set;
based on the original image set, performing image preprocessing by adopting a wavelet transformation denoising technology and local self-adaptive histogram equalization to generate an optimized image set;
based on the optimized image set, feature extraction is carried out by adopting a ResNet model and an image analysis technology, and an image feature data set is generated;
based on the image characteristic data set, performing image quality evaluation by adopting a peak signal-to-noise ratio algorithm, and generating an image quality evaluation report;
based on the image quality evaluation report, intelligent screening is carried out by adopting a support vector machine intelligent classification algorithm and a pattern recognition technology, and a screened image set is generated;
Based on the screened image set, carrying out image annotation by adopting a natural language processing technology based on a transducer model, and generating an image annotation data set;
based on the image annotation data set, performing image storage and retrieval by adopting a distributed NoSQL database and an image retrieval algorithm to generate an image database;
based on the image database, adopting MongoDB technology to manage and optimize the annotation information, and generating an annotation information management library;
and based on the annotation information management library, searching by adopting a query optimization technology and an intelligent search engine to generate a comprehensive query optimization and search system.
10. The image processing method according to claim 9, characterized in that: the optimized image set specifically refers to an image subjected to denoising and contrast adjustment, the image data set comprises edge, texture, shape and color information of the image, the image quality evaluation report comprises evaluation of definition, exposure, noise, color and contrast, the screened image set specifically refers to an image without blurring and moderate exposure, and the image annotation data set comprises an image description text and a keyword label.
CN202311651983.0A 2023-12-04 2023-12-04 Image processing system and image processing method Withdrawn CN117475121A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202311651983.0A CN117475121A (en) 2023-12-04 2023-12-04 Image processing system and image processing method

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202311651983.0A CN117475121A (en) 2023-12-04 2023-12-04 Image processing system and image processing method

Publications (1)

Publication Number Publication Date
CN117475121A true CN117475121A (en) 2024-01-30

Family

ID=89629406

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202311651983.0A Withdrawn CN117475121A (en) 2023-12-04 2023-12-04 Image processing system and image processing method

Country Status (1)

Country Link
CN (1) CN117475121A (en)

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN117788468A (en) * 2024-02-26 2024-03-29 江西福松和安医疗科技有限公司 Laryngeal image processing method, laryngeal image processing system, laryngoscope and adjustable airway establishing device
CN117994160A (en) * 2024-04-01 2024-05-07 南昌理工学院 Image processing method and system
CN117994160B (en) * 2024-04-01 2024-06-04 南昌理工学院 Image processing method and system

Cited By (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN117788468A (en) * 2024-02-26 2024-03-29 江西福松和安医疗科技有限公司 Laryngeal image processing method, laryngeal image processing system, laryngoscope and adjustable airway establishing device
CN117788468B (en) * 2024-02-26 2024-04-30 江西福松和安医疗科技有限公司 Laryngeal image processing method, laryngeal image processing system, laryngoscope and adjustable airway establishing device
CN117994160A (en) * 2024-04-01 2024-05-07 南昌理工学院 Image processing method and system
CN117994160B (en) * 2024-04-01 2024-06-04 南昌理工学院 Image processing method and system

Similar Documents

Publication Publication Date Title
Wiley et al. Computer vision and image processing: a paper review
Pape et al. 3-D histogram-based segmentation and leaf detection for rosette plants
Wang et al. Detection and counting of immature green citrus fruit based on the Local Binary Patterns (LBP) feature using illumination-normalized images
JP5941674B2 (en) Cell contour forming apparatus and method, and cell contour forming program
CN104463195A (en) Printing style digital recognition method based on template matching
CN111797653A (en) Image annotation method and device based on high-dimensional image
CN117475121A (en) Image processing system and image processing method
CN107704878B (en) Hyperspectral database semi-automatic establishment method based on deep learning
KR102027878B1 (en) Method for recognizing art objects in video combining deep learning technology and image feature extraction technology
WO2013030850A2 (en) Chemical structure recognition tool
CN109977899B (en) Training, reasoning and new variety adding method and system for article identification
Zhang et al. Research on the correlation between image enhancement and underwater object detection
CN113129236A (en) Single low-light image enhancement method and system based on Retinex and convolutional neural network
CN113409355A (en) Moving target identification system and method based on FPGA
CN202815869U (en) Vehicle microcomputer image and video data extraction apparatus
CN116823694B (en) Infrared and visible light image fusion method and system based on multi-focus information integration
Wu et al. Review of imaging device identification based on machine learning
CN112258453A (en) Positioning landmark detection method for industrial fault inspection robot
CN112308162A (en) Image big data similarity comparison method and system
Liu et al. Deseal: Semantic-aware seal2clear attention for document seal removal
Kurapati et al. A Review: Text Extraction from Stone Inscriptions and Translating to Modern Language
CN114926761B (en) Action recognition method based on space-time smoothing characteristic network
CN118014887B (en) Intelligent folding house remote monitoring method based on Internet of things
CN111191550B (en) Visual perception device and method based on automatic dynamic adjustment of image sharpness
Rajaboina et al. An Empirical Review for Character Recognition in Video Frames using Text Detection Algorithms

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
WW01 Invention patent application withdrawn after publication
WW01 Invention patent application withdrawn after publication

Application publication date: 20240130