CN118196310B - Image processing method for image recognition calculation - Google Patents
Image processing method for image recognition calculation Download PDFInfo
- Publication number
- CN118196310B CN118196310B CN202410615990.3A CN202410615990A CN118196310B CN 118196310 B CN118196310 B CN 118196310B CN 202410615990 A CN202410615990 A CN 202410615990A CN 118196310 B CN118196310 B CN 118196310B
- Authority
- CN
- China
- Prior art keywords
- image
- specific structural
- structural features
- images
- feature
- 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.)
- Active
Links
- 238000004364 calculation method Methods 0.000 title claims abstract description 19
- 238000003672 processing method Methods 0.000 title claims abstract description 17
- 238000012545 processing Methods 0.000 claims abstract description 47
- 230000000007 visual effect Effects 0.000 claims abstract description 18
- 238000004458 analytical method Methods 0.000 claims description 30
- 238000000034 method Methods 0.000 claims description 27
- 239000013598 vector Substances 0.000 claims description 26
- 238000013507 mapping Methods 0.000 claims description 16
- 238000005457 optimization Methods 0.000 claims description 10
- 230000036544 posture Effects 0.000 claims description 9
- 238000012216 screening Methods 0.000 claims description 9
- 239000000203 mixture Substances 0.000 claims description 4
- 238000005516 engineering process Methods 0.000 abstract description 7
- 239000007787 solid Substances 0.000 abstract description 6
- 238000010276 construction Methods 0.000 abstract description 4
- 230000008569 process Effects 0.000 description 8
- 230000008901 benefit Effects 0.000 description 3
- 238000004422 calculation algorithm Methods 0.000 description 2
- 230000008859 change Effects 0.000 description 2
- 238000006243 chemical reaction Methods 0.000 description 2
- 238000001514 detection method Methods 0.000 description 2
- 238000000605 extraction Methods 0.000 description 2
- 230000010354 integration Effects 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 230000001360 synchronised effect Effects 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000015572 biosynthetic process Effects 0.000 description 1
- 239000003086 colorant Substances 0.000 description 1
- 238000004590 computer program Methods 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 238000011156 evaluation Methods 0.000 description 1
- 239000000284 extract Substances 0.000 description 1
- 230000006870 function Effects 0.000 description 1
- 238000010921 in-depth analysis Methods 0.000 description 1
- 230000007246 mechanism Effects 0.000 description 1
- 230000003278 mimic effect Effects 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 230000003068 static effect Effects 0.000 description 1
- 238000003786 synthesis reaction Methods 0.000 description 1
- 230000009466 transformation Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T17/00—Three dimensional [3D] modelling, e.g. data description of 3D objects
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T19/00—Manipulating 3D models or images for computer graphics
- G06T19/20—Editing of 3D images, e.g. changing shapes or colours, aligning objects or positioning parts
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/70—Determining position or orientation of objects or cameras
- G06T7/77—Determining position or orientation of objects or cameras using statistical methods
-
- 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
- G06V10/16—Image acquisition using multiple overlapping images; Image stitching
-
- 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
- G06V10/32—Normalisation of the pattern dimensions
-
- 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
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Multimedia (AREA)
- Computer Graphics (AREA)
- Software Systems (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Evolutionary Biology (AREA)
- Probability & Statistics with Applications (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Architecture (AREA)
- Life Sciences & Earth Sciences (AREA)
- Computer Hardware Design (AREA)
- General Engineering & Computer Science (AREA)
- Geometry (AREA)
- Image Analysis (AREA)
Abstract
The invention is applicable to the field of image recognition, and provides an image processing method used in image recognition calculation, which comprises the following steps: identifying a preset specific structural feature in a plurality of images; calculating the view angle of the specific structural feature according to the spatial relationship; keeping the specific structural feature size in each picture consistent; performing position processing on all the images; building a three-dimensional space; optimizing and adjusting the joint of the specific structural feature in the three-dimensional space; and outputting the three-dimensional images. The consistency and the accurate alignment of the same specific structural features in different images are ensured through a refined position processing technology, and the image frames are intelligently cut, so that the consistency among the images is enhanced, the precision of the subsequent three-dimensional construction stage is improved, the visual effect is optimized, the user can observe and analyze more conveniently, the efficiency and the practicability of the whole image processing method are remarkably improved, and a solid foundation is laid for reconstructing a high-quality three-dimensional model.
Description
Technical Field
The invention belongs to the field of image recognition, and particularly relates to an image processing method used in image recognition calculation.
Background
The field of computer vision and image processing relates to techniques that enable a computer to interpret and understand image and video data from cameras and other optical sensors. Research in this area is extensive, including aspects of image acquisition, analysis, processing, and understanding, to mimic the functions of the human visual system, enabling the computer to perform visual tasks such as recognition, tracking, and classification.
At present, when three-dimensional modeling is performed through images, if multi-angle shooting of the same equipment is adopted, the time required is too complicated, and if multi-ginseng and shooting of different equipment are adopted, as images come from a plurality of mobile equipment, camera parameters and image capturing settings of each equipment may be different, so that the received images have differences in size. The same specific structural features in each image may be located differently in the image due to differences in the angle and distance of capture, which presents challenges for subsequent image alignment and three-dimensional reconstruction.
Disclosure of Invention
The invention aims to provide an image processing method used in image recognition calculation, which aims to solve the technical problems in the prior art determined in the background art.
The present invention is embodied in a method of image processing for use in image recognition computation, the method comprising:
Respectively receiving images from a plurality of mobile devices, and respectively identifying preset specific structural features in the images;
carrying out angle analysis on specific structural features in each image, deducing the attitude angle of the specific structural features in each image relative to a camera, and calculating the view angle of the specific structural features according to the spatial relationship;
identifying and recording the image sizes of the specific structural features with the same angle in all the images according to the view angle relation of the specific structural features, calculating the intermediate value size, carrying out size processing on all the images according to the intermediate value size, and keeping the specific structural feature sizes in each image consistent through the association degree between the specific structural features with different angles;
recognizing the coordinates of specific structural features in the image after the size processing in the image, and carrying out position processing on all the images by taking the midpoint of the image size as an origin;
Building a three-dimensional space, mapping specific structural features in each image into the three-dimensional space, and detecting the alignment degree of the specific structural features in the three-dimensional space;
identifying structural texture information of specific structural features in the image, and optimally adjusting joints of the specific structural features in the three-dimensional space according to the identified texture information;
And fusing the images in the three-dimensional space and outputting the fused images into a three-dimensional image.
As a further aspect of the present invention, the receiving images from a plurality of mobile devices respectively, and identifying preset specific structural features in the plurality of images respectively specifically includes:
scanning and analyzing the preset image, reading all characteristic attributes in the preset image, and screening out existing specific structural characteristics;
Respectively receiving images from a plurality of mobile devices, analyzing the received images, and extracting all characteristic information contained in the images;
screening all the characteristic information to obtain images all containing specific structural characteristics;
And analyzing the image containing the specific structural features, judging the definition of the specific structural features in the image, and deleting the image with the definition lower than the identification standard.
As a further aspect of the present invention, the performing an angle analysis on the specific structural feature in each image after the size and the position processing specifically includes:
analyzing the specific structural features in the provided image, and identifying the positions and the shapes of the specific structural features in the image;
based on the result of the feature analysis, comparing the specific structural features with known attitude angle information, and judging the relative angles of the specific structural features in the image;
Based on the inferred relative angles, in combination with known camera parameters and feature location information, the viewing angle of the particular structural feature in the camera coordinate system is calculated.
As a further aspect of the present invention, the identifying and recording the image size of the specific structural feature in all the images, calculating an intermediate value size, and performing size processing on all the images according to the intermediate value size, specifically includes:
based on the results of the angle analysis, the dimensions of the specific structural features at the same viewing angle are identified in all images and recorded;
Calculating the size intermediate value of the specific structural feature under the view angle according to the recorded size of the specific structural feature in different images;
Based on the intermediate value size, the specific structural feature sizes in the images of all the same view angles are adjusted to be the intermediate value size, the specific structural feature sizes of different view angles are analyzed after adjustment is completed, the association degree between the specific structural features of different angles is compared, and the sizes of the specific structural features are secondarily adjusted.
As a further aspect of the present invention, the identifying coordinates of the specific structural feature in the image after the size processing, and performing the position processing on all the images with the midpoint of the image size as the origin, specifically includes:
analyzing the images with the processed sizes, taking the whole images as a coordinate system, taking the points in the images as the origin of coordinates, and establishing a coordinate system with the same size for each image, wherein the coordinate system contains complete specific structural features;
recognizing coordinate values of specific structural features in each image in an image coordinate system, and moving the image to an origin position by taking the coordinate values as a moving center;
And identifying the distance between the frame of each image and the frame of the coordinate system, cutting all the images by taking the farthest distance of each frame as a reference, and obtaining a final image with specific structural features at the middle point of the image and the same size of each image.
As a further aspect of the present invention, the mapping the specific structural feature in each image to the three-dimensional space, and detecting the alignment degree of the specific structural feature in the three-dimensional space, specifically includes:
creating a virtual three-dimensional space in a three-dimensional coordinate system based on the results of the angle analysis;
For specific structural features in each image, combining the results of angle analysis, and mapping the positions of the specific structural features in the image coordinate system to corresponding positions in a three-dimensional space according to the visual angle information;
In the three-dimensional space, detecting the alignment degree between specific structural features mapped to the three-dimensional space by different images, and carrying out secondary adjustment on the positions and the postures of the specific structural features.
As a further aspect of the present invention, the optimizing and adjusting the joint of the specific structural feature in the three-dimensional space specifically includes:
Analyzing and extracting textures of specific structural features and surrounding areas in the image;
Detecting smoothness of joints of the specific structural features in the three-dimensional space according to the mapping of the specific structural features in the three-dimensional space and in combination with the identified texture information;
and optimizing and adjusting the position, the posture and the shape of the joint according to the texture information of the joint and the positions of the adjacent specific structural features.
As a further aspect of the present invention, the fusing of images in a three-dimensional space and outputting the fused images into a three-dimensional stereoscopic image specifically includes:
and according to the position and texture information of the specific structural features in the three-dimensional space, overlapping and fusing the images subjected to optimization and adjustment in the three-dimensional space to form a unified three-dimensional image.
The beneficial effects of the invention are as follows:
the consistency and the accurate alignment of the same specific structural features in different images are ensured through a refined position processing technology, and the image frames are intelligently cut, so that the consistency among the images is enhanced, the precision of the subsequent three-dimensional construction stage is improved, the visual effect is optimized, the user can observe and analyze more conveniently, the efficiency and the practicability of the whole image processing method are remarkably improved, and a solid foundation is laid for reconstructing a high-quality three-dimensional model.
Drawings
FIG. 1 is a flowchart of an image processing method for use in image recognition computation according to an embodiment of the present invention;
FIG. 2 is a flowchart of an embodiment of the present invention for respectively receiving images from a plurality of mobile devices and respectively identifying preset specific structural features in the plurality of images;
FIG. 3 is a flow chart of an embodiment of the present invention for performing an angular analysis on a specific structural feature in each image;
FIG. 4 is a flowchart of identifying and recording image sizes of specific structural features at the same angle in all images according to the view angle relationship of the specific structural features according to an embodiment of the present invention;
FIG. 5 is a flowchart of a position process for all images according to an embodiment of the present invention;
FIG. 6 is a flow chart of mapping specific structural features in each image into a three-dimensional space, respectively, for building the three-dimensional space according to an embodiment of the present invention;
fig. 7 is a flowchart for identifying structural texture information of a specific structural feature in an image according to an embodiment 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.
It will be understood that the terms "first," "second," and the like, as used herein, may be used to describe various elements, but these elements are not limited by these terms unless otherwise specified. These terms are only used to distinguish one element from another element. For example, a first xx script may be referred to as a second xx script, and similarly, a second xx script may be referred to as a first xx script, without departing from the scope of this disclosure.
Fig. 1 is a flowchart of an image processing method for use in image recognition computation according to an embodiment of the present invention, as shown in fig. 1, the method includes:
s100, respectively receiving images from a plurality of mobile devices, and respectively identifying preset specific structural features in the images;
The system scans the preset image, identifies and extracts all characteristic attributes, and screens out information containing specific structural characteristics. Next, the system receives and parses the images transmitted by different mobile devices, applying a multi-layer feature extraction strategy, not only extracting basic features such as colors, textures, but also understanding higher level semantic content. And then, classifying and comparing all the characteristic information through an intelligent screening mechanism to ensure that only images containing specific structural characteristics are left, thereby reducing irrelevant data processing and improving efficiency. And finally, carrying out definition evaluation on the screened images, and eliminating the images which do not meet the quality standard.
This tandem procedure shows a significant advance in the automation, intelligent processing of the system, which can accommodate a variety of equipment and shooting environments, ensure real-time performance and have continuous optimization capabilities. In addition, by ensuring high quality of the input data, the S100 step lays a solid foundation for the accuracy and reliability of the entire system.
S200, carrying out angle analysis on specific structural features in each image, deducing the attitude angle of the specific structural features in each image relative to a camera, and calculating the view angles of the specific structural features according to the spatial relationship;
This step is mainly responsible for deducing the pose angle of a specific structural feature in each image relative to the camera. The step accurately calculates the observation angles of the specific structural features in the camera coordinate system by deeply analyzing the angles of the specific structural features, and brings remarkable advantages for the whole image processing flow. First, accurate angular analysis enhances the system's ability to understand three-dimensional space, providing a solid foundation for subsequent three-dimensional mapping. Second, by normalizing the viewing angle of the same feature between different images, the S200 step ensures consistency of image size in further processing, which is critical for constructing high quality three-dimensional stereoscopic images.
In addition, the step optimizes the alignment of the features in the three-dimensional space and the adjustment of the joints of the subsequent texture information by accurately identifying the observation angles of the features, thereby remarkably improving the accuracy of three-dimensional reconstruction. The accurate angle analysis also improves the efficiency of the whole image processing flow, particularly reduces the need of repeated adjustment when processing a large amount of image data, and accelerates the processing speed.
In summary, the step S200 provides powerful support for generating accurate and consistent three-dimensional images through high attention and accurate calculation of the angle details.
S300, identifying and recording the image sizes of the specific structural features with the same angle in all images according to the view angle relation of the specific structural features, calculating the intermediate value size, carrying out size processing on all images according to the intermediate value size, and keeping the specific structural feature sizes in each image consistent through the association degree between the specific structural features with different angles;
This step uses the results of the angular analysis of the specific structural features obtained from step S200 to perform a series of accurate dimensional processing operations to ensure consistency and accuracy of the specific structural features during the three-dimensional model reconstruction process.
The system first uses the angle analysis results of step S200 to identify and record the dimensions of specific structural features in all images at the same viewing angle. In this way, the system can build a detailed database of the same feature sizes in different images, providing data support for further size processing.
The system then uses the collected data to calculate intermediate values for the dimensions of the particular structural features at each particular viewing angle. This median size is calculated based on a large amount of data and thus can represent the average size of the structural feature in most cases, providing a reliable basis for unifying size criteria.
Finally, the system performs a size adjustment of specific structural features of the same view angle in all images based on the intermediate size, ensuring that the sizes of the same features in different images are consistent. In addition, the system can analyze the association degree between the structural features under different view angles so as to ensure that the dimensional adjustment is performed, and meanwhile, the correlation of the specific structural features under different angles can be considered, so that the dimensional fine secondary adjustment is performed. This secondary correlation-based adjustment ensures that the dimensions of a particular structural feature remain consistent and accurate from any angle of view.
The significant advancement and advantage of the step S300 is that it ensures a high degree of accuracy and consistency in the three-dimensional reconstruction process. Through accurate size processing, the step S300 effectively eliminates the characteristic size difference of specific structures caused by visual angle change, and improves the quality of the three-dimensional model after different images are fused. In addition, this step optimizes the visual effect of the three-dimensional model so that the particular structural features of the model remain highly accurate and natural when viewed from any angle. In practical application, the method not only improves user experience, but also greatly enhances the application value and practicability of the three-dimensional reconstruction technology in various scenes.
S400, recognizing coordinates of specific structural features in the image after the size processing in the image, and performing position processing on all the images by taking the middle point of the image size as an origin;
The step ensures the consistency and accurate alignment of the same specific structural features in different images through a fine position processing technology, and provides a solid foundation for the subsequent three-dimensional space reconstruction. The step firstly analyzes the images with the processed sizes, and establishes a coordinate system with uniform sizes by taking the center of the images as the origin, thereby ensuring the integrity and consistency of specific structural features in all the images. Then, by precisely locating the coordinates of the specific structural features in each image and moving these features to the origin of coordinates based thereon, precise alignment of features between different images is achieved.
In addition, S400 involves the step of intelligently cropping the image frames to ensure uniformity in size of all images while leaving the specific structural feature in the center of the image. The processing means not only strengthen the consistency among images and improve the precision of the subsequent three-dimensional construction stage, but also optimize the visual effect, so that the user is more convenient in observation and analysis.
Through careful position processing, S400 remarkably improves the efficiency and practicality of the whole image processing method, and lays a solid foundation for reconstructing a high-quality three-dimensional model.
S500, building a three-dimensional space, mapping specific structural features in each image into the three-dimensional space, and detecting the alignment degree of the specific structural features in the three-dimensional space;
The present step involves a stereo reconstruction in three-dimensional space, ensuring that multi-source images captured from various angles and positions can be aligned and fused in three-dimensional space accurately. Successful execution of this step has a decisive influence on the generation of high-precision and high-quality three-dimensional models.
In step S500, a finely calibrated three-dimensional virtual coordinate system is first created, which is the basis for all subsequent mapping operations. The virtual three-dimensional space considers the factors such as proportion sense, visual angle and light in the real world, and ensures the authenticity and reliability of the three-dimensional model.
Then, the two-dimensional coordinates in the picture are converted into coordinates in the three-dimensional space by the posture angle of the specific structural feature in each image inferred in the previous step S200. The conversion process not only considers the azimuth and the angle of the specific structural characteristics, but also fully utilizes the visual angle information to adjust the relative position of the specific structure in the three-dimensional space, so that images captured from different angles can be correctly butted.
And finally, accurately aligning, detecting and adjusting the specific structural features in the three-dimensional space. If the specific structural features mapped by different images are found to have tiny dislocation or angle deviation, the system can finely adjust the positions and the attitudes of the features according to the requirements until all the features are perfectly aligned in the three-dimensional space. The alignment is not only visual, but also structural, meaning that the three-dimensional model can exactly match the original physical world in terms of shape, size, scale, mutual position, etc.
This step allows for efficient integration of different perspective images from multiple mobile devices such that the construction of the three-dimensional model is no longer limited to a single perspective or location. The multi-angle image integration technology remarkably improves the accuracy, richness and fidelity of the three-dimensional model. And, through advanced algorithm, S500 step has improved processing speed and efficiency by a wide margin, has shortened the conversion time from the image to three-dimensional model. This provides great convenience for various fields such as virtual reality, game development, architectural design, etc., and can quickly and accurately convert real-world objects and scenes into the digital world, providing users with the potential for immersive experience and in-depth analysis.
S600, identifying structural texture information of specific structural features in the image, and optimally adjusting joints of the specific structural features in the three-dimensional space according to the identified texture information;
The method optimizes the joint of specific structural features in the three-dimensional model through detailed texture analysis and adjustment, so that the visual sense realism of the final three-dimensional image is remarkably improved. The method comprises the steps of firstly carrying out deep analysis and extraction on the specific structural features in the image and the textures of the surrounding areas, and precisely capturing the details of the textures by using an advanced image processing algorithm, wherein the details comprise the identification of dynamic textures, thereby laying a foundation for subsequent smoothness detection and optimization adjustment. Then, the system evaluates the smoothness of the joint in the three-dimensional space according to the texture information, and discovers and marks the area of the texture discontinuity or fracture by using the technologies of texture continuity check and the like. Then, by performing precise optimization adjustment on the areas, such as texture transformation, deformation processing or complex image synthesis technology, the visual consistency of the joint area is ensured, and the splicing trace is effectively eliminated. The process not only greatly improves the visual sense of reality of the model, but also saves a great amount of time and labor cost due to high automation, shows high adaptability and effective processing capacity for dynamic textures, and can ensure high-quality output even at the juncture of the textures and the complex structures.
Through the application of the step S600, the three-dimensional image processing achieves a new height in terms of efficiency, accuracy, quality and sense of reality.
S700, fusing the images in the three-dimensional space and outputting the fused images into a three-dimensional image.
Fig. 2 is a flowchart of receiving images from a plurality of mobile devices and identifying preset specific structural features in the plurality of images, respectively, according to an embodiment of the present invention, as shown in fig. 2, where the receiving images from the plurality of mobile devices and identifying preset specific structural features in the plurality of images respectively specifically includes:
S110, scanning and analyzing a preset image, reading all characteristic attributes in the preset image, and screening out existing specific structural characteristics;
It should be noted here that, as for the preset image, the preset image here is not an image received by the mobile device, but a standard image set in advance. It can be understood that by performing scanning analysis on a preset standard image and further performing overall feature attribute analysis, specific structural features are screened out for subsequent comparison with feature information in an image received by the mobile device.
S120, respectively receiving images from a plurality of mobile devices, analyzing the received images, and extracting all characteristic information contained in the images;
s130, screening all the characteristic information to screen out images all containing specific structural characteristics;
the step S130 specifically includes the following steps:
Step S1301, a feature code is built for a specific structural feature in a preset image to obtain a preset feature code;
Step S1302, extracting, for each image received from the mobile device, a corresponding feature code to obtain an image feature code;
step S1303, performing similarity calculation on the image feature codes and the preset feature codes to obtain feature similarity;
The calculation formula of the feature similarity is expressed as follows:
；
Wherein, Represent the firstFeature similarity of the individual image feature codes to the preset feature codes,Representing the first predetermined feature codeThe number of components of the composition,Represent the firstCoding of individual image featuresThe number of components of the composition,Representing the dimension of the feature code.
It should be noted that, for the feature similarityIts value ranges from-1 to 1, with a value closer to 1 indicating that the two features are more similar, a value closer to-1 indicating that the two features are more different, and a value of 0 indicating that the two features are completely uncorrelated. Coding of preset featuresIndividual componentsThe preset feature is a quantized numerical representation of the specific structural feature to be identified, and each component of the vector represents an attribute of the feature.And (3) withAnd respectively representing the preset feature vector and the modulo (length) of the image feature vector, and normalizing the feature vector to ensure that the similarity calculation is more accurate.
In step S1304, when it is determined that the feature similarity is greater than the similarity threshold, it is determined that the image contains a specific structural feature.
And S140, analyzing the image containing the specific structural features, judging the definition of the specific structural features in the image, and deleting the image with the definition lower than the identification standard.
Fig. 3 is a flowchart of performing angle analysis on specific structural features in each image according to an embodiment of the present invention, and as shown in fig. 3, the performing angle analysis on specific structural features in each image after size and position processing specifically includes:
s210, analyzing the specific structural features in the provided image, and identifying the positions and the shapes of the specific structural features in the image;
S220, comparing the specific structural features with known attitude angle information based on the result of the feature analysis, and judging the relative angles of the specific structural features in the image;
In step S220, the calculation formula of the relative angle of the specific structural feature in the image is expressed as:
；
Wherein, Representing the relative angle of a particular structural feature in the image,Feature vectors representing particular structural features extracted from the image,The reference vector is represented by a reference vector,Representing feature vectorsIs provided with a die for the mold,Representing reference vectorsIs a mold of (a).
It should be noted that, for the feature vectorThe direction and size representing a particular structural feature extracted from an image is a directional and length quantity, typically represented as a multi-dimensional vector.
S230, according to the deduced relative angle, combining the known camera parameters and the feature position information, calculating the observation angle of the specific structural feature in the camera coordinate system.
Wherein, the calculation formula of the observation angle of the specific structural feature in the camera coordinate system is expressed as follows:
；
Wherein, Representing the viewing angle of a particular structural feature in the camera coordinate system,Representing the vertical offset of the feature point relative to the center of the image,Representing the horizontal offset of the feature point relative to the center of the image,Representing the adjustment coefficient.
It should be noted here that for the adjustment factorFor fine tuning the viewing angle according to specific parameters of the camera, such as focal length, pixel size, etc.Representing the angles (in radians) used to calculate the feature points in the image coordinate system relative to the center of the image.Representing the relative angle of a particular structural feature in an imageNormalized to the range of [0, 2 ].
Fig. 4 is a flowchart of identifying and recording image sizes of specific structural features at the same angle in all images according to a view angle relationship of the specific structural features provided in an embodiment of the present invention, as shown in fig. 4, where the identifying and recording image sizes of the specific structural features in all images, calculating an intermediate value size, and performing size processing on all images according to the intermediate value size, specifically includes:
S310, based on the result of the angle analysis, identifying the size of the specific structural feature under the same viewing angle in all images and recording the same;
S320, calculating a size intermediate value of the specific structural feature under the view angle according to the recorded size of the specific structural feature in different images;
Wherein, the calculation formula of the size intermediate value of the specific structural feature is expressed as:
；
Wherein, Intermediate values of dimensions representing features of a particular structure,Representing the first in an imageThe dimensions of the individual features of the particular structure,Representing the first in an imageThe weight corresponding to the size of each particular structural feature,Representing the total amount of a particular structural feature in the image.
S330, based on the intermediate value size, the specific structural feature sizes in the images of all the same view angles are adjusted to the intermediate value size, the specific structural feature sizes of different view angles are analyzed after the adjustment is completed, the association degree between the specific structural features of different angles is compared, and the sizes of the specific structural features are secondarily adjusted.
Fig. 5 is a flowchart of performing position processing on all images according to an embodiment of the present invention, as shown in fig. 5, where coordinates of a specific structural feature in an image after the size processing are identified, and a midpoint of the image size is taken as an origin, and performing position processing on all images, where the method specifically includes:
S410, analyzing the images with the processed sizes, taking the whole images as a coordinate system, taking the middle points of the images as the origin of coordinates, and establishing a coordinate system with the same size for each image, wherein the coordinate system contains complete specific structural features;
s420, recognizing coordinate values of specific structural features in each image in an image coordinate system, and moving the image to an origin position by taking the coordinate values as a movement center;
s430, identifying the distance between the frame of each image and the frame of the coordinate system, cutting all the images by taking the farthest distance of each frame as a reference, and obtaining a final image with specific structural features at the middle point of the image and the same size of each image.
Fig. 6 is a flowchart of mapping specific structural features in each image to a three-dimensional space, and as shown in fig. 6, the mapping of the specific structural features in each image to the three-dimensional space and the detection of the alignment degree of the specific structural features in the three-dimensional space specifically include:
S510, creating a virtual three-dimensional space in a three-dimensional coordinate system based on the result of the angle analysis;
S520, mapping the position of the specific structural features in each image in an image coordinate system to the corresponding position in the three-dimensional space according to the visual angle information by combining the results of the angle analysis;
in this step, the expression of the coordinate position of the specific structural feature in the three-dimensional space is:
；
Wherein, Respectively represent the firstThe abscissa and ordinate of each specific structural feature in the three-dimensional space in the horizontal direction,Represent the firstThe coordinates of the individual specific structural features in the vertical direction,Representing the abscissa and ordinate respectively of a particular structural feature in a two-dimensional image coordinate system,Represent the firstThe relative angle of the individual specific structural features in the image,Represent the firstThe azimuth of each particular structural feature relative to the camera,Represent the firstPitch angle of the specific structural feature relative to the camera.
Here, the azimuth angle of the camera is describedThe rotation angle of the camera on the horizontal plane is represented and reflects the deflection of the shooting direction of the camera relative to a certain fixed reference (such as the north direction). Pitch angle of cameraRepresenting the tilt angle of the camera relative to the horizontal, which is 0 if the camera is placed perfectly horizontally; if the camera is tilted up or down, this angle will change accordingly.
And S530, detecting the alignment degree between specific structural features of different images mapped to the three-dimensional space in the three-dimensional space, and carrying out secondary adjustment on the positions and the postures of the specific structural features.
Fig. 7 is a flowchart for identifying structural texture information of a specific structural feature in an image according to an embodiment of the present invention, and as shown in fig. 7, the optimizing adjustment of a seam of the specific structural feature in a three-dimensional space specifically includes:
s610, analyzing and extracting textures of specific structural features and surrounding areas in the image;
S620, detecting smoothness of joints of the specific structural features in the three-dimensional space according to the mapping of the specific structural features in the three-dimensional space and in combination with the identified texture information;
And S630, optimizing and adjusting the position, the posture and the shape of the joint according to the texture information of the joint and the positions of the adjacent specific structural features.
In step S630, in the step of optimizing and adjusting the position, posture and shape of the joint, the calculation formula for determining the optimized joint position is expressed as:
；
Wherein, Indicating the position of the joint after the optimization,Indicating the initial position of the seam,The weight factor is represented by a weight factor,A gradient representing the degree of texture variance,Representing vectors between adjacent specific structural features,Representing the distance between adjacent specific structural features,Representing the geometric vector of the joint.
It should be noted that, for the weight factorThe method is used for adjusting contributions of different influencing factors to joint position optimization, and can be adjusted according to specific application requirements and scenes so as to find the optimal joint position. Gradient for texture variabilityIs a vector pointing in the fastest direction where texture differences decrease. During the optimization process, the seam can be moved in this direction to reduce texture discontinuities. For vectors between adjacent specific structural featuresThe vector points to the nearest particular structural feature and by taking into account the vector, the seam location can be adjusted to a location that is more coordinated with the surrounding features. Geometric vector for jointThe shape of the joint can be adjusted to be smoother and more natural by considering the vector, which can be defined according to geometric characteristics such as curvature and length of the joint.
In addition, the fusing the images in the three-dimensional space and outputting the fused images into a three-dimensional stereoscopic image specifically includes:
and according to the position and texture information of the specific structural features in the three-dimensional space, overlapping and fusing the images subjected to optimization and adjustment in the three-dimensional space to form a unified three-dimensional image.
It should be understood that, although the steps in the flowcharts of the embodiments of the present invention are shown in order as indicated by the arrows, these steps are not necessarily performed in order as indicated by the arrows. The steps are not strictly limited to the order of execution unless explicitly recited herein, and the steps may be executed in other orders. Moreover, at least some of the steps in various embodiments may include multiple sub-steps or stages that are not necessarily performed at the same time, but may be performed at different times, nor do the order in which the sub-steps or stages are performed necessarily performed in sequence, but may be performed alternately or alternately with at least a portion of the sub-steps or stages of other steps or other steps.
Those skilled in the art will appreciate that all or part of the processes in the methods of the above embodiments may be implemented by a computer program for instructing relevant hardware, where the program may be stored in a non-volatile computer readable storage medium, and where the program, when executed, may include processes in the embodiments of the methods described above. Any reference to memory, storage, database, or other medium used in embodiments provided herein may include non-volatile and/or volatile memory. The nonvolatile memory can include Read Only Memory (ROM), programmable ROM (PROM), electrically Programmable ROM (EPROM), electrically Erasable Programmable ROM (EEPROM), or flash memory. Volatile memory can include Random Access Memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in a variety of forms such as Static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double Data Rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link (SYNCHLINK) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), among others.
The technical features of the above-described embodiments may be arbitrarily combined, and all possible combinations of the technical features in the above-described embodiments are not described for brevity of description, however, as long as there is no contradiction between the combinations of the technical features, they should be considered as the scope of the description.
The foregoing examples illustrate only a few embodiments of the invention and are described in detail herein without thereby limiting the scope of the invention. It should be noted that it will be apparent to those skilled in the art that several variations and modifications can be made without departing from the spirit of the invention, which are all within the scope of the invention. Accordingly, the scope of protection of the present invention is to be determined by the appended claims.
The foregoing description of the preferred embodiments of the invention is not intended to be limiting, but rather is intended to cover all modifications, equivalents, and alternatives falling within the spirit and principles of the invention.
Claims (9)
1. An image processing method for use in image recognition computation, the method comprising the steps of:
Respectively receiving images from a plurality of mobile devices, and respectively identifying preset specific structural features in the images;
carrying out angle analysis on specific structural features in each image, deducing the attitude angle of the specific structural features in each image relative to a camera, and calculating the view angle of the specific structural features according to the spatial relationship;
identifying and recording the image sizes of the specific structural features with the same angle in all the images according to the view angle relation of the specific structural features, calculating the intermediate value size, carrying out size processing on all the images according to the intermediate value size, and keeping the specific structural feature sizes in each image consistent through the association degree between the specific structural features with different angles;
recognizing the coordinates of specific structural features in the image after the size processing in the image, and carrying out position processing on all the images by taking the midpoint of the image size as an origin;
Building a three-dimensional space, mapping specific structural features in each image into the three-dimensional space, and detecting the alignment degree of the specific structural features in the three-dimensional space;
identifying structural texture information of specific structural features in the image, and optimally adjusting joints of the specific structural features in the three-dimensional space according to the identified texture information;
fusing the images in the three-dimensional space and outputting the fused images into a three-dimensional image;
The method for respectively receiving the images from the plurality of mobile devices and respectively identifying the preset specific structural features in the plurality of images specifically comprises the following steps:
scanning and analyzing the preset image, reading all characteristic attributes in the preset image, and screening out existing specific structural characteristics;
Respectively receiving images from a plurality of mobile devices, analyzing the received images, and extracting all characteristic information contained in the images;
screening all the characteristic information to obtain images all containing specific structural characteristics;
And analyzing the image containing the specific structural features, judging the definition of the specific structural features in the image, and deleting the image with the definition lower than the identification standard.
2. The image processing method for use in image recognition computation according to claim 1, wherein the method of screening out images all including specific structural features by screening out all feature information comprises the steps of:
Establishing feature codes for specific structural features in a preset image to obtain preset feature codes;
extracting a corresponding feature code for each image received from the mobile device to obtain an image feature code;
Performing similarity calculation on the image feature codes and the preset feature codes to obtain feature similarity;
The calculation formula of the feature similarity is expressed as follows:
；
Wherein, Represent the firstFeature similarity of the individual image feature codes to the preset feature codes,Representing the first predetermined feature codeThe number of components of the composition,Represent the firstCoding of individual image featuresThe number of components of the composition,Representing the dimension of the feature code.
3. The image processing method for use in image recognition computation according to claim 2, wherein the method for performing angle analysis on specific structural features in each image specifically comprises the steps of:
analyzing the specific structural features in the provided image, and identifying the positions and the shapes of the specific structural features in the image;
based on the result of the feature analysis, comparing the specific structural features with known attitude angle information, and judging the relative angles of the specific structural features in the image;
Calculating the observation angle of the specific structural feature in a camera coordinate system according to the inferred relative angle by combining known camera parameters and feature position information;
Wherein, the calculation formula of the relative angle of the specific structural feature in the image is expressed as follows:
；
Wherein, Representing the relative angle of a particular structural feature in the image,Feature vectors representing particular structural features extracted from the image,The reference vector is represented by a reference vector,Representing feature vectorsIs provided with a die for the mold,Representing reference vectorsIs a mold of (2);
the calculation formula of the observation angle of the specific structural feature in the camera coordinate system is expressed as follows:
；
Wherein, Representing the viewing angle of a particular structural feature in the camera coordinate system,Representing the vertical offset of the feature point relative to the center of the image,Representing the horizontal offset of the feature point relative to the center of the image,Representing the adjustment coefficient.
4. An image processing method for use in image recognition computation according to claim 3, wherein the method for recognizing and recording image sizes of specific structural features of the same angle in all images, computing an intermediate value size, and performing size processing on all images based on the intermediate value size comprises the steps of:
based on the results of the angle analysis, the dimensions of the specific structural features at the same viewing angle are identified in all images and recorded;
Calculating the size intermediate value of the specific structural feature under the view angle according to the recorded size of the specific structural feature in different images;
based on the intermediate value size, adjusting the specific structural feature sizes in the images of all the same visual angles to the intermediate value size, analyzing the specific structural feature sizes of different visual angles after the adjustment is completed, comparing the association degree between the specific structural features of different angles, and secondarily adjusting the sizes of the specific structural features;
Wherein, the calculation formula of the size intermediate value of the specific structural feature is expressed as:
；
Wherein, Intermediate values of dimensions representing features of a particular structure,Representing the first in an imageThe dimensions of the individual features of the particular structure,Representing the first in an imageThe weight corresponding to the size of each particular structural feature,Representing the total amount of a particular structural feature in the image.
5. The method for image processing in image recognition computation according to claim 4, wherein the recognizing coordinates of the specific structural feature in the image after the size processing, and performing the position processing on all the images with the midpoint of the image size as the origin, specifically comprises:
analyzing the images with the processed sizes, taking the whole images as a coordinate system, taking the points in the images as the origin of coordinates, and establishing a coordinate system with the same size for each image, wherein the coordinate system contains complete specific structural features;
recognizing coordinate values of specific structural features in each image in an image coordinate system, and moving the image to an origin position by taking the coordinate values as a moving center;
And identifying the distance between the frame of each image and the frame of the coordinate system, cutting all the images by taking the farthest distance of each frame as a reference, and obtaining a final image with specific structural features at the middle point of the image and the same size of each image.
6. The image processing method for use in image recognition computing according to claim 5, wherein the mapping the specific structural features in each image into the three-dimensional space and detecting the alignment of the specific structural features in the three-dimensional space respectively specifically comprises:
creating a virtual three-dimensional space in a three-dimensional coordinate system based on the results of the angle analysis;
For specific structural features in each image, combining the results of angle analysis, and mapping the positions of the specific structural features in the image coordinate system to corresponding positions in a three-dimensional space according to the visual angle information;
In the three-dimensional space, detecting the alignment degree between specific structural features mapped to the three-dimensional space by different images, and carrying out secondary adjustment on the positions and the postures of the specific structural features;
wherein, the expression of the coordinate position of the specific structural feature in the three-dimensional space is:
；
Wherein, Respectively represent the firstThe abscissa and ordinate of each specific structural feature in the three-dimensional space in the horizontal direction,Represent the firstThe coordinates of the individual specific structural features in the vertical direction,Representing the abscissa and ordinate respectively of a particular structural feature in a two-dimensional image coordinate system,Represent the firstThe relative angle of the individual specific structural features in the image,Represent the firstThe azimuth of each particular structural feature relative to the camera,Represent the firstPitch angle of the specific structural feature relative to the camera.
7. The method for image processing in image recognition computing according to claim 6, wherein the optimizing the joint of the specific structural feature in the three-dimensional space comprises:
Analyzing and extracting textures of specific structural features and surrounding areas in the image;
Detecting smoothness of joints of the specific structural features in the three-dimensional space according to the mapping of the specific structural features in the three-dimensional space and in combination with the identified texture information;
and optimizing and adjusting the position, the posture and the shape of the joint according to the texture information of the joint and the positions of the adjacent specific structural features.
8. The image processing method according to claim 7, wherein in the step of optimally adjusting the position, posture, shape of the joint based on texture information of the joint and the positions of the adjacent specific structural features, the calculation formula for determining the optimized joint position is expressed as:
；
Wherein, Indicating the position of the joint after the optimization,Indicating the initial position of the seam,The weight factor is represented by a weight factor,A gradient representing the degree of texture variance,Representing vectors between adjacent specific structural features,Representing the distance between adjacent specific structural features,Representing the geometric vector of the joint.
9. The image processing method for use in image recognition computation according to claim 8, wherein the fusing of the images in the three-dimensional space and outputting as a three-dimensional stereoscopic image specifically comprises:
and according to the position and texture information of the specific structural features in the three-dimensional space, overlapping and fusing the images subjected to optimization and adjustment in the three-dimensional space to form a unified three-dimensional image.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410615990.3A CN118196310B (en) | 2024-05-17 | 2024-05-17 | Image processing method for image recognition calculation |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410615990.3A CN118196310B (en) | 2024-05-17 | 2024-05-17 | Image processing method for image recognition calculation |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN118196310A CN118196310A (en) | 2024-06-14 |
| CN118196310B true CN118196310B (en) | 2024-07-19 |
Family
ID=91406951
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202410615990.3A Active CN118196310B (en) | 2024-05-17 | 2024-05-17 | Image processing method for image recognition calculation |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN118196310B (en) |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110199293A (en) * | 2016-12-14 | 2019-09-03 | 玛涅蒂马瑞利有限公司 | Object identifying based on universal adaptive 3D model |
| CN115033967A (en) * | 2022-06-27 | 2022-09-09 | 北京航空航天大学 | Building template real-time modeling method based on point cloud data |
Family Cites Families (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US9197877B2 (en) * | 2011-11-22 | 2015-11-24 | Universitat De Valéncia | Smart pseudoscopic-to-orthoscopic conversion (SPOC) protocol for three-dimensional (3D) display |
| KR101666959B1 (en) * | 2015-03-25 | 2016-10-18 | ㈜베이다스 | Image processing apparatus having a function for automatically correcting image acquired from the camera and method therefor |
| CN115984530A (en) * | 2022-09-30 | 2023-04-18 | 江苏理工学院 | Large-size shaft part image acquisition device and acquisition method |
-
2024
- 2024-05-17 CN CN202410615990.3A patent/CN118196310B/en active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110199293A (en) * | 2016-12-14 | 2019-09-03 | 玛涅蒂马瑞利有限公司 | Object identifying based on universal adaptive 3D model |
| CN115033967A (en) * | 2022-06-27 | 2022-09-09 | 北京航空航天大学 | Building template real-time modeling method based on point cloud data |
Also Published As
| Publication number | Publication date |
|---|---|
| CN118196310A (en) | 2024-06-14 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US11354851B2 (en) | Damage detection from multi-view visual data | |
| CN112771539B (en) | Employing three-dimensional data predicted from two-dimensional images using neural networks for 3D modeling applications | |
| CN111126304B (en) | Augmented reality navigation method based on indoor natural scene image deep learning | |
| JP6220486B1 (en) | 3D model generation system, 3D model generation method, and program | |
| CN109934847B (en) | Method and device for estimating posture of weak texture three-dimensional object | |
| EP3242275B1 (en) | Using photo collections for three dimensional modeling | |
| JP6196416B1 (en) | 3D model generation system, 3D model generation method, and program | |
| AU2020211387B2 (en) | Damage detection from multi-view visual data | |
| KR20170008638A (en) | Three dimensional content producing apparatus and three dimensional content producing method thereof | |
| CN113298708A (en) | Three-dimensional house type generation method, device and equipment | |
| Pacheco et al. | Reconstruction of high resolution 3D objects from incomplete images and 3D information | |
| JP5503510B2 (en) | Posture estimation apparatus and posture estimation program | |
| CN117115358B (en) | Automatic digital person modeling method and device | |
| CN118247429A (en) | Air-ground cooperative rapid three-dimensional modeling method and system | |
| CN118196310B (en) | Image processing method for image recognition calculation | |
| CN116433811A (en) | Skeleton intelligent design system of virtual 3D role | |
| JP3791186B2 (en) | Landscape modeling device | |
| CN114677439A (en) | Camera pose determination method and device, electronic equipment and storage medium | |
| CN117115274B (en) | Method, device, equipment and storage medium for determining three-dimensional information | |
| CN116433852B (en) | Data processing method, device, equipment and storage medium | |
| CN118570424A (en) | Virtual reality tour guide system | |
| Liu et al. | A Keyframe Extraction Approach for 3D Videogrammetry Based on Baseline Constraints | |
| CN117173381A (en) | Image correction method and device and digital person automatic modeling method and device | |
| CN117857766A (en) | Video compression method based on key object 3D modeling | |
| CN115661883A (en) | Face direction correction method and device, computer equipment and storage medium |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |