CN112672043A - High-quality precise panoramic imaging method and system based on single lens reflex - Google Patents
High-quality precise panoramic imaging method and system based on single lens reflex Download PDFInfo
- Publication number
- CN112672043A CN112672043A CN202011496515.7A CN202011496515A CN112672043A CN 112672043 A CN112672043 A CN 112672043A CN 202011496515 A CN202011496515 A CN 202011496515A CN 112672043 A CN112672043 A CN 112672043A
- Authority
- CN
- China
- Prior art keywords
- image
- data
- acquiring
- task
- single lens
- 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.)
- Granted
Links
Images
Landscapes
- Studio Devices (AREA)
- Image Processing (AREA)
Abstract
The invention provides a high-quality precise panoramic imaging method and system based on a single lens reflex, which comprises the following steps: acquiring an image set of the single lens reflex, acquiring an execution task based on the image set, analyzing the execution task, and acquiring task analysis data; acquiring a current working mode of the single lens reflex, and determining an execution instruction by combining the task analysis data and the working mode; acquiring the execution instruction, and controlling the single lens reflex to execute image downloading on the image set or process the image set into a panoramic picture according to the execution instruction; by acquiring the image set of the single camera, image downloading or accurate splicing of the image set is realized according to a specified task, high-precision processing of images is realized, and high-precision splicing or downloading of photos is realized.
Description
Technical Field
The invention relates to the technical field of panoramic imaging, in particular to a high-quality precise panoramic imaging method and system based on a single lens reflex.
Background
At present, in an image acquisition device with multiple streams running on a single lens in the market, an image is formed by using an optical imaging principle, a negative film is used for recording the image, images of all angles are obtained by moving the single lens, and then the images are spliced to form a panoramic image.
However, the image acquisition device with a single lens is used for acquiring panoramic photos, so that a large amount of manpower and time are wasted, the quality of the synthesized images is poor, the splicing process is complicated, and the splicing efficiency is low.
Therefore, the invention provides a high-quality precise panoramic imaging method and system based on a single lens reflex.
Disclosure of Invention
The invention provides a high-quality and precise panoramic imaging method and system based on a single lens reflex, which are used for processing and downloading images and synthesizing high-precision panoramic photos.
A high-quality precise panoramic imaging method based on a single lens reflex comprises the following steps:
acquiring an image set of the single lens reflex, acquiring an execution task based on the image set, analyzing the execution task, and acquiring task analysis data;
acquiring a current working mode of the single lens reflex, and determining an execution instruction by combining the task analysis data and the working mode;
and acquiring the execution instruction, and controlling the SLR camera to execute image downloading on the image set or process the image set into a panoramic picture according to the execution instruction.
Preferably, after the task of executing the single lens reflex is acquired and before the parsing, the method for high-quality precise panoramic imaging based on a single lens reflex further includes:
acquiring a binary code set corresponding to an execution task based on a basic execution rule of the execution task of the single lens reflex;
acquiring the starting time of the execution task, and acquiring the total space amount of the binary coding set;
determining a duration of the execution task based on a start time of the execution task and a total amount of space of the binary-coded set;
and determining the analysis speed required for starting analysis of the execution task according to the duration.
Preferably, the specific working process of analyzing the execution task by the high-quality precise panoramic imaging method based on the single lens reflex comprises the following steps:
acquiring a data protocol of the execution task, and acquiring source data of the execution task according to the data protocol;
defining an analysis identifier for the source data of the execution task according to a preset basic analysis function;
meanwhile, determining a data analysis frame header of the source data according to the analysis identifier of the source data;
determining a data type identifier of the source data based on the data analysis frame header, and meanwhile, establishing a corresponding analysis rule according to the data type identifier;
establishing a mapping relation between the analysis rule and the source data of the execution task;
and analyzing the execution task based on the mapping relation, and acquiring the task analysis data.
Preferably, the method for high-quality precise panoramic imaging based on a single lens reflex camera determines a working process of executing the instruction according to the task analysis data and the working mode, and includes:
acquiring a first working mode of the single lens reflex, and extracting mode data corresponding to the first working mode;
performing data matching on the mode data and the task analysis data based on a preset matching rule;
if the mode data is matched with the task analysis data, judging that the current first working mode of the single lens reflex is in accordance with the execution task of the single lens reflex, and meanwhile, determining that the execution instruction is the current first working mode of the single lens reflex;
otherwise, acquiring a transcoding task request in a preset period, and deleting the first working mode of the single lens reflex at present;
distributing a read-write lock for the transcoding task request, and performing read-write operation on a second working mode of the single lens reflex;
converting the first working mode into the second working mode according to the transcoding task request;
meanwhile, the execution instruction is determined to be a second working mode of the single lens reflex.
Preferably, a high-quality precise panoramic imaging method based on a single lens reflex,
the first working mode of the single lens reflex camera is as follows: a picture downloading mode;
the second working mode of the single lens reflex camera is as follows: and (5) a panoramic splicing mode.
Preferably, the specific step process of performing data matching on the mode data and the task analysis data based on a preset matching rule includes:
step 1: storing the mode data and the task analysis data into a matched data packet;
meanwhile, acquiring a first dimension of the mode data in the matching data packet and a second dimension of the task analysis data;
step 2: acquiring a hierarchical relationship between the first dimension and the second dimension, and determining a matching rule of the mode data and the task analysis data according to the hierarchical relationship;
and step 3: according to the matching rule, performing data matching on the mode data and the task analysis data in the matching data packet;
and 4, step 4: acquiring a first matching code corresponding to the matched mode data and the matched task analysis data, and extracting the corresponding relation of the first matching code;
and 5: judging whether the corresponding relation of the first matching code is consistent with the matching rule or not;
if so, selecting the mode data and the task analysis data corresponding to the first matching code with the highest priority as an optimal matching result;
otherwise, reducing the stage number of the matching code, obtaining a second matching code, and repeating the step 4-5, wherein the second matching code is the final matching code.
Preferably, when the single lens reflex is in a panoramic stitching mode, the panoramic stitching working process includes:
acquiring a target image to be synthesized in the single-phase inverter, and acquiring a synthesis position coordinate of the target image;
determining a detection area corresponding to the target image based on the synthesis position coordinates, generating a countermeasure network related to the detection area, and setting a network loss function;
based on gray scale compensation, acquiring pixel data of the target image, training the pixel data in the countermeasure network according to the network loss function, and simultaneously carrying out image clustering on the target image by utilizing a clustering algorithm to acquire an image set;
acquiring a synthesis model of the image set based on the trained pixel data and the clustered image set;
performing transfer learning on the synthetic model to obtain a target area synthetic model;
extracting edge feature points of the target area synthesis model, and calculating the distance between the edge of the target area synthesis model and the edge feature points;
meanwhile, determining the synthesis sequence of the target images in the image set according to the distance and the obtained target area synthesis model;
determining a texture image block of the target image in the image set according to the target region synthesis model;
and splicing the target images based on the texture image blocks of the target images according to the synthesis sequence to obtain a panoramic spliced synthetic image.
Preferably, the high-quality precise panoramic imaging method based on a single lens reflex further includes, after the target images are stitched, the steps of:
extracting the synthetic edge noise of the spliced synthetic image, calculating the image fusion rate of the spliced synthetic image according to the synthetic edge noise, and meanwhile, calculating the image integration degree of the target image according to the image fusion rate, wherein the specific working process comprises the following steps:
graying the spliced composite image and acquiring grayscale gradient data of the spliced composite image;
placing the gray gradient data in a preset neural convolution network for training and learning to obtain the synthetic edge noise of the spliced synthetic image;
calculating the image fusion rate of the spliced composite image based on the composite edge noise;
wherein η represents an image fusion rate of the stitched composite image, k represents a composite edge noise of the stitched composite image, h represents an image gray value of the stitched composite image, m represents an image gradient value of the stitched composite image, ζ represents an image resolution of the stitched composite image, σ represents a stitching error rate of the target image, ξ represents an image fusion coefficient of the stitched composite image, g represents an error norm of the target image, and d represents an edge blur degree of the stitched composite image;
acquiring the wave band of the spliced composite image according to the image fusion rate of the spliced composite image, and acquiring the wave band interpolation of the spliced composite image according to the wave band;
calculating the image integration degree of the target image based on the image fusion rate of the spliced composite image and the band interpolation;
wherein P represents an image integration degree of the target image, η represents an image fusion rate of the stitched composite image, xiRepresenting the ith spectral value of the spliced composite graph, i representing the number of the spectral values of the spliced composite graph, N representing the total number of the spectral values of the spliced composite graph, mu representing the mean value of the spectral values of the spliced composite graph, f representing the band interpolation of the spliced composite graph, v representing the splicing speed of the target image, q representing the smoothing rate of the spliced composite graph, and tau representing the compactness of the spliced composite graph;
according to the image integration degree of the target image, performing splicing quality estimation on the spliced composite image according to the standard integration degree;
if the image integration level is equal to the standard integration level, outputting the spliced composite image;
otherwise, splicing the target images again until the standard integration level is met.
A single lens reflex based high quality precision panoramic imaging system comprising:
the data processing module is used for acquiring an execution task of the single lens reflex, analyzing the execution task and acquiring task analysis data;
the data conversion module is used for acquiring the current working mode of the single lens reflex and determining an execution instruction by combining the task analysis data and the working mode;
and the data acquisition module is used for acquiring the execution instruction and executing the current task of the single lens reflex according to the execution instruction.
Additional features and advantages of the invention will be set forth in the description which follows, and in part will be obvious from the description, or may be learned by practice of the invention. The objectives and other advantages of the invention will be realized and attained by the structure particularly pointed out in the written description and claims hereof as well as the appended drawings.
The technical solution of the present invention is further described in detail by the accompanying drawings and embodiments.
Drawings
The accompanying drawings, which are included to provide a further understanding of the invention and are incorporated in and constitute a part of this specification, illustrate embodiments of the invention and together with the description serve to explain the principles of the invention and not to limit the invention. In the drawings:
FIG. 1 is a flow chart of a high-quality precise panoramic imaging method based on a single lens reflex in an embodiment of the present invention;
fig. 2 is a structural diagram of a high-quality precise panoramic imaging system based on a single lens reflex according to an embodiment of the present invention.
Detailed Description
The preferred embodiments of the present invention will be described in conjunction with the accompanying drawings, and it will be understood that they are described herein for the purpose of illustration and explanation and not limitation.
Example 1:
the invention provides a high-quality precise panoramic imaging method based on a single lens reflex, which comprises the following steps of:
acquiring an image set of the single lens reflex, acquiring an execution task based on the image set, analyzing the execution task, and acquiring task analysis data;
acquiring a current working mode of the single lens reflex, and determining an execution instruction by combining the task analysis data and the working mode;
and acquiring the execution instruction, and controlling the SLR camera to execute image downloading on the image set or process the image set into a panoramic picture according to the execution instruction.
In this embodiment, the execution task is to control the operation mode of the current single lens reflex camera.
In this embodiment, the operation modes of the single lens reflex camera include: an image downloading mode and an image splicing mode.
In this embodiment, the image set refers to a set formed by target images selected in the single lens reflex, and may be obtained in a clustering manner.
The beneficial effects of the above technical scheme are:
by acquiring the image set of the single lens reflex and executing image downloading or accurate splicing of the image set according to the specified task, not only can the high-precision downloading of the images be realized, but also the high-precision panoramic photo can be obtained.
Example 2:
on the basis of embodiment 1, the invention provides a high-quality precise panoramic imaging method based on a single lens reflex, which further comprises the following steps after the task of the single lens reflex is executed and before the analysis is performed:
acquiring a binary code set corresponding to an execution task based on a basic execution rule of the execution task of the single lens reflex;
acquiring the starting time of the execution task, and acquiring the total space amount of the binary coding set;
determining a duration of the execution task based on a start time of the execution task and a total amount of space of the binary-coded set;
and determining the analysis speed required for starting analysis of the execution task according to the duration.
In this embodiment, the execution rule may be determined based on the execution task in order to obtain the corresponding binary encoding set.
In this embodiment, the starting time for executing the task and the total space amount of the binary coding set are obtained, so that the starting time is used as a starting point for executing the task, an end point for executing the task is determined according to the total space amount of the binary coding set, and a duration for executing the task is determined according to the starting point and the end point.
The working principle of the technical scheme is as follows:
the time for executing the task can be determined by acquiring the starting time for executing the task and the total space amount of the binary coding set, so that the analysis speed required by the starting analysis of the task is determined, and the accurate control of the time rate in the analysis process is facilitated.
Example 3:
on the basis of embodiment 1, the invention provides a high-quality precise panoramic imaging method based on a single lens reflex, and a specific working process for analyzing the execution task comprises the following steps:
acquiring a data protocol of the execution task, and acquiring source data of the execution task according to the data protocol;
defining an analysis identifier for the source data of the execution task according to a preset basic analysis function;
meanwhile, determining a data analysis frame header of the source data according to the analysis identifier of the source data;
determining a data type identifier of the source data based on the data analysis frame header, and meanwhile, establishing a corresponding analysis rule according to the data type identifier;
establishing a mapping relation between the analysis rule and the source data of the execution task;
and analyzing the execution task based on the mapping relation, and acquiring the task analysis data.
In this embodiment, the data protocol may be a Protocol Data Unit (PDU) that is to be built at each layer of the transport system in a layered network structure, such as in the Open Systems Interconnection (OSI) model, where the PDU contains information from the upper layers, as well as information appended to the entities of the current layer.
In this embodiment, the parsing frame header may be a protocol data unit of a data link layer, which includes data type, data control information, and so on.
In this embodiment, the data type identifier may be of the float type, int type or floating point type.
In this embodiment, the mapping relationship may be that each parsing rule corresponds to one source data, or that one parsing rule corresponds to multiple source data.
The working principle of the technical scheme is as follows:
the method comprises the steps of acquiring source data of an execution task through a data protocol of the execution task, accurately acquiring an analysis identifier, accurately acquiring a data type identifier according to the head of an analysis frame of the source data, determining an analysis rule, establishing a mapping relation between the analysis rule and the source data of the execution task, realizing accurate analysis of the execution task, and improving data processing efficiency.
Example 4:
on the basis of embodiment 1, the invention provides a high-quality precise panoramic imaging method based on a single lens reflex, which determines a working process of an execution instruction according to the task analysis data and the working mode, and comprises the following steps:
acquiring a first working mode of the single lens reflex, and extracting mode data corresponding to the first working mode;
performing data matching on the mode data and the task analysis data based on a preset matching rule;
if the mode data is matched with the task analysis data, judging that the current first working mode of the single lens reflex is in accordance with the execution task of the single lens reflex, and meanwhile, determining that the execution instruction is the current first working mode of the single lens reflex;
otherwise, acquiring a transcoding task request in a preset period, and deleting the first working mode of the single lens reflex at present;
distributing a read-write lock for the transcoding task request, and performing read-write operation on a second working mode of the single lens reflex;
converting the first working mode into the second working mode according to the transcoding task request;
meanwhile, the execution instruction is determined to be a second working mode of the single lens reflex.
In this embodiment, the mode data includes execution data of the operation mode and identification data of the operation mode.
In this embodiment, the matching rule may be a matching rule set based on the result of the training of the matching data.
In this embodiment, the first operating mode may be a picture downloading mode, and the second operating mode may be a panorama stitching mode.
In this embodiment, the first working mode of the single lens reflex is deleted to maintain data consistency when data is converted, and avoid data interference, thereby realizing conversion of the working mode.
In this embodiment, the read-write lock is to ensure data security when the data receives the transcoding task request.
The beneficial effects of the above technical scheme are:
through data matching of the mode data and the analysis data, whether the current working mode of the single lens reflex is consistent with the execution task or not can be determined, if so, an instruction is executed, otherwise, data conversion is carried out, and the working mode is converted, so that the conversion of the working mode of the single lens reflex is accurately finished, and the working efficiency of the single lens reflex is improved.
Example 5:
on the basis of embodiment 4, the invention provides a high-quality precise panoramic imaging method based on a single lens reflex,
the first working mode of the single lens reflex camera is as follows: a picture downloading mode;
the second working mode of the single lens reflex camera is as follows: and (5) a panoramic splicing mode.
The beneficial effects of the above technical scheme are: and precisely defining the working mode of the single lens reflex to finish the working content of the single lens reflex.
Example 6:
on the basis of embodiment 4, the invention provides a high-quality precise panoramic imaging method based on a single lens reflex, and the specific step process of performing data matching on the mode data and the task analysis data based on a preset matching rule comprises the following steps:
step 1: storing the mode data and the task analysis data into a matched data packet;
meanwhile, acquiring a first dimension of the mode data in the matching data packet and a second dimension of the task analysis data;
step 2: acquiring a hierarchical relationship between the first dimension and the second dimension, and determining a matching rule of the mode data and the task analysis data according to the hierarchical relationship;
and step 3: according to the matching rule, performing data matching on the mode data and the task analysis data in the matching data packet;
and 4, step 4: acquiring a first matching code corresponding to the matched mode data and the matched task analysis data, and extracting the corresponding relation of the first matching code;
and 5: judging whether the corresponding relation of the first matching code is consistent with the matching rule or not;
if so, selecting the mode data and the task analysis data corresponding to the first matching code with the highest priority as an optimal matching result;
otherwise, reducing the stage number of the matching code, obtaining a second matching code, and repeating the step 4-5, wherein the second matching code is the final matching code.
In this embodiment, the matching data packet may be established according to a collection of pattern data and parsing data, so as to provide a space for data matching and avoid interference of irrelevant data.
In this embodiment, the obtaining of the hierarchical relationship may be established based on a relationship between a first dimension of the schema data and a second dimension of the parsing data, for example: when the first dimension is equal to the second dimension, the hierarchical relationship is an equal relationship, when the first dimension is smaller than the second dimension, the hierarchical relationship is an analytic relationship, and when the first dimension is larger than the second dimension, the hierarchical relationship is a mode relationship.
In this embodiment, the correspondence relationship of the matching codes includes: the correlation degree, the first dimension, the second dimension, the hierarchical relationship and other basic contents of the mode data and the analysis data.
In this embodiment, the first matching code with the highest priority is obtained to obtain the optimal matching result, and the data matching degree at this time is the highest.
In this embodiment, the first dimension may be an organization form constructed from a plurality of mode data, and may be in the form of one-dimensional data.
In this embodiment, the second dimension may be an organization form formed by parsing data according to a plurality of tasks, and the second dimension may be one-dimensional data.
In this embodiment, the first matching code may be formed according to a data segment in which the pattern data matches with data in the task analysis data, and the first matching code is obtained by performing binary coding on the data in the data segment.
In this embodiment, the matching rules may be matching rules established based on regular expressions, and they are mainly composed of a group of characters describing the character string features.
In this embodiment, the reducing of the number of stages of the matching code may be reducing of matching goodness of fit of the mode data and the task analysis data, and re-encoding the first matching code, thereby reducing the number of stages of the matching code.
In this embodiment, the second matching code may be a second matching code obtained by reducing the number of matching code stages.
The beneficial effects of the above technical scheme are:
by storing the pattern data and the analysis data into the matching data packet, the integrity of the data is protected, the first dimension of the pattern data and the second dimension of the analysis data can be quickly obtained, the hierarchical relationship is further determined, the hierarchical relationship is beneficial to obtaining the matching rule, the first matching code is further obtained, and the optimal matching result is further determined.
Example 7:
on the basis of embodiment 5, the invention provides a high-quality precise panoramic imaging method based on a single lens reflex, and when the single lens reflex is in a panoramic stitching mode, the working process of panoramic stitching comprises the following steps:
acquiring a target image to be synthesized in the single-phase inverter, and acquiring a synthesis position coordinate of the target image;
determining a detection area corresponding to the target image based on the synthesis position coordinates, generating a countermeasure network related to the detection area, and setting a network loss function;
based on gray scale compensation, acquiring pixel data of the target image, training the pixel data in the countermeasure network according to the network loss function, and simultaneously carrying out image clustering on the target image by utilizing a clustering algorithm to acquire an image set;
acquiring a synthesis model of the image set based on the trained pixel data and the clustered image set;
performing transfer learning on the synthetic model to obtain a target area synthetic model;
extracting edge feature points of the target area synthesis model, and calculating the distance between the edge of the target area synthesis model and the edge feature points;
meanwhile, determining the synthesis sequence of the target images in the image set according to the distance and the obtained target area synthesis model;
determining a texture image block of the target image in the image set according to the target region synthesis model;
and splicing the target images based on the texture image blocks of the target images according to the synthesis sequence to obtain a panoramic spliced synthetic image.
In this embodiment, the synthesized position coordinates are determined based on the edge points of the target image, and are used to locate the detection area.
In this embodiment, the detection area corresponding to the target image is determined based on the synthesized position coordinates, for example, the target image may be positioned by the synthesized position coordinates, and then the detection area may be determined with the range of (-10, 10) of the synthesized position coordinates.
In this embodiment, the countermeasure network is generated to protect data in the detection area, so that erosion of interference data can be effectively prevented.
In this embodiment, the transfer learning of the synthetic model may be performed by selecting parameters useful for the synthetic model from the synthetic model, for example, performing effective weight distribution on the parameters of the synthetic model, and making the synthetic model approach the synthetic model of the target region, thereby establishing a high-precision synthetic model.
In this embodiment, the gray scale compensation refers to an adjustment process of gray scale brightness of the target image.
In this embodiment, the composite model may be synthesized by edge features of each image in the image set, so that the accumulation of the synthesis forms the minimum synthesis unit, and a complete composite model may be formed by associating the graphic file formed by the minimum synthesis unit with the attribute file recording the synthesis transformation once.
In this embodiment, the target region synthesis model may be a synthesis model that is obtained by performing transfer learning on the synthesis model, that is, learning the synthesis model, where the synthesis model formed by the target region is the target region synthesis model.
The beneficial effects of the above technical scheme are:
the method comprises the steps of obtaining the coordinates of the synthesis position of a target image, further determining the detection area corresponding to the target image, generating a countermeasure network to protect the data of the target image, accurately obtaining a synthesis model in the countermeasure network through a network loss function, effectively obtaining the synthesis model of the target area through transfer learning of the synthesis model, achieving high accuracy of the synthesis model, further determining the synthesis sequence of the target image according to the distance between the edge of the synthesis model of the target area and the edge feature point, and accurately achieving splicing of the target image through texture blocks and the synthesis sequence of the target image, so that the splicing effect is improved, and the image splicing efficiency is improved.
Example 8:
on the basis of embodiment 7, the present invention provides a high-quality precise panoramic imaging method based on a single lens reflex, and after stitching the target images, the method further includes:
extracting the synthetic edge noise of the spliced synthetic image, calculating the image fusion rate of the spliced synthetic image according to the synthetic edge noise, and meanwhile, calculating the image integration degree of the target image according to the image fusion rate, wherein the specific working process comprises the following steps:
graying the spliced composite image and acquiring grayscale gradient data of the spliced composite image;
placing the gray gradient data in a preset neural convolution network for training and learning to obtain the synthetic edge noise of the spliced synthetic image;
calculating the image fusion rate of the spliced composite image based on the composite edge noise;
wherein η represents an image fusion rate of the stitched composite image, k represents a composite edge noise of the stitched composite image, h represents an image gray value of the stitched composite image, m represents an image gradient value of the stitched composite image, ζ represents an image resolution of the stitched composite image, σ represents a stitching error rate of the target image, ξ represents an image fusion coefficient of the stitched composite image, g represents an error norm of the target image, and d represents an edge blur degree of the stitched composite image;
acquiring the wave band of the spliced composite image according to the image fusion rate of the spliced composite image, and acquiring the wave band interpolation of the spliced composite image according to the wave band;
calculating the image integration degree of the target image based on the image fusion rate of the spliced composite image and the band interpolation;
wherein P represents an image integration degree of the target image, η represents an image fusion rate of the stitched composite image, xiRepresenting the ith spectral value of the spliced composite graph, i representing the number of the spectral values of the spliced composite graph, N representing the total number of the spectral values of the spliced composite graph, mu representing the mean value of the spectral values of the spliced composite graph, f representing the band interpolation of the spliced composite graph, v representing the splicing speed of the target image, q representing the smoothing rate of the spliced composite graph, and tau representing the compactness of the spliced composite graph;
according to the image integration degree of the target image, performing splicing quality estimation on the spliced composite image according to the standard integration degree;
if the image integration level is equal to the standard integration level, outputting the spliced composite image;
otherwise, splicing the target images again until the standard integration level is met.
In this embodiment, the composite edge noise is obtained by combining a gaussian smoothing filter with a laplacian operator by the gaussian-laplacian operator.
In this embodiment, the mosaic image is grayed to obtain a grayscale image of the mosaic composite image, and the grayscale range is (0, 255).
In this embodiment, the error norm is a linear space over the exponential domain.
In this embodiment, the band interpolation is obtained for the purpose of accurate analysis of the mosaic.
The beneficial effects of the above technical scheme are:
through the synthetic edge noise of obtaining the splicing synthetic picture, and then can realize carrying out accurate analysis to the image rate of splicing synthetic picture, and then can obtain the image integration degree of target image, can realize the concatenation quality estimation to the synthetic splicing picture, when not conform to standard integration degree, splice again, improved the precision of the synthetic panorama photo of single opposition machine.
Example 9:
the invention provides a high-quality precise panoramic imaging system based on a single lens reflex, as shown in figure 2, comprising:
the data processing module is used for acquiring an execution task of the single lens reflex, analyzing the execution task and acquiring task analysis data;
the data conversion module is used for acquiring the current working mode of the single lens reflex and determining an execution instruction by combining the task analysis data and the working mode;
and the data acquisition module is used for acquiring the execution instruction and executing the current task of the single lens reflex according to the execution instruction.
The beneficial effects of the above technical scheme are:
by acquiring the image set of the single camera, image downloading or accurate splicing of the image set is realized according to a specified task, high-precision processing of images is realized, and high-precision splicing or downloading of photos is realized.
It will be apparent to those skilled in the art that various changes and modifications may be made in the present invention without departing from the spirit and scope of the invention. Thus, if such modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention is also intended to include such modifications and variations.
Claims (9)
1. A high-quality precise panoramic imaging method based on a single lens reflex is characterized by comprising the following steps:
acquiring an image set of the single lens reflex, acquiring an execution task based on the image set, analyzing the execution task, and acquiring task analysis data;
acquiring a current working mode of the single lens reflex, and determining an execution instruction by combining the task analysis data and the working mode;
and acquiring the execution instruction, and controlling the SLR camera to execute image downloading on the image set or process the image set into a panoramic picture according to the execution instruction.
2. The method for high-quality precise panoramic imaging based on the single lens reflex camera as claimed in claim 1, wherein after the task of the single lens reflex camera is executed and before the parsing, the method further comprises:
acquiring a binary code set corresponding to an execution task based on a basic execution rule of the execution task of the single lens reflex;
acquiring the starting time of the execution task, and acquiring the total space amount of the binary coding set;
determining a duration of the execution task based on a start time of the execution task and a total amount of space of the binary-coded set;
and determining the analysis speed required for starting analysis of the execution task according to the duration.
3. The method for high-quality precise panoramic imaging based on the single lens reflex camera as claimed in claim 1, wherein the specific work process for analyzing the execution task comprises:
acquiring a data protocol of the execution task, and acquiring source data of the execution task according to the data protocol;
defining an analysis identifier for the source data of the execution task according to a preset basic analysis function;
meanwhile, determining a data analysis frame header of the source data according to the analysis identifier of the source data;
determining a data type identifier of the source data based on the data analysis frame header, and meanwhile, establishing a corresponding analysis rule according to the data type identifier;
establishing a mapping relation between the analysis rule and the source data of the execution task;
and analyzing the execution task based on the mapping relation, and acquiring the task analysis data.
4. The method for high-quality precise panoramic imaging based on the single lens reflex camera as claimed in claim 1, wherein the determining of the work process of executing the instructions according to the task analysis data and the work mode comprises:
acquiring a first working mode of the single lens reflex, and extracting mode data corresponding to the first working mode;
performing data matching on the mode data and the task analysis data based on a preset matching rule;
if the mode data is matched with the task analysis data, judging that the current first working mode of the single lens reflex is in accordance with the execution task of the single lens reflex, and meanwhile, determining that the execution instruction is the current first working mode of the single lens reflex;
otherwise, acquiring a transcoding task request in a preset period, and deleting the first working mode of the single lens reflex at present;
distributing a read-write lock for the transcoding task request, and performing read-write operation on a second working mode of the single lens reflex;
converting the first working mode into the second working mode according to the transcoding task request;
meanwhile, the execution instruction is determined to be a second working mode of the single lens reflex.
5. The method of claim 4, wherein the high-quality precise panoramic imaging method based on the single lens reflex camera,
the first working mode of the single lens reflex camera is as follows: a picture downloading mode;
the second working mode of the single lens reflex camera is as follows: and (5) a panoramic splicing mode.
6. The method for high-quality precise panoramic imaging based on the single lens reflex camera as claimed in claim 4, wherein the specific step process of performing data matching on the mode data and the task analysis data based on a preset matching rule comprises:
step 1: storing the mode data and the task analysis data into a matched data packet;
meanwhile, acquiring a first dimension of the mode data in the matching data packet and a second dimension of the task analysis data;
step 2: acquiring a hierarchical relationship between the first dimension and the second dimension, and determining a matching rule of the mode data and the task analysis data according to the hierarchical relationship;
and step 3: according to the matching rule, performing data matching on the mode data and the task analysis data in the matching data packet;
and 4, step 4: acquiring a first matching code corresponding to the matched mode data and the matched task analysis data, and extracting the corresponding relation of the first matching code;
and 5: judging whether the corresponding relation of the first matching code is consistent with the matching rule or not;
if so, selecting the mode data and the task analysis data corresponding to the first matching code with the highest priority as an optimal matching result;
otherwise, reducing the stage number of the matching code, obtaining a second matching code, and repeating the step 4-5, wherein the second matching code is the final matching code.
7. The method for high-quality precise panoramic imaging based on the single lens reflex according to claim 5, wherein when the single lens reflex is in a panoramic stitching mode, the working process of panoramic stitching comprises the following steps:
acquiring a target image to be synthesized in the single-phase inverter, and acquiring a synthesis position coordinate of the target image;
determining a detection area corresponding to the target image based on the synthesis position coordinates, generating a countermeasure network related to the detection area, and setting a network loss function;
based on gray scale compensation, acquiring pixel data of the target image, training the pixel data in the countermeasure network according to the network loss function, and simultaneously carrying out image clustering on the target image by utilizing a clustering algorithm to acquire an image set;
acquiring a synthesis model of the image set based on the trained pixel data and the clustered image set;
performing transfer learning on the synthetic model to obtain a target area synthetic model;
extracting edge feature points of the target area synthesis model, and calculating the distance between the edge of the target area synthesis model and the edge feature points;
meanwhile, determining the synthesis sequence of the target images in the image set according to the distance and the obtained target area synthesis model;
determining a texture image block of the target image in the image set according to the target region synthesis model;
and splicing the target images based on the texture image blocks of the target images according to the synthesis sequence to obtain a panoramic spliced synthetic image.
8. The method for high-quality precise panoramic imaging based on the single lens reflex according to claim 7, wherein after the target images are spliced, the method further comprises:
extracting the synthetic edge noise of the spliced synthetic image, calculating the image fusion rate of the spliced synthetic image according to the synthetic edge noise, and meanwhile, calculating the image integration degree of the target image according to the image fusion rate, wherein the specific working process comprises the following steps:
graying the spliced composite image and acquiring grayscale gradient data of the spliced composite image;
placing the gray gradient data in a preset neural convolution network for training and learning to obtain the synthetic edge noise of the spliced synthetic image;
calculating the image fusion rate of the spliced composite image based on the composite edge noise;
wherein η represents an image fusion rate of the stitched composite image, k represents a composite edge noise of the stitched composite image, h represents an image gray value of the stitched composite image, m represents an image gradient value of the stitched composite image, ζ represents an image resolution of the stitched composite image, σ represents a stitching error rate of the target image, ξ represents an image fusion coefficient of the stitched composite image, g represents an error norm of the target image, and d represents an edge blur degree of the stitched composite image;
acquiring the wave band of the spliced composite image according to the image fusion rate of the spliced composite image, and acquiring the wave band interpolation of the spliced composite image according to the wave band;
calculating the image integration degree of the target image based on the image fusion rate of the spliced composite image and the band interpolation;
wherein P represents an image integration degree of the target image, η represents an image fusion rate of the stitched composite image, xiI-th spectral value representing said stitched synthesis map, i representing said stitched synthesisThe number of spectral values of the graph, N represents the total number of spectral values of the spliced composite graph, mu represents the mean value of the spectral values of the spliced composite graph, f represents the band interpolation of the spliced composite graph, v represents the splicing speed of the target image, q represents the smoothing rate of the spliced composite graph, and tau represents the compactness of the spliced composite graph;
according to the image integration degree of the target image, performing splicing quality estimation on the spliced composite image according to the standard integration degree;
if the image integration level is equal to the standard integration level, outputting the spliced composite image;
otherwise, splicing the target images again until the standard integration level is met.
9. A high-quality precise panoramic imaging system based on a single lens reflex camera is characterized by comprising:
the data processing module is used for acquiring an execution task of the single lens reflex, analyzing the execution task and acquiring task analysis data;
the data conversion module is used for acquiring the current working mode of the single lens reflex and determining an execution instruction by combining the task analysis data and the working mode;
and the data acquisition module is used for acquiring the execution instruction and executing the current task of the single lens reflex according to the execution instruction.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011496515.7A CN112672043B (en) | 2020-12-17 | 2020-12-17 | High-quality precise panoramic imaging method and system based on single lens reflex |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011496515.7A CN112672043B (en) | 2020-12-17 | 2020-12-17 | High-quality precise panoramic imaging method and system based on single lens reflex |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN112672043A true CN112672043A (en) | 2021-04-16 |
| CN112672043B CN112672043B (en) | 2021-09-14 |
Family
ID=75404831
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202011496515.7A Active CN112672043B (en) | 2020-12-17 | 2020-12-17 | High-quality precise panoramic imaging method and system based on single lens reflex |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112672043B (en) |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20040233274A1 (en) * | 2000-07-07 | 2004-11-25 | Microsoft Corporation | Panoramic video |
| CN108769755A (en) * | 2018-04-27 | 2018-11-06 | 湖北工业大学 | High-resolution full view frequency live streaming camera system and method |
| CN108933899A (en) * | 2018-08-22 | 2018-12-04 | Oppo广东移动通信有限公司 | Panorama shooting method, device, terminal and computer readable storage medium |
| CN109600556A (en) * | 2019-02-18 | 2019-04-09 | 武汉大学 | A kind of high quality precision omnidirectional imaging system and method based on slr camera |
| CN111367598A (en) * | 2018-12-26 | 2020-07-03 | 北京奇虎科技有限公司 | Action instruction processing method and device, electronic equipment and computer-readable storage medium |
-
2020
- 2020-12-17 CN CN202011496515.7A patent/CN112672043B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20040233274A1 (en) * | 2000-07-07 | 2004-11-25 | Microsoft Corporation | Panoramic video |
| CN108769755A (en) * | 2018-04-27 | 2018-11-06 | 湖北工业大学 | High-resolution full view frequency live streaming camera system and method |
| CN108933899A (en) * | 2018-08-22 | 2018-12-04 | Oppo广东移动通信有限公司 | Panorama shooting method, device, terminal and computer readable storage medium |
| CN111367598A (en) * | 2018-12-26 | 2020-07-03 | 北京奇虎科技有限公司 | Action instruction processing method and device, electronic equipment and computer-readable storage medium |
| CN109600556A (en) * | 2019-02-18 | 2019-04-09 | 武汉大学 | A kind of high quality precision omnidirectional imaging system and method based on slr camera |
Also Published As
| Publication number | Publication date |
|---|---|
| CN112672043B (en) | 2021-09-14 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| RU2698402C1 (en) | Method of training a convolutional neural network for image reconstruction and a system for forming an image depth map (versions) | |
| CN109583483B (en) | Target detection method and system based on convolutional neural network | |
| CN110248096A (en) | Focusing method and device, electronic equipment, computer readable storage medium | |
| CN112766160A (en) | Face replacement method based on multi-stage attribute encoder and attention mechanism | |
| CN111428575B (en) | Tracking method for fuzzy target based on twin network | |
| CN111723691B (en) | Three-dimensional face recognition method and device, electronic equipment and storage medium | |
| KR20060058147A (en) | Image collation system and image collation method | |
| CN109376641B (en) | Moving vehicle detection method based on unmanned aerial vehicle aerial video | |
| CN113011401B (en) | Face image posture estimation and correction method, system, medium and electronic equipment | |
| CN110191287A (en) | Focusing method and device, electronic equipment, computer readable storage medium | |
| CN111768452A (en) | Non-contact automatic mapping method based on deep learning | |
| CN115996322B (en) | Image data management method for digital video shooting | |
| CN111915735B (en) | Depth optimization method for three-dimensional structure outline in video | |
| CN111680573B (en) | Face recognition method, device, electronic equipment and storage medium | |
| CN112929626A (en) | Three-dimensional information extraction method based on smartphone image | |
| CN111681271B (en) | Multichannel multispectral camera registration method, system and medium | |
| CN112672043B (en) | High-quality precise panoramic imaging method and system based on single lens reflex | |
| CN113628334A (en) | Visual SLAM method, device, terminal equipment and storage medium | |
| CN115063813B (en) | Training method and training device of alignment model aiming at character distortion | |
| Su et al. | Gpr-net: Multi-view layout estimation via a geometry-aware panorama registration network | |
| CN115496859A (en) | Three-dimensional scene motion trend estimation method based on scattered point cloud cross attention learning | |
| CN114863021A (en) | Simulation data set analysis method and system based on three-dimensional reconstruction scene | |
| CN115456870A (en) | Multi-image splicing method based on external parameter estimation | |
| CN115035193A (en) | Bulk grain random sampling method based on binocular vision and image segmentation technology | |
| CN113807327A (en) | Deep learning side face image processing method and system based on light compensation |
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 |