CN114494636A - Method and device for automatically constructing house type based on picture and computer equipment - Google Patents
Method and device for automatically constructing house type based on picture and computer equipment Download PDFInfo
- Publication number
- CN114494636A CN114494636A CN202111638391.6A CN202111638391A CN114494636A CN 114494636 A CN114494636 A CN 114494636A CN 202111638391 A CN202111638391 A CN 202111638391A CN 114494636 A CN114494636 A CN 114494636A
- Authority
- CN
- China
- Prior art keywords
- elements
- house type
- wall
- window
- door
- 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.)
- Pending
Links
Images
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
- G06T17/10—Constructive solid geometry [CSG] using solid primitives, e.g. cylinders, cubes
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/243—Classification techniques relating to the number of classes
- G06F18/2431—Multiple classes
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/10—Geometric CAD
- G06F30/13—Architectural design, e.g. computer-aided architectural design [CAAD] related to design of buildings, bridges, landscapes, production plants or roads
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/20—Image enhancement or restoration by the use of local operators
- G06T5/30—Erosion or dilatation, e.g. thinning
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/187—Segmentation; Edge detection involving region growing; involving region merging; involving connected component labelling
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2210/00—Indexing scheme for image generation or computer graphics
- G06T2210/04—Architectural design, interior design
Abstract
The invention discloses a method and a device for automatically constructing a house type based on pictures and computer equipment, wherein the method comprises the following steps: carrying out house type element classification on the house type picture by utilizing a semantic segmentation model; extracting candidate key points in the family element classification result by using a key point detection model, screening the candidate key points, and breaking and clustering the wall body based on the determined key points to construct wall body elements; detecting an image connected domain according to the classification result of the house type elements to construct door elements, window elements and column elements, and associating the three elements with wall elements; and correcting the constructed wall elements, door elements, window elements and column elements by combining the correlation results to obtain a constructed house type diagram. The method can improve the accuracy and efficiency of the house type automatic construction.
Description
Technical Field
The invention belongs to the field of house type design, and particularly relates to a method and a device for automatically constructing house type based on pictures and computer equipment.
Background
Home decoration design is a time consuming and laborious task, and a complete design scheme usually requires a significant amount of time for the designer to design and modify. With the continuous progress of computer technology, intelligent design tools are also continuously innovated and optimized, and designers are helped to improve the design efficiency.
Common intelligent design capabilities include that a set of house types and some furniture are given, reasonable placement of the furniture is automatically achieved through an algorithm, and for example, a set of house types are given, house type structures are intelligently identified through the algorithm, and therefore automatic layout of water and electricity point positions is achieved. However, no matter the intelligent placement of home decoration or the automatic layout of water and electricity points, the basic house type element description cannot be separated, the user firstly needs to directly operate a visual house type scheme (2d or 3d scene) in a design tool to perform subsequent design links, and all the steps cannot be separated from the extraction and construction of the house type elements.
The traditional house type construction usually adopts a copy mode to restore the house type information described in the drawing in a stroke-by-stroke mode in a design tool based on a given house type drawing through a great amount of manpower. Furthermore, based on industry specifications, house layout often has some similarity in structure, such as building wall thickness often at 120mm or 240 mm. For a large number of similar house-type drawings, the practitioner needs to spend almost the same time to mechanically repeat the whole drawing process, and the design efficiency is extremely low.
In order to improve the design efficiency, the conventional house type automatic construction method usually implements recognition of the house type elements by performing some fixed image operations on the house type graph based on fixed rules, such as a method for generating a three-dimensional house type based on photographed house type graph recognition disclosed in patent document CN105279787A, and further such as a method for automatically extracting the house type elements based on machine vision disclosed in patent document CN 108763606A. However, conventional rule-based recognition has difficulty achieving satisfactory results due to the wide source of family diagrams and their diverse styles.
In summary, the requirements of high-quality and efficient household type scheme construction cannot be well met no matter the household type is drawn manually through a large amount of manpower or the household type is identified through rules.
Disclosure of Invention
In view of the above technical problems, an object of the present invention is to provide a method and an apparatus for automatically constructing a house type based on pictures, which improve accuracy and efficiency of automatically constructing a house type.
In order to achieve the above object, a first aspect of the present invention provides an automatic house type construction method based on pictures, including the following steps:
carrying out house type element classification on the house type picture by utilizing a semantic segmentation model;
extracting candidate key points in the family element classification result by using a key point detection model, screening the candidate key points, and breaking and clustering the wall body based on the determined key points to construct wall body elements;
detecting an image connected domain according to the classification result of the house type elements to construct door elements, window elements and column elements, and associating the three elements with wall elements;
and correcting the constructed wall elements, door elements, window elements and column elements by combining the correlation results to obtain a constructed house type diagram.
In one embodiment, before the user type element classification, the user type picture is subjected to normalization operation, including the steps of resetting the picture size of the user type picture and carrying out pixel normalization on the user type picture according to a standard value and a variance;
and inputting the house type picture after the normalization operation into a semantic segmentation model for carrying out house type element classification.
Before extracting candidate key points, carrying out binarization and normalization operation on the house type element classification result, wherein the normalization operation comprises resetting the picture size of the house type element classification result and carrying out pixel standardization on the house type element classification result according to a standard value and a variance;
and inputting the classification result of the house type elements after the normalization operation into a key point detection model for candidate key point extraction.
In one embodiment, the semantic segmentation sub-model is based on a PSPNet model and is constructed through parameter optimization; the keypoint detection model is based on a CenterNet model and is constructed through parameter optimization.
In one embodiment, candidate key points are screened according to a set threshold to determine key points, after wall marks in the classification result of the type elements are broken based on the key points, pixels belonging to the same wall are clustered, and the thickness of a starting pixel, an ending pixel and a pixel region of each cluster is calculated to construct wall elements.
In one embodiment, after the wall elements are constructed, the bearing wall area in the house type element classification result is extracted, the overlapping area of the wall elements and the bearing wall area is judged, and the wall elements larger than a set first overlapping area threshold value are used as the bearing wall elements.
In one embodiment, the detecting the connected image domain according to the classification result of the house type elements includes:
clustering adjacent pixels with the same pixel value into a pixel connected region according to the house type element classification result, carrying out image expansion on the pixel region, screening the image expansion result according to a set threshold value, and combining the pixel connected region obtained by screening with corresponding pixel classification to construct a door element, a window element and a column element.
In one embodiment, associating the door element, the window element, the column element with the wall element comprises:
and after contour extraction is carried out on each pixel connected region, the overlapping area of the wall element and the door element, the window element and the column element is calculated, and if the overlapping area exceeds a second overlapping area threshold value, the wall element and the door element, the window element and the column element are associated.
In one embodiment, the correcting the constructed wall element, door element, window element and column element by combining the correlation result comprises:
calculating closed rings for wall elements, wherein each closed ring forms a room element, performing collinear judgment on all the wall elements, combining the same type of walls, and taking the thickness of the longest wall in the collinear walls as the thickness of the combined wall;
aiming at the door elements and the window elements, calculating door center lines and window center lines according to the door and window attached walls and carrying out normalization processing according to the correlation results of the door elements and the window elements and the wall elements, and comprising the following steps: the lengths of the door elements and the window elements cannot exceed the lengths of the wall bodies, collinear doors and collinear windows are combined, the door types and the window types of the combined structures are set according to the door elements and the window elements with the largest areas, the door elements or the window elements with the illegal intersection of the center lines of the doors and the windows are removed, and the door elements or the window elements with the larger areas in the intersection elements are reserved;
aiming at the column elements, when the column elements and the wall body have intersection, calculating a circumscribed rectangle of a contour formed after the column is cut by the wall body as the outer contour of the column; and when the column elements do not intersect with the wall body, calculating the circumscribed rectangle of the column elements as the outer contour.
In order to achieve the above object, a second embodiment of the invention provides a device for automatically constructing a picture-based house type, comprising:
the family element classification module is used for classifying the family elements of the family pictures by utilizing the semantic segmentation model;
the family element construction module is used for extracting candidate key points in the family element classification result by using the key point detection model, screening the candidate key points, and breaking and clustering the wall body based on the determined key points to construct wall body elements; the system is also used for detecting image connected domains according to the classification result of the house type elements so as to construct door elements, window elements and column elements and associate the three elements with wall elements;
and the house type correction module corrects the constructed wall elements, door elements, window elements and column elements by combining the correlation results to obtain a constructed house type graph.
To achieve the above object, a third aspect of the embodiments provides a computer device, which includes a memory, a processor, and a computer program stored in the memory and executable on the processor, and when the processor executes the computer program, the processor implements the steps of the picture-based house-type automatic construction method according to the first aspect.
Compared with the prior art, the invention has the beneficial effects that at least:
by adopting a machine learning mode, the characteristics of the house type elements are learned through mass house type picture data, so that a semantic segmentation model and a key point detection model have stronger generalization performance, and the house type element classification and key point detection of various styles of house type pictures can be realized;
the method has the advantages that the key point detection model is utilized to realize instance-level segmentation of the wall elements, and a single wall element is constructed based on the instance-level segmentation result, so that a large amount of invalid and complex post-processing logics are avoided, and the importing efficiency and accuracy are improved;
the automatic construction of the house type is realized by combining a semantic segmentation model to realize the classification of the house type elements, the construction of wall elements realized by a key point detection model, the construction of door elements, window elements and column elements realized by connected domain detection and the correction processing of the elements, the time for drawing the house type by a user is saved, and the house type drawing efficiency is improved.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the embodiments or the description of the prior art will be briefly described below, it is obvious that the drawings in the following description are only some embodiments of the present invention, and it is obvious for those skilled in the art that other drawings can be obtained according to these drawings without creative efforts.
FIG. 1 is a flow chart of an automatic construction method of a picture-based house type provided by an embodiment;
FIG. 2 is a diagram illustrating the result of the classification of the house type elements of the house type picture according to the embodiment;
FIG. 3 shows the construction of wall elements according to the present embodiment;
fig. 4 is a schematic structural diagram of a picture-based house-type automatic construction device according to an embodiment.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention will be further described in detail with reference to the accompanying drawings and examples. It should be understood that the detailed description and specific examples, while indicating the scope of the invention, are intended for purposes of illustration only and are not intended to limit the scope of the invention.
In order to solve the problems that a designer cannot search a house type and needs to spend a large amount of time to manually draw the house type and the traditional house type construction effect based on rule import is poor, the embodiment provides an end-to-end automatic house type construction method and device based on pictures.
FIG. 1 is a flowchart illustrating an embodiment of a method for automatically constructing a picture-based house type. As shown in fig. 1, the method for automatically constructing a house type based on pictures provided by the embodiment includes the following steps:
The house type element classification refers to performing pixel classification on any input house type picture, and determining the house type elements according to pixel classification results, wherein the house type elements comprise wall elements, door elements, window elements, column elements and the like, and areas corresponding to the elements are wall areas, door areas, window areas, column areas and the like.
Specifically, the house type element classification process comprises the following steps:
firstly, carrying out normalization operation on the house type picture, specifically resetting the picture size of the house type picture, and carrying out pixel standardization on the house type picture according to a standard value and a variance;
then, inputting the house type picture subjected to normalization operation into a semantic segmentation model, calculating to obtain the probability that each pixel belongs to different semantic categories, comparing the probabilities of the different semantic categories aiming at each pixel point, and selecting the maximum value of the probabilities as the house type element category of the current pixel point, so that the house type element classification is realized, and fig. 2 is a visual diagram of an exemplary house type element classification result.
In the embodiment, the semantic segmentation model is constructed based on a PSPNet model, and the specific construction process comprises the following steps: and taking the house type picture with the house type element real label as sample data, inputting the sample data into the PSPNet model, optimizing the PSPNet model parameters by taking the minimum cross entropy loss function of the prediction label of the house type element and the real label output by the model as an optimization target, and after the optimization is finished, extracting the PSPNet model determined by the parameters as a semantic segmentation model.
The semantic segmentation model is constructed through massive pictures and scheme learning, so that the semantic segmentation model has strong generalization capability, can realize the classification of the house type elements of the house type pictures of different styles, can ensure the classification accuracy, and can improve the classification efficiency.
And 2, building the house type elements, namely detecting and constructing wall elements, door elements, window elements and column elements by using the key point detection model and the connected domain, and associating the door elements, the window elements and the column elements with the wall elements.
In the embodiment, candidate key points in the family element classification result are extracted by using the key point detection model, and after the candidate key points are screened, the wall is broken and clustered based on the determined key points to construct wall elements. The specific process comprises the following steps:
firstly, binarization and normalization operations are carried out on the house type element classification result, wherein the normalization operations comprise resetting the picture size of the house type element classification result and carrying out pixel standardization on the house type element classification result according to a standard value and a variance.
Then, inputting the normalized user type element classification result picture into a key point detection model, obtaining a plurality of candidate key points and confidence degrees thereof through calculation, and screening a plurality of key points meeting conditions from the candidate key points based on a certain threshold value.
And then, after breaking the wall marks in the picture of the classification result of the user type elements based on the key points, clustering the pixels belonging to the same wall, and calculating the initial pixels, the final pixels and the thickness of the pixel region of each cluster to construct the wall elements. In an embodiment, the thicknesses of the start pixel, the end pixel and the pixel region may be determined according to the position relationship of the pixel points in the cluster.
And finally, extracting a bearing wall area in the house type element classification result, judging the overlapping area of the wall element and the bearing wall area, and taking the wall element larger than a set first overlapping area threshold value as the bearing wall element.
In the embodiment, the first overlap area threshold is set according to an actual application situation, and is not limited specifically. The key point detection model is constructed based on a CenterNet model, and the specific construction process comprises the following steps: and taking the house type picture marked with the key points of the house type area as sample data, inputting the sample data into the CenterNet model, integrating the original loss function of the CenterNet model to optimize the parameters of the CenterNet model, and after the optimization is finished, extracting the CenterNet model determined by the parameters as the key point detection model. Fig. 3 exemplarily shows candidate keypoints obtained by performing keypoint detection by using the keypoint detection model, and the visualization result of the wall keypoints determined after screening is shown in fig. 3.
In the embodiment, the image connected domain detection is carried out according to the classification result of the house type elements so as to construct door elements, window elements and column elements, and the three elements are associated with wall elements. The specific process comprises the following steps:
firstly, according to the subdivision categories of the house type elements, the subdivision categories comprise a sliding door, a single-door, a double-door, a common window, a bay window, a French window, a flue, a common pillar and the like, and an image area with the corresponding pixel category of the subdivision categories is screened from the classification results of the house type elements.
Then, clustering adjacent pixels with the same pixel value into a pixel connected region, carrying out image expansion on the pixel region, screening image expansion results according to a set threshold value, and combining the pixel connected region obtained by screening with corresponding pixel classification to construct a door element, a window element and a column element.
And finally, after contour extraction is carried out on each pixel communication region, the overlapping area of the wall element and the door element, the window element and the column element is calculated, if the overlapping area exceeds a second overlapping area threshold value, the wall element and the door element, the window element and the column element are associated, and therefore association binding of the door element, the window element and the column element and the wall instance is achieved.
And 3, correcting the house type, namely correcting the constructed wall elements, door elements, window elements and column elements by combining the correlation results to obtain a constructed house type diagram.
The house type elements constructed through step 2 are relatively rough, so secondary correction is also needed to be carried out on the house type elements. The specific correction process comprises:
aiming at wall elements, closed rings can be calculated for the wall elements by adopting a ring search algorithm, each closed ring forms a room element, all the wall elements are subjected to collinear judgment, the walls of the same type are combined, and the thickness of the longest wall in the collinear walls is taken as the thickness of the combined wall;
aiming at the door elements and the window elements, calculating door center lines and window center lines according to the door and window attached walls and carrying out normalization processing according to the correlation results of the door elements and the window elements and the wall elements, and comprising the following steps: the lengths of the door elements and the window elements cannot exceed the lengths of the wall bodies, collinear doors and collinear windows are combined, the door types and the window types of the combined structures are set according to the door elements and the window elements with the largest areas, the door elements or the window elements with the illegal intersection of the center lines of the doors and the windows are removed, and the door elements or the window elements with the larger areas in the intersection elements are reserved;
aiming at the column elements, when the column elements and the wall body have intersection, the outline of the wall body can cut the outline of the column, and a circumscribed rectangle of the outline formed after the column is cut by the wall body is calculated to be used as the outline of the column; and when the column elements do not intersect with the wall body, calculating the circumscribed rectangle of the column elements as the outer contour.
And finally, all the house type elements in the given input house type picture are extracted, wherein the house type elements comprise a wall body, a door, a window, a column and a room, and the obtained two-dimensional vector information of the house type elements is subjected to certain visualization in a design tool to obtain a complete house type scheme for layout operation of a user.
Compared with the traditional mode of identifying the house type elements in the pictures through rules, the house type automatic construction method based on the pictures provided by the embodiment adopts a machine learning mode, learns the features of the house type elements through mass house type picture data, enables the semantic segmentation model and the key point detection model to have stronger generalization performance, and can realize the house type element classification and key point detection of the house type pictures with various styles.
Compared with other recognition modes combining semantic segmentation and post-processing rules, the method for automatically constructing the house type based on the pictures provided by the embodiment realizes instance-level segmentation of the wall elements by using the key point detection model, and constructs a single wall element based on the instance-level segmentation result, so that a large amount of invalid and complex post-processing logics are avoided, and the importing efficiency and the importing accuracy are improved.
The automatic house type construction method based on the pictures provided by the embodiment combines the semantic segmentation model to realize the house type element classification, the wall element construction realized by the key point detection model, the door element, window element and column element construction realized by the connected domain detection, and the element correction processing, realizes the automatic house type construction, saves the house type drawing time of the user, and improves the house type drawing efficiency.
Fig. 4 is a schematic structural diagram of a picture-based house-type automatic configuration device according to an embodiment. As shown in fig. 4, the house type automatic configuration apparatus according to the embodiment includes:
the family element classification module is used for classifying the family elements of the family pictures by utilizing the semantic segmentation model;
the family element construction module is used for extracting candidate key points in the family element classification result by using the key point detection model, screening the candidate key points, and breaking and clustering the wall body based on the determined key points to construct wall body elements; the system is also used for detecting image connected domains according to the classification result of the house type elements so as to construct door elements, window elements and column elements and associate the three elements with wall elements;
and the house type correction module corrects the constructed wall elements, door elements, window elements and column elements by combining the correlation results to obtain a constructed house type graph.
It should be noted that, when the home automation configuration device provided in the foregoing embodiment performs the home automation configuration, the division of each function module is used as an example, and the function distribution may be performed by different function modules as needed, that is, the internal structure of the terminal or the server is divided into different function modules to perform all or part of the functions described above. In addition, the house type automatic construction device provided by the above embodiment and the house type automatic construction method embodiment belong to the same concept, and the specific implementation process is described in detail in the house type automatic construction method embodiment, and is not described again here.
An embodiment further provides a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the above-mentioned picture-based house type automatic configuration method when executing the computer program, and the method includes the following steps:
And 2, building the house type elements, namely detecting and constructing wall elements, door elements, window elements and column elements by using the key point detection model and the connected domain, and associating the door elements, the window elements and the column elements with the wall elements.
And 3, correcting the house type, namely correcting the constructed wall elements, door elements, window elements and column elements by combining the correlation results to obtain a constructed house type diagram.
It should be noted that the computer memory may be volatile memory at the near end, such as RAM, non-volatile memory, such as ROM, FLASH, floppy disk, mechanical hard disk, etc., or may be a remote storage cloud. The computer processor may be a Central Processing Unit (CPU), a Microprocessor (MPU), a Digital Signal Processor (DSP), or a Field Programmable Gate Array (FPGA), i.e., steps of the picture-based house-type automatic construction may be implemented by these processors.
The above-mentioned embodiments are intended to illustrate the technical solutions and advantages of the present invention, and it should be understood that the above-mentioned embodiments are only the most preferred embodiments of the present invention, and are not intended to limit the present invention, and any modifications, additions, equivalents, etc. made within the scope of the principles of the present invention should be included in the scope of the present invention.
Claims (10)
1. A house type automatic construction method based on pictures is characterized by comprising the following steps:
carrying out house type element classification on the house type picture by utilizing a semantic segmentation model;
extracting candidate key points in the family element classification result by using a key point detection model, screening the candidate key points, and breaking and clustering the wall body based on the determined key points to construct wall body elements;
detecting an image connected domain according to the classification result of the house type elements to construct door elements, window elements and column elements, and associating the three elements with wall elements;
and correcting the constructed wall elements, door elements, window elements and column elements by combining the correlation results to obtain a constructed house type diagram.
2. The method of claim 1, wherein prior to the classification of the house type elements, performing a normalization operation on the house type picture, including resizing the house type picture, and performing pixel normalization on the house type picture according to a standard value and a variance;
and inputting the house type picture after the normalization operation into a semantic segmentation model for carrying out house type element classification.
Before extracting candidate key points, carrying out binarization and normalization operation on the house type element classification result, wherein the normalization operation comprises resetting the picture size of the house type element classification result and carrying out pixel standardization on the house type element classification result according to a standard value and a variance;
and inputting the classification result of the house type elements after the normalization operation into a key point detection model for candidate key point extraction.
3. The automatic house type construction method based on pictures as claimed in claim 1, characterized in that the semantic segmentation sub-model is constructed based on PSPNet model and through parameter optimization;
the keypoint detection model is based on a CenterNet model and is constructed through parameter optimization.
4. The method according to claim 1, wherein candidate key points are screened according to a set threshold to determine key points, after wall marks in the classification result of the household elements are broken based on the key points, pixels belonging to the same wall are clustered, and the thickness of a start pixel, an end pixel and a pixel region of each cluster is calculated to construct the wall elements.
5. The automatic house type construction method based on pictures as claimed in claim 1, characterized in that after the wall elements are constructed, the bearing wall area in the house type element classification result is extracted, the overlapping area of the wall elements and the bearing wall area is judged, and the wall elements larger than the set first overlapping area threshold are used as the bearing wall elements.
6. The method for automatically constructing a house type based on pictures according to claim 1, wherein the detecting of the image connected domain according to the house type element classification result comprises:
clustering adjacent pixels with the same pixel value into a pixel connected region according to the house type element classification result, carrying out image expansion on the pixel region, screening the image expansion result according to a set threshold value, and combining the pixel connected region obtained by screening with corresponding pixel classification to construct a door element, a window element and a column element.
7. The method of claim 6, wherein associating the door element, the window element, and the pillar element with the wall element comprises:
and after contour extraction is carried out on each pixel connected region, the overlapping area of the wall element and the door element, the window element and the column element is calculated, and if the overlapping area exceeds a second overlapping area threshold value, the wall element and the door element, the window element and the column element are associated.
8. The method for automatically constructing a house type based on pictures according to claim 1, wherein the correcting the constructed wall elements, door elements, window elements and column elements by combining the correlation results comprises:
calculating closed rings for wall elements, wherein each closed ring forms a room element, performing collinear judgment on all the wall elements, combining the same type of walls, and taking the thickness of the longest wall in the collinear walls as the thickness of the combined wall;
aiming at the door elements and the window elements, according to the correlation results of the door elements and the window elements and the wall elements, calculating door center lines and window center lines according to the walls attached to the doors and the windows, and performing normalization processing, wherein the method comprises the following steps of: the lengths of the door elements and the window elements cannot exceed the lengths of the wall bodies, collinear doors and collinear windows are combined, the door types and the window types of the combined structures are set according to the door elements and the window elements with the largest areas, the door elements or the window elements with the illegal intersection of the center lines of the doors and the windows are removed, and the door elements or the window elements with the larger areas in the intersection elements are reserved;
aiming at the column elements, when the column elements and the wall body have intersection, calculating a circumscribed rectangle of a contour formed after the column is cut by the wall body as the outer contour of the column; and when the column elements do not intersect with the wall body, calculating the circumscribed rectangle of the column elements as the outer contour.
9. An apparatus for automatically constructing a house type based on pictures, comprising:
the family element classification module is used for classifying the family elements of the family pictures by utilizing the semantic segmentation model;
the family element construction module is used for extracting candidate key points in the family element classification result by using the key point detection model, screening the candidate key points, and breaking and clustering the wall body based on the determined key points to construct wall body elements; the system is also used for detecting image connected domains according to the classification result of the house type elements so as to construct door elements, window elements and column elements and associate the three elements with wall elements;
and the house type correction module corrects the constructed wall elements, door elements, window elements and column elements by combining the correlation results to obtain a constructed house type graph.
10. A computer device comprising a memory, a processor and a computer program stored in said memory and executable on said processor, characterized in that said processor implements the steps of the picture-based house-type auto-construction method according to any of claims 1-8 when executing said computer program.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111638391.6A CN114494636A (en) | 2021-12-29 | 2021-12-29 | Method and device for automatically constructing house type based on picture and computer equipment |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111638391.6A CN114494636A (en) | 2021-12-29 | 2021-12-29 | Method and device for automatically constructing house type based on picture and computer equipment |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN114494636A true CN114494636A (en) | 2022-05-13 |
Family
ID=81508637
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202111638391.6A Pending CN114494636A (en) | 2021-12-29 | 2021-12-29 | Method and device for automatically constructing house type based on picture and computer equipment |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114494636A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115205418A (en) * | 2022-09-15 | 2022-10-18 | 武汉智筑完美家居科技有限公司 | Home graph reconstruction method and device, electronic equipment and storage medium |
-
2021
- 2021-12-29 CN CN202111638391.6A patent/CN114494636A/en active Pending
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115205418A (en) * | 2022-09-15 | 2022-10-18 | 武汉智筑完美家居科技有限公司 | Home graph reconstruction method and device, electronic equipment and storage medium |
| CN115205418B (en) * | 2022-09-15 | 2022-12-13 | 武汉智筑完美家居科技有限公司 | Household graph reconstruction method and device, electronic equipment and storage medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Hui et al. | Effective building extraction from high-resolution remote sensing images with multitask driven deep neural network | |
| Zhang et al. | Deep hierarchical guidance and regularization learning for end-to-end depth estimation | |
| Lv et al. | Residential floor plan recognition and reconstruction | |
| Lu et al. | Data-driven floor plan understanding in rural residential buildings via deep recognition | |
| CN111125397B (en) | Cloth image retrieval method based on convolutional neural network | |
| CN111862119A (en) | Semantic information extraction method based on Mask-RCNN | |
| CN110929696A (en) | Remote sensing image semantic segmentation method based on multi-mode attention and self-adaptive fusion | |
| CN111611643A (en) | Family type vectorization data obtaining method and device, electronic equipment and storage medium | |
| WO2021248269A1 (en) | High-precision identification method and system for transformer device | |
| CN112068555A (en) | Voice control type mobile robot based on semantic SLAM method | |
| CN101872415A (en) | Video copying detection method being suitable for IPTV | |
| CN114724169A (en) | Identification method and system of two-dimensional drawing component | |
| Xu et al. | 3D point cloud plane segmentation method based on RANSAC and support vector machine | |
| Zabulis et al. | Correspondence-free pose estimation for 3D objects from noisy depth data | |
| Dong et al. | Learning regional purity for instance segmentation on 3d point clouds | |
| CN114494636A (en) | Method and device for automatically constructing house type based on picture and computer equipment | |
| CN110097603B (en) | Fashionable image dominant hue analysis method | |
| CN111782817A (en) | Knowledge graph construction method and device for information system and electronic equipment | |
| CN113239785A (en) | Method, system, storage medium and electronic device for automatically identifying house type based on convolutional neural network | |
| Perez-Perez et al. | Artificial neural network for semantic segmentation of built environments for automated Scan2BIM | |
| Tang et al. | Automatic structural scene digitalization | |
| Liu et al. | Moving object detection based on improved ViBe algorithm | |
| CN110705339A (en) | C-C3D-based sign language identification method | |
| CN114707174A (en) | Data processing method and device, electronic equipment and storage medium | |
| Collins et al. | Towards applicable Scan-to-BIM and Scan-to-Floorplan: an end-to-end experiment |
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 |