CN111985518A - Door and window detection method and model training method and device thereof - Google Patents
Door and window detection method and model training method and device thereof Download PDFInfo
- Publication number
- CN111985518A CN111985518A CN202010100580.7A CN202010100580A CN111985518A CN 111985518 A CN111985518 A CN 111985518A CN 202010100580 A CN202010100580 A CN 202010100580A CN 111985518 A CN111985518 A CN 111985518A
- Authority
- CN
- China
- Prior art keywords
- door
- window
- model
- training
- detection
- 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
- 238000001514 detection method Methods 0.000 title claims abstract description 116
- 238000012549 training Methods 0.000 title claims abstract description 105
- 238000000034 method Methods 0.000 title claims abstract description 76
- 238000003062 neural network model Methods 0.000 claims abstract description 40
- 238000000605 extraction Methods 0.000 claims abstract description 12
- 238000012545 processing Methods 0.000 claims abstract description 7
- 230000011218 segmentation Effects 0.000 claims description 21
- 238000010586 diagram Methods 0.000 claims description 12
- 238000004590 computer program Methods 0.000 claims description 7
- 238000013461 design Methods 0.000 abstract description 23
- 238000012938 design process Methods 0.000 abstract description 6
- 238000012360 testing method Methods 0.000 abstract 1
- 230000008569 process Effects 0.000 description 21
- 238000005034 decoration Methods 0.000 description 15
- 230000000694 effects Effects 0.000 description 8
- 238000004891 communication Methods 0.000 description 7
- 230000006870 function Effects 0.000 description 4
- 238000005457 optimization Methods 0.000 description 4
- 230000008878 coupling Effects 0.000 description 3
- 238000010168 coupling process Methods 0.000 description 3
- 238000005859 coupling reaction Methods 0.000 description 3
- 239000002245 particle Substances 0.000 description 3
- 238000013528 artificial neural network Methods 0.000 description 2
- 238000009877 rendering Methods 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 238000006243 chemical reaction Methods 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 238000013527 convolutional neural network Methods 0.000 description 1
- 230000004069 differentiation Effects 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000000802 evaporation-induced self-assembly Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/004—Artificial life, i.e. computing arrangements simulating life
- G06N3/006—Artificial life, i.e. computing arrangements simulating life based on simulated virtual individual or collective life forms, e.g. social simulations or particle swarm optimisation [PSO]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/40—Extraction of image or video features
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- General Physics & Mathematics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- General Engineering & Computer Science (AREA)
- Evolutionary Computation (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Molecular Biology (AREA)
- Bioinformatics & Computational Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Software Systems (AREA)
- Health & Medical Sciences (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- General Health & Medical Sciences (AREA)
- Evolutionary Biology (AREA)
- Computing Systems (AREA)
- Mathematical Physics (AREA)
- Multimedia (AREA)
- Image Analysis (AREA)
Abstract
The invention provides a door and window detection method and a device for model training thereof, relating to the technical field of image processing, wherein the model training method comprises the following steps: determining a door and window sample image based on a two-dimensional plane house-type graph in a preset door and window training set; carrying out feature extraction on the door and window sample image, and inputting an extracted door and window feature result into an initial neural network model for training; and when the training result of the initial neural network model meets a preset expected threshold value, obtaining a model for door and window detection. And inputting the two-dimensional plane house type graph to be detected into a door and window detection model which is trained in advance, and outputting a door and window detection result. This scheme is at the in-process that door and window detected, and the testing result contains door and window position and door and window type accessible multidimension's door and window and detects, can further promote door and window's speed and the degree of accuracy in the house type picture, has alleviated the problem that can occupy more design time when door and window discernment in the house ornamentation design process.
Description
Technical Field
The invention relates to the technical field of image processing, in particular to a door and window detection method and a device for model training thereof.
Background
The door and window is an important link in the home decoration design, and because the door and window directly influences the lighting of house types and the access of people, the door and window is one of the home decoration parts with the highest utilization rate in daily life, whether the door and window design is scientific, reasonable or not is directly related to the overall effect of the home decoration design. Because the door and window occupies less area in the house type picture, the characteristics are relatively unobvious, the related home decoration design tools in the prior art have very low door and window identification accuracy, even common windows such as common windows and bay windows cannot be automatically identified, and the positions of the door and window can only be changed by designers in a manual adjustment mode, so that the door and window occupies more design time in the home decoration design process, and the design efficiency is low.
Disclosure of Invention
In view of the above, an object of the present invention is to provide a method and an apparatus for detecting doors and windows and a method and an apparatus for training models thereof, wherein trained related neural network models are used to identify doors and windows in a house type diagram, and target area detection and semantic segmentation detection are used to perform multi-dimensional detection of the doors and windows, so that the speed and accuracy of the doors and windows in the house type diagram are further improved, and the problem that more design time is occupied during door and window identification in a home decoration design process is solved.
In a first aspect, an embodiment of the present invention provides a model training method for door and window detection, where the method includes:
determining a door and window sample image based on a two-dimensional plane house-type graph in a preset door and window training set;
carrying out feature extraction on the door and window sample image, and inputting an extracted door and window feature result into an initial neural network model for training;
and when the training result of the initial neural network model meets a preset expected threshold value, obtaining a model for door and window detection.
In some embodiments, the step of determining the door/window sample image based on the two-dimensional planar floor plan in the preset door/window training set includes:
traversing the two-dimensional plane house type graph in the door and window training set, and identifying doors and windows in the two-dimensional plane house type graph to obtain a door and window identification result;
determining the door and window sample image containing the door and window in the door and window identification result as a door and window positive sample image; determining the door and window sample image which does not contain the door and window in the door and window identification result as a door and window negative sample image;
and taking the door and window positive sample image and the door and window negative sample image as door and window sample images.
In some embodiments, the step of extracting the features of the door and window sample image and inputting the extracted door and window feature result into the initial neural network model for training includes:
initializing a neural network model, wherein the network model comprises a target detection model and a semantic segmentation model;
respectively inputting the door and window sample images into a target detection model and a semantic segmentation model for training, and outputting the identification results of the door and window sample images;
judging the identification result of the door and window sample image and the door and window real result of the door and window sample image to obtain a judgment result;
and adjusting the preset training parameters of the neural network model according to the judgment result.
In some embodiments, the target detection model is a yolo v3 neural network model; the semantic segmentation model is a mask-rcnn neural network model.
In a second aspect, an embodiment of the present invention provides a door and window detecting method, where the method includes:
acquiring a two-dimensional planar house pattern to be detected;
inputting a two-dimensional plane house type graph to be detected into a door and window detection model which is trained in advance, and outputting a door and window detection result; the door and window detection model is obtained by training through the model training method for door and window detection in any one of the embodiments of the first aspect.
In some embodiments, the step of inputting the two-dimensional flat house type diagram to be detected into the door and window detection model which is trained in advance and outputting the door and window detection result includes:
detecting the two-dimensional plane house type graph to be detected through a target detection model in a door and window detection model to obtain the door and window position in the two-dimensional plane house type graph to be detected;
detecting the two-dimensional plane house type graph to be detected through a semantic segmentation model in a door and window detection model to obtain the door and window type in the two-dimensional plane house type graph to be detected;
and marking the door and window position and the door and window type in the two-dimensional plane house type graph to be detected, and outputting the identification result of the door and window sample image.
In a third aspect, an embodiment of the present invention provides a model training apparatus for door and window detection, where the apparatus includes:
the system comprises a sample acquisition module, a data acquisition module and a data processing module, wherein the sample acquisition module is used for determining a door and window sample image based on a two-dimensional plane house-type diagram in a preset door and window training set;
the model training module is used for extracting the characteristics of the door and window sample images and inputting the extracted door and window characteristic results into the initial neural network model for training;
and the model acquisition module is used for obtaining a model for door and window detection when the training result of the preset neural network model meets a preset expected threshold value.
In a fourth aspect, an embodiment of the present invention provides a door and window detecting apparatus, including:
the acquisition module of the two-dimensional house type graph to be detected is used for acquiring the two-dimensional house type graph to be detected;
the door and window detection module is used for inputting the two-dimensional plane house type graph to be detected into a door and window detection model which is trained in advance and outputting a door and window detection result; the door and window detection model is obtained by training through the model training method for door and window detection in any one of the first aspect.
In a fifth aspect, an embodiment of the present invention provides an electronic device, including: a processor and a storage device; the storage means has stored thereon a computer program which, when executed by the processor, performs the method as provided in the first and second aspects.
In a sixth aspect, the present invention provides a computer-readable storage medium, on which a computer program is stored, where the computer program is executed by a processor to perform the methods provided in the first and second aspects.
The embodiment of the invention has the following beneficial effects: the embodiment of the invention provides a door and window detection method and a device for model training thereof.A door and window sample image is determined based on a two-dimensional plane house-type diagram in a preset door and window training set in the model training process for door and window detection; and then, carrying out feature extraction on the door and window sample image, and obtaining a model for door and window detection when the training result of the initial neural network model meets a preset expected threshold value. When the trained door and window detection model is used for detecting doors and windows, firstly, a two-dimensional plane house type graph to be detected is obtained; and then inputting the two-dimensional plane house type graph to be detected into a door and window detection model which is trained in advance, and outputting a door and window detection result. In the process of detecting the door and window, target area detection and semantic segmentation detection are adopted to carry out multi-dimensional detection on the door and window, and the detection result comprises the door and window position and the door and window type. Through the door and window detection of multidimension degree, can further promote door and window's speed and the degree of accuracy in the house type picture, alleviated the house ornamentation design in-process and can occupy more design time's problem when door and window discernment.
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.
In order to make the aforementioned and other objects, features and advantages of the present invention comprehensible, preferred embodiments accompanied with figures are described in detail below.
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 description of the embodiments or the prior art will be briefly described below, and it is obvious that the drawings in the following description are some embodiments of the present invention, and other drawings can be obtained by those skilled in the art without creative efforts.
FIG. 1 is a flowchart of a model training method for door and window detection according to an embodiment of the present invention;
fig. 2 is a flowchart of step S101 in the model training method for door and window detection according to the embodiment of the present invention;
fig. 3 is a flowchart of step S102 in the model training method for door and window detection according to the embodiment of the present invention;
FIG. 4 is a flowchart of a door/window detecting method according to an embodiment of the present invention;
fig. 5 is a flowchart of step S402 in the door/window detecting method according to the embodiment of the present invention;
FIG. 6 is a schematic structural diagram of a model training device for door and window detection according to an embodiment of the present invention;
fig. 7 is a schematic structural view of a door/window detecting device according to an embodiment of the present invention;
fig. 8 is a schematic structural diagram of an electronic device according to an embodiment of the present invention.
Icon:
601-a sample acquisition module; 602-a model training module; 603-a model acquisition module; 701-acquiring a two-dimensional planar house figure to be detected; 702-a door and window detection module; 101-a processor; 102-a memory; 103-a bus; 104-communication interface.
Detailed Description
To make the objects, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings, and it is apparent that the described embodiments are some, but not all embodiments of the present invention. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
The door and window is an important link in the house decoration design, and is one of the house decoration parts with the highest utilization rate in daily life because the door and window directly influences the lighting of house types and the access of people. Because the door and window and other parts have a plurality of associations and restrictions, a plurality of parameters need to be considered, once deviation occurs in the design parameter setting of the door and window, chain reaction occurs in the home decoration design, the final home decoration design effect is influenced, and whether the door and window design is scientifically and reasonably directly related to the overall effect of the home decoration design can be seen.
Because the door and window occupies less area in the house type picture, the characteristics are relatively unobvious, the related house decoration design tools in the prior art have very low door and window identification accuracy, even common windows such as common windows and bay windows can not be automatically identified, and the positions of the door and window can only be changed by a designer in a manual adjustment mode, so that the door and window occupy more design time in the house decoration design process, the design efficiency is low, and it is seen that an efficient door and window identification mode is also lacked in the existing house decoration design process.
In view of the above problems in the door and window identification process in the field of home design, the present invention provides a door and window detection method and a method and an apparatus for model training thereof, where the technology may be applied to the door and window detection process in home design, and may be implemented by using related software or hardware, which is described below by way of an embodiment.
To facilitate understanding of the embodiment, first, a detailed description is given of a model training method for door and window detection disclosed in the embodiment of the present invention, where a flowchart of the method is shown in fig. 1, and the method includes:
and S101, determining a door and window sample image based on a two-dimensional plane house-type graph in a preset door and window training set.
The preset door and window training set is a set containing various door and window attributes, and the door and window attribute data is parameter data used for describing structures, sizes and the like of doors and windows, for example, the door and window attribute data can comprise types of the doors and windows and opening directions of the doors and windows. The door and window type is a kind of door and window, and taking a window as an example, the door and window type can comprise a common window and a bay window. The opening direction of the window is determined by the position of the window, such as left, right, up, or down. The opening direction of the window is usually related to the type attribute and other attributes of the window, the window is usually opened left or right, and the shaft connecting the window and the wall surface is arranged on the left or right side; the half-open window deploys the window axis on both the left and right sides. It should be noted that most window opening modes are not provided on the upper and lower sides.
The attribute data can be stored in a related attribute library after being collected and sorted, and the attribute library can be established by utilizing a database in the field of computers and is controlled by utilizing the computers.
And S102, extracting the characteristics of the door and window sample image, and inputting the extracted door and window characteristic result into an initial neural network model for training.
The initial neural network model is initialized before the training layer is input, and the state of the neural network model at the moment can be the state that the initialization process is just completed and the training is not started; or may be a state already in training.
The characteristic extraction of the door and window sample image is realized by adopting a related image recognition algorithm, for example, the door and window outline in the door and window sample image can be extracted by an outline extraction algorithm, and then the extraction result is subjected to characteristic distinguishing; the special color in the door and window sample image can be extracted; the pre-stored characteristic patterns can also be extracted through a matching algorithm. In the process of feature extraction, multi-dimensional detection of doors and windows can be carried out through target area detection and semantic segmentation detection.
After the door and window sample image with the extracted features is input to the neural network model, relevant parameters of the model are changed through relevant operation, and therefore the identification precision of the model is improved. For example, a penalty factor of the neural network model is optimized in the training process of the training image layer, the penalty factor is a parameter for representing the tolerance of the error, and the larger the value of the penalty factor is, the more intolerable the error occurs, and the overfitting phenomenon is relatively easier to occur; conversely, the smaller the value of the penalty factor, the more the under-fitting phenomenon is relatively easy to occur.
The process of inputting the door and window sample image subjected to the feature extraction into the preset neural network model for training also comprises the optimization of other parameters of the model, and the details are not repeated herein.
And step S103, obtaining a model for detecting the door and window when the training result of the initial neural network model meets a preset expected threshold value.
The model optimizes the relevant parameters in the training process, and the particle swarm optimization algorithm can be adopted to further optimize and calculate the relevant parameters in the neural network model in the optimization process. The particle swarm optimization algorithm is also called as a particle swarm algorithm, and can complete training of connection weights, structural design, learning rule adjustment, feature selection, initialization of the connection weights, rule extraction and the like in an artificial neural network.
In the training process of the model, the door and window sample image with the extracted features is input into a preset neural network model to obtain an output result, and the output result is judged so as to determine whether the performance of the model meets the requirements. For example, the training process of the model can be determined according to the numerical value of the loss function, and when the numerical value of the loss function reaches a preset threshold value, the performance of the model is considered to meet the requirement, so that the training of the model can be stopped, and the model for detecting the door and the window is obtained.
In the model training method for door and window detection provided by the embodiment of the invention, the characteristic extraction is carried out on the door and window sample image in the model training process, so that the model training is completed, the multi-dimensional detection of the door and window can be realized by target area detection and semantic segmentation detection, the speed and the accuracy of the door and window in the house type graph are further improved, and the problem that more design time is occupied during door and window identification in the house decoration design process is solved.
In some embodiments, the step S101 of determining a window and door sample image based on a two-dimensional planar floor plan in a preset window and door training set, as shown in fig. 2, includes:
step S201, traversing the two-dimensional plane house type graph in the door and window training set, and identifying doors and windows in the two-dimensional plane house type graph to obtain a door and window identification result.
The identification of the door and window includes two types of information: identification of location and identification of category. The identification of the door and window positions identifies the two-dimensional plane house type graph in the door and window training set through a correlation matching algorithm and a contour algorithm, the identification result is coordinate data and length data, the identification result is a closed area, and the area represents the door and window positions. The above-mentioned area can also be realized by means of manual calibration.
The identification of the door and window category is used to determine what the door and window are, for example, the window can be divided into a normal window and a bay window, and the window and the door also need to be distinguished as a door or a window, so the above process is a type distinction. The type can be distinguished through a pre-trained relevant model, the construction process of the model is not repeated in the embodiment, and doors and windows contained in a two-dimensional plane house type graph in a door and window training set can be distinguished through the model. The above-mentioned differentiation result can also be implemented manually.
Step S202, determining the door and window sample image containing the door and window in the door and window identification result as a door and window positive sample image; and determining the door and window sample image which does not contain the door and window in the door and window identification result as a door and window negative sample image.
Step S203, the door and window positive sample image and the door and window negative sample image are used as the door and window sample image.
The positive sample image and the negative sample image are sample types necessary for model training, and if the model only selects the positive sample image or only selects the negative sample image, the sample diversity degree of the model is low, and the effect of the model in various complex scenes is influenced. Therefore, in the training process of the model, as many sample types of the model as possible need to be selected.
In some embodiments, the step S102 of extracting features of the door and window sample image and inputting the extracted door and window feature result into the initial neural network model for training includes, as shown in fig. 3:
step S301, initializing a neural network model, wherein the network model comprises a target detection model and a semantic segmentation model.
The target detection model is used for detecting the door and window positions in the door and window sample image; and the semantic segmentation model is used for detecting the door and window types in the door and window sample image. In the specific implementation process, the target detection model is a yolo v3 neural network model; the semantic segmentation model is a mask-rcnn neural network model. The YoLO (You Only Look at Once) neural network model can combine target region prediction and target category prediction into one, and a target detection task is regarded as a regression problem of the target region prediction and the category prediction. The method adopts a single neural network to directly predict the object boundary and the class probability, and realizes end-to-end object detection, so that the identification performance is higher. The yolo algorithm of version v3 is used in the object detection model of the present embodiment.
The mask-rcnn neural network model is based on fast R-CNN, one branch is added to the original two branches (classification + coordinate regression) for semantic segmentation, and the loss change brought by adding the mask branch indirectly influences the effect of a main network. In the present embodiment, the type of the door and window can be identified.
Step S302, respectively inputting the door and window sample images into a target detection model and a semantic segmentation model for training, and outputting the identification results of the door and window sample images.
In the training process of the target detection model and the semantic segmentation model, the recognition result of the door and window sample image is output, and the recognition result is used as a parameter of model training and needs to be trained on the corresponding model again.
And step S303, judging the identification result of the door and window sample image and the door and window real result of the door and window sample image to obtain a judgment result.
The identification result of the door and window sample image is a result output by the model and is not a real result, the output result needs to be judged according to the real result of the door and window sample image, and the process can be realized through manual judgment.
And step S304, adjusting the preset training parameters of the neural network model according to the judgment result.
If the identification result of the door and window sample image is inconsistent with the door and window real result of the door and window sample image in the judgment result, the model performance is low, and relevant parameters need to be adjusted to further train the model; if the recognition result of the door and window sample image is consistent with the door and window real result of the door and window sample image, the model performance is good, and the training degree of the model can be properly adjusted in the subsequent training process.
An embodiment of the present invention provides a door and window detection method, as shown in fig. 4, the method includes:
step S401, a two-dimensional plane house type graph to be detected is obtained.
The two-dimensional flat house type graph to be detected is used as an input image and is different from an image source in a model training process, the two-dimensional flat house type graph to be detected can be obtained by drawing in a drawing tool in the home field, and house type data can also be obtained by rendering after being input into a related rendering tool.
And S402, inputting the two-dimensional plane house type graph to be detected into a door and window detection model which is trained in advance, and outputting a door and window detection result.
The door and window detection model is obtained through training by the model training method for the door and window detection model mentioned in the embodiment. In some embodiments, this step is illustrated in fig. 5, and comprises:
step S501, detecting the two-dimensional plane house type graph to be detected through a target detection model in a door and window detection model to obtain the door and window position in the two-dimensional plane house type graph to be detected.
And S502, detecting the two-dimensional plane house type graph to be detected through a semantic segmentation model in a door and window detection model to obtain the door and window type in the two-dimensional plane house type graph to be detected.
Because the door and window detection model comprises the target detection model and the semantic segmentation model, the two-dimensional flat house type graph to be detected is respectively input into the target detection model and the semantic segmentation model in the steps S501 and S502, and the corresponding door and window position and the corresponding door and window type are obtained. The above steps can be performed in an interchangeable order or simultaneously.
And S503, marking the door and window position and the door and window type in the two-dimensional plane house type image to be detected, and outputting the identification result of the door and window sample image.
The door and window position result can be an array comprising a vertex coordinate and a line segment length, the coordinate and the line segment form a closed area, and corresponding doors and windows in the two-dimensional plane house type graph can be calibrated.
The door and window type can be drawn near the door and window position, and can also be stored in the related attribute value of the image as a parameter, and the attribute value is in one-to-one correspondence with the door and window position data.
In the embodiment of the door/window detection method, the implementation principle and the generated technical effect of the door/window detection model are the same as those of the embodiment of the model training method for door/window detection, and for brief description, corresponding contents in the embodiment of the method can be referred to where the embodiment is not mentioned.
Corresponding to the embodiment of the model training method for door and window detection, the embodiment further provides a model training device for door and window detection, as shown in fig. 6, the device includes:
the sample acquisition module 601 is configured to determine a door and window sample image based on a two-dimensional planar house type map in a preset door and window training set;
the model training module 602 is configured to perform feature extraction on the door and window sample image, and input an extracted door and window feature result into the initial neural network model for training;
the model obtaining module 603 obtains a model for door and window detection when a training result of the preset neural network model meets a preset expected threshold.
The model training device for door and window detection provided by the embodiment of the invention has the same implementation principle and technical effect as the embodiment of the model training method for door and window detection, and for brief description, corresponding contents in the embodiment of the method can be referred to where the embodiment is not mentioned.
Corresponding to the embodiment of the door and window detecting method, the embodiment further provides a door and window detecting device, as shown in fig. 7, the device includes:
a two-dimensional house type diagram acquisition module 701 for acquiring a two-dimensional house type diagram to be detected;
the door and window detection module 702 is used for inputting the two-dimensional plane house type graph to be detected into a door and window detection model which is trained in advance, and outputting a door and window detection result; the door and window detection model is obtained by training through the model training method for door and window detection in any one of the first aspect.
The door and window detecting device provided by the embodiment of the invention has the same implementation principle and technical effect as the embodiment of the door and window detecting method, and for brief description, the corresponding content in the embodiment of the method can be referred to where the embodiment is not mentioned.
The embodiment also provides an electronic device, a schematic structural diagram of which is shown in fig. 8, and the electronic device includes a processor 101 and a memory 102; the memory 102 is configured to store one or more computer instructions, and the one or more computer instructions are executed by the processor to implement the model training method for door and window detection and the door and window detection method.
The electronic device shown in fig. 8 further comprises a bus 103 and a communication interface 104, and the processor 101, the communication interface 104 and the memory 102 are connected through the bus 103.
The Memory 102 may include a high-speed Random Access Memory (RAM) and may also include a non-volatile Memory (non-volatile Memory), such as at least one disk Memory. Bus 103 may be an ISA bus, PCI bus, EISA bus, or the like. The bus may be divided into an address bus, a data bus, a control bus, etc. For ease of illustration, only one double-headed arrow is shown in FIG. 8, but that does not indicate only one bus or one type of bus.
The communication interface 104 is configured to connect with at least one user terminal and other network units through a network interface, and send the packaged IPv4 message or IPv4 message to the user terminal through the network interface.
The processor 101 may be an integrated circuit chip having signal processing capabilities. In implementation, the steps of the above method may be performed by integrated logic circuits of hardware or instructions in the form of software in the processor 101. The Processor 101 may be a general-purpose Processor, and includes a Central Processing Unit (CPU), a Network Processor (NP), and the like; the device can also be a Digital Signal Processor (DSP), an Application Specific Integrated Circuit (ASIC), a Field Programmable Gate Array (FPGA) or other Programmable logic device, a discrete Gate or transistor logic device, or a discrete hardware component. The various methods, steps, and logic blocks disclosed in the embodiments of the present disclosure may be implemented or performed. A general purpose processor may be a microprocessor or the processor may be any conventional processor or the like. The steps of the method disclosed in connection with the embodiments of the present disclosure may be directly implemented by a hardware decoding processor, or implemented by a combination of hardware and software modules in the decoding processor. The software module may be located in ram, flash memory, rom, prom, or eprom, registers, etc. storage media as is well known in the art. The storage medium is located in the memory 102, and the processor 101 reads the information in the memory 102 and completes the steps of the method of the foregoing embodiment in combination with the hardware thereof.
Embodiments of the present invention further provide a computer-readable storage medium, on which a computer program is stored, where the computer program, when executed by a processor, performs the steps of the method of the foregoing embodiments.
In the several embodiments provided in the present application, it should be understood that the disclosed system, apparatus, and method may be implemented in other ways. The above-described embodiments of the apparatus are merely illustrative, and for example, a division of a unit is merely a division of one logic function, and there may be other divisions when actually implemented, and for example, a plurality of units or components may be combined or integrated into another system, or some features may be omitted, or not executed. In addition, the shown or discussed mutual coupling or direct coupling or communication connection may be an indirect coupling or communication connection of devices or units through some communication interfaces, and may be in an electrical, mechanical or other form.
Units described as separate parts may or may not be physically separate, and parts displayed as units may or may not be physical units, may be located in one place, or may be distributed on a plurality of network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of the embodiment.
In addition, functional units in the embodiments of the present invention may be integrated into one processing unit, or each unit may exist alone physically, or two or more units are integrated into one unit.
The functions, if implemented in the form of software functional units and sold or used as a stand-alone product, may be stored in a non-volatile computer-readable storage medium executable by a processor. Based on such understanding, the technical solution of the present invention may be embodied in the form of a software product, which is stored in a storage medium and includes instructions for causing a computer device (which may be a personal computer, a server, or a network device) to execute all or part of the steps of the method according to the embodiments of the present invention. And the aforementioned storage medium includes: a U-disk, a removable hard disk, a Read-Only Memory (ROM), a Random Access Memory (RAM), a magnetic disk or an optical disk, and other various media capable of storing program codes.
Finally, it should be noted that: although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art will understand that: any person skilled in the art can modify or easily conceive the technical solutions described in the foregoing embodiments or equivalent substitutes for some technical features within the technical scope of the present disclosure; such modifications, changes or substitutions do not depart from the spirit and scope of the embodiments of the present invention, and they should be construed as being included therein. Therefore, the protection scope of the present invention shall be subject to the protection scope of the claims.
Claims (10)
1. A model training method for door and window detection is characterized by comprising the following steps:
determining a door and window sample image based on a two-dimensional plane house-type graph in a preset door and window training set;
extracting the characteristics of the door and window sample image, and inputting the extracted door and window characteristic result into an initial neural network model for training;
and when the training result of the initial neural network model meets a preset expected threshold value, obtaining a model for door and window detection.
2. The model training method for door and window detection according to claim 1, wherein the step of determining the door and window sample image based on the two-dimensional planar floor plan in the preset door and window training set comprises:
traversing the two-dimensional plane house type graph in the door and window training set, and identifying doors and windows in the two-dimensional plane house type graph to obtain a door and window identification result;
determining the door and window sample image containing doors and windows in the door and window identification result as a door and window positive sample image; determining the door and window sample image which does not contain doors and windows in the door and window identification result as a door and window negative sample image;
and taking the door and window positive sample image and the door and window negative sample image as door and window sample images.
3. The model training method for door and window detection according to claim 1, wherein the step of performing feature extraction on the door and window sample image, inputting the extracted door and window feature result into an initial neural network model for training comprises:
initializing the neural network model, wherein the network model comprises a target detection model and a semantic segmentation model;
respectively inputting the door and window sample images into the target detection model and the semantic segmentation model for training, and outputting the identification results of the door and window sample images;
judging the identification result of the door and window sample image and the door and window real result of the door and window sample image to obtain a judgment result;
and adjusting the training parameters of the preset neural network model according to the judgment result.
4. The model training method for door and window detection according to claim 3, wherein the target detection model is a yolo v3 neural network model; the semantic segmentation model is a mask-rcnn neural network model.
5. A method of detecting a door or window, the method comprising:
acquiring a two-dimensional planar house pattern to be detected;
inputting the two-dimensional plane house type graph to be detected into a door and window detection model which is trained in advance, and outputting a door and window detection result; the door and window detection model is obtained by training through the model training method for door and window detection in any one of claims 1 to 4.
6. The door and window detection method according to claim 5, wherein the step of inputting the two-dimensional flat house type diagram to be detected into a door and window detection model which is trained in advance and outputting a door and window detection result comprises:
the two-dimensional plane house type graph to be detected is detected through a target detection model in the door and window detection model, and the door and window position in the two-dimensional plane house type graph to be detected is obtained;
the two-dimensional planar house type graph to be detected is detected through a semantic segmentation model in the door and window detection model, and the door and window type in the two-dimensional planar house type graph to be detected is obtained;
and marking the door and window position and the door and window type in the two-dimensional plane house type graph to be detected, and outputting the identification result of the door and window sample image.
7. A model training device for door and window detection, the device comprising:
the system comprises a sample acquisition module, a data acquisition module and a data processing module, wherein the sample acquisition module is used for determining a door and window sample image based on a two-dimensional plane house-type diagram in a preset door and window training set;
the model training module is used for extracting the characteristics of the door and window sample images and inputting the extracted door and window characteristic results into an initial neural network model for training;
and the model acquisition module is used for obtaining a model for door and window detection when the training result of the preset neural network model meets a preset expected threshold value.
8. A door and window detecting device, the device comprising:
the acquisition module of the two-dimensional house type graph to be detected is used for acquiring the two-dimensional house type graph to be detected;
the door and window detection module is used for inputting the two-dimensional plane house type graph to be detected into a door and window detection model which is trained in advance and outputting a door and window detection result; the door and window detection model is obtained by training through the model training method for door and window detection in any one of claims 1 to 4.
9. An electronic device, comprising: a processor and a storage device; the storage means having stored thereon a computer program which, when executed by the processor, performs the steps of the method according to any one of claims 1 to 5.
10. A computer-readable storage medium, on which a computer program is stored, which, when being executed by a processor, carries out the steps of the method according to any one of the claims 1 to 5.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010100580.7A CN111985518A (en) | 2020-02-18 | 2020-02-18 | Door and window detection method and model training method and device thereof |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010100580.7A CN111985518A (en) | 2020-02-18 | 2020-02-18 | Door and window detection method and model training method and device thereof |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN111985518A true CN111985518A (en) | 2020-11-24 |
Family
ID=73441750
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010100580.7A Pending CN111985518A (en) | 2020-02-18 | 2020-02-18 | Door and window detection method and model training method and device thereof |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111985518A (en) |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113920314A (en) * | 2021-09-30 | 2022-01-11 | 北京百度网讯科技有限公司 | Semantic segmentation and model training method, device, equipment and storage medium |
| CN113946900A (en) * | 2021-12-21 | 2022-01-18 | 深圳小库科技有限公司 | Method for quickly recommending similar house types based on house type profiles and distribution characteristics |
| CN114219590A (en) * | 2022-02-23 | 2022-03-22 | 广亚铝业有限公司 | Door and window information service system and method based on server and client APP |
| CN114677604A (en) * | 2022-04-20 | 2022-06-28 | 电子科技大学 | Window state detection method based on machine vision |
| CN116624065A (en) * | 2023-07-20 | 2023-08-22 | 山东智赢门窗科技有限公司 | Automatic folding regulation and control method for intelligent doors and windows |
| CN116876950A (en) * | 2023-09-05 | 2023-10-13 | 山东智赢门窗科技有限公司 | Intelligent door and window control system and method, computer equipment and storage medium |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2017007626A1 (en) * | 2015-07-09 | 2017-01-12 | Qualcomm Incorporated | Context-based priors for object detection in images |
| CN109993797A (en) * | 2019-04-04 | 2019-07-09 | 广东三维家信息科技有限公司 | Door and window method for detecting position and device |
| CN110059690A (en) * | 2019-03-28 | 2019-07-26 | 广州智方信息科技有限公司 | Floor plan semanteme automatic analysis method and system based on depth convolutional neural networks |
| CN110084293A (en) * | 2019-04-18 | 2019-08-02 | 贝壳技术有限公司 | A kind of determination method and apparatus in complete bright pattern house |
| CN110348368A (en) * | 2019-07-08 | 2019-10-18 | 创新奇智(北京)科技有限公司 | A kind of artificial intelligence analytic method, computer-readable medium and the system of floor plan |
| EP3579148A1 (en) * | 2018-06-08 | 2019-12-11 | Guangdong Oppo Mobile Telecommunications Corp., Ltd. | Method and apparatus for training, classification model, mobile terminal, and readable storage medium |
-
2020
- 2020-02-18 CN CN202010100580.7A patent/CN111985518A/en active Pending
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2017007626A1 (en) * | 2015-07-09 | 2017-01-12 | Qualcomm Incorporated | Context-based priors for object detection in images |
| EP3579148A1 (en) * | 2018-06-08 | 2019-12-11 | Guangdong Oppo Mobile Telecommunications Corp., Ltd. | Method and apparatus for training, classification model, mobile terminal, and readable storage medium |
| CN110059690A (en) * | 2019-03-28 | 2019-07-26 | 广州智方信息科技有限公司 | Floor plan semanteme automatic analysis method and system based on depth convolutional neural networks |
| CN109993797A (en) * | 2019-04-04 | 2019-07-09 | 广东三维家信息科技有限公司 | Door and window method for detecting position and device |
| CN110084293A (en) * | 2019-04-18 | 2019-08-02 | 贝壳技术有限公司 | A kind of determination method and apparatus in complete bright pattern house |
| CN110348368A (en) * | 2019-07-08 | 2019-10-18 | 创新奇智(北京)科技有限公司 | A kind of artificial intelligence analytic method, computer-readable medium and the system of floor plan |
Cited By (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113920314A (en) * | 2021-09-30 | 2022-01-11 | 北京百度网讯科技有限公司 | Semantic segmentation and model training method, device, equipment and storage medium |
| CN113920314B (en) * | 2021-09-30 | 2022-09-02 | 北京百度网讯科技有限公司 | Semantic segmentation and model training method, device, equipment and storage medium |
| CN113946900A (en) * | 2021-12-21 | 2022-01-18 | 深圳小库科技有限公司 | Method for quickly recommending similar house types based on house type profiles and distribution characteristics |
| CN113946900B (en) * | 2021-12-21 | 2022-03-29 | 深圳小库科技有限公司 | Method for quickly recommending similar house types based on house type profiles and distribution characteristics |
| CN114219590A (en) * | 2022-02-23 | 2022-03-22 | 广亚铝业有限公司 | Door and window information service system and method based on server and client APP |
| CN114677604A (en) * | 2022-04-20 | 2022-06-28 | 电子科技大学 | Window state detection method based on machine vision |
| CN114677604B (en) * | 2022-04-20 | 2023-04-07 | 电子科技大学 | Window state detection method based on machine vision |
| CN116624065A (en) * | 2023-07-20 | 2023-08-22 | 山东智赢门窗科技有限公司 | Automatic folding regulation and control method for intelligent doors and windows |
| CN116624065B (en) * | 2023-07-20 | 2023-10-13 | 山东智赢门窗科技有限公司 | Automatic folding regulation and control method for intelligent doors and windows |
| CN116876950A (en) * | 2023-09-05 | 2023-10-13 | 山东智赢门窗科技有限公司 | Intelligent door and window control system and method, computer equipment and storage medium |
| CN116876950B (en) * | 2023-09-05 | 2023-12-05 | 山东智赢门窗科技有限公司 | Intelligent door and window control system and method, computer equipment and storage medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111985518A (en) | Door and window detection method and model training method and device thereof | |
| CN112052787B (en) | Target detection method and device based on artificial intelligence and electronic equipment | |
| CN109784391B (en) | Multi-model-based sample labeling method and device | |
| CN110110715A (en) | Text detection model training method, text filed, content determine method and apparatus | |
| CN112434721A (en) | Image classification method, system, storage medium and terminal based on small sample learning | |
| CN111062495A (en) | Machine learning method and related device | |
| CN111340954B (en) | House type wall drawing method and model training method and device thereof | |
| CN108460346B (en) | Fingerprint identification method and device | |
| CN110889437B (en) | Image processing method and device, electronic equipment and storage medium | |
| CN112465709B (en) | Image enhancement method, device, storage medium and equipment | |
| CN108520263B (en) | Panoramic image identification method and system and computer storage medium | |
| CN113657202A (en) | Component identification method, training set construction method, device, equipment and storage medium | |
| CN113554065A (en) | Three-dimensional building model component classification method and device | |
| CN114092947B (en) | Text detection method and device, electronic equipment and readable storage medium | |
| CN111738290B (en) | Image detection method, model construction and training method, device, equipment and medium | |
| CN113128604A (en) | Page element identification method and device, electronic equipment and storage medium | |
| CN107077617B (en) | Fingerprint extraction method and device | |
| CN116798053A (en) | Icon generation method and device | |
| CN113808192B (en) | House pattern generation method, device, equipment and storage medium | |
| CN115713669A (en) | Image classification method and device based on inter-class relation, storage medium and terminal | |
| CN116777814A (en) | Image processing method, apparatus, computer device, storage medium, and program product | |
| CN113642667A (en) | Enhancement strategy determination method and device, electronic equipment and storage medium | |
| CN110135464B (en) | Image processing method and device, electronic equipment and storage medium | |
| CN113886578A (en) | Form classification method and device | |
| CN110111298A (en) | Intelligent house type size verification method, apparatus, equipment and readable storage medium storing program for executing |
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 |