Floor plan generation server, floor plan generation method, and floor plan generation program

Abstract

A floor plan generation server, a floor plan generation method, and a floor plan generation program are provided that can generate floor plans in a predetermined format from drawings that are made public as objects of trading. [Solution] The real estate property information acquisition unit 131 of the floor plan generation server 100 acquires real estate property information that associates property information for each real estate property with drawings related to the real estate property, the floor plan determination unit 132 determines whether the drawing is a floor plan, the floor plan generation unit 134 generates a floor plan in a predetermined format from the drawing if it determines that the drawing is a floor plan, and the quality determination unit 135 determines whether the drawing, which is a floor plan in a predetermined format, is of a certain quality, and if it determines that the drawing is of a certain quality, the property information and the drawing, which is a floor plan in the predetermined format, are associated and stored in the real estate floor plan information storage unit 121.

Description

[Technical Field]

[0001] The present invention relates to a floor plan generation server, a floor plan generation method, and a floor plan generation program. [Background technology]

[0002] Floor plans of real estate properties are important in real estate transactions because they allow customers to visualize their life there. However, because they are created by real estate sales companies and real estate brokerage companies, they are often created in various formats. However, there has been a demand for floor plans to be provided in a consistent format to make it easier to compare real estate properties. To solve this problem, a technology has been disclosed that reads design drawings, acquires image data, and creates vector data of the floor plan from the image data (see Patent Document 1). [Prior art documents] [Patent documents]

[0003] [Patent Document 1] Japanese Patent Publication No. 2022-157539 Summary of the Invention [Problem to be solved by the invention]

[0004] However, since not all real estate properties on the real estate market are new constructions, it is rare to be able to obtain blueprints. In addition, the applicant conducted testing to generate floor plans in a predetermined format from various publicly available drawings using generative AI, which has seen remarkable development in recent years, but found it difficult to generate floor plans of sufficient quality.

[0005] The present invention has been made in consideration of the above, and aims to provide a floor plan generation server, a floor plan generation method, and a floor plan generation program that can generate floor plans in a predetermined format from drawings that are made public as objects of trading. [Means for solving the problem]

[0006] In order to solve the above-mentioned problems, the present invention is characterized by obtaining real estate property information that associates property information for each real estate property with a drawing related to the real estate property, determining whether the drawing is a floor plan, and if it is determined that the drawing is a floor plan, generating a floor plan in a predetermined format from the drawing, determining whether the floor plan in the predetermined format is of a certain quality, and if it is determined that the floor plan in the predetermined format is of a certain quality, associating the property information with the floor plan in the predetermined format and storing them in a real estate floor plan information storage unit. [Effects of the Invention]

[0007] According to the present invention configured as described above, it is possible to generate a floor plan in a predetermined format from a drawing that has been made public as a trading object. [Brief explanation of the drawings]

[0008] [Figure 1] 1 is an explanatory diagram showing an example of a floor plan generating system 10 according to an embodiment of the present invention. [Figure 2] 1 is a block diagram showing the configuration of a floor plan generating server 100 according to an embodiment of the present invention. [Figure 3] 1 is an explanatory diagram showing an example of the data configuration of a real estate floor plan information storage unit 121 and a standard floor plan storage unit 126. FIG. [Figure 4] 10 is a flowchart showing the procedure of a floor plan generation process executed by the floor plan generation server 100. [Figure 5] 10 is a flowchart showing the procedure of a learned model re-learning process executed by the floor plan generation server 100. [Figure 6] 10 is a flowchart showing the procedure of a floor plan generation process executed by the floor plan generation server 100. [Figure 7] 10 is a flowchart showing the procedure of a floor plan generation process executed by the floor plan generation server 100. DETAILED DESCRIPTION OF THE INVENTION

[0009] Below, examples of the floor plan generation server, floor plan generation method, and floor plan generation program according to the present application will be described with reference to the accompanying drawings. The following explanations are examples of embodiments of the present application, and the floor plan generation server, floor plan generation method, and floor plan generation program according to the present application are not limited to these examples.

[0010] 1 is an explanatory diagram showing an example of a floor plan generation system 10 according to this embodiment. The floor plan generation system 10 includes a floor plan generation server 100, web servers 200-1 to 200-n (hereinafter referred to as web servers 200), and information terminal devices 300-1 to 300-m (hereinafter referred to as information terminal devices 300).

[0011] The floor plan generation server 100 is a computer that generates floor plans in a predetermined format from drawings in various formats. The drawings in various formats are acquired by crawling the web server 200, receiving drawings sent from other devices, or reading drawings stored in a storage medium. The floor plan generation server 100 transmits and receives data to and from the web server 200 and the information terminal device 300 via a network N. The network N is any communication network, such as the Internet, an intranet, a local area network (LAN), a virtual private network (VPN), or a mobile communication network, or a combination thereof, and some or all of these may be wired or wireless.

[0012] The web server 200 is a computer that publishes information related to real estate by real estate information providers, real estate sales and rental brokerage companies, etc. The floor plan generation server 100 periodically crawls and acquires real estate property information to be published on the web server 200.

[0013] The information terminal device 300 is a computer operated by a user who wishes to purchase or rent real estate, and may be a smartphone, tablet terminal, personal computer, etc. The information terminal device 300 launches a browser, accesses the floor plan generation server 100, and searches for and displays real estate information.

[0014] 2 is a block diagram showing the configuration of the floor plan generation server 100 according to this embodiment. The floor plan generation server 100 includes a communication unit 110, a storage unit 120, a control unit 130, and an operation display unit 140.

[0015] The communication unit 110 transmits and receives data between the web server 200 and the information terminal device 300. The communication unit 110 periodically crawls to receive real estate property information and the like published by the web server 200. The communication unit 110 receives real estate property information and the like transmitted from the web server 200 to be stored in the memory unit of the web server 200.

[0016] The memory unit 120 includes a real estate floor plan information memory unit 121, an NG drawing (floor plan) memory unit 122, an NG drawing (unnecessary elements) memory unit 123, an NG drawing (quality) memory unit 124, a trained model 125, and a standard floor plan memory unit 126.

[0017] Fig. 3 is an explanatory diagram showing an example of the data configuration of the real estate floor plan information storage unit 121 and the standard floor plan storage unit 126. Fig. 3(a) is an explanatory diagram showing an example of the data configuration of the real estate floor plan information storage unit 121. The real estate floor plan information storage unit 121 stores property information, floor plans, quality NG drawings, and other information in association with each other.

[0018] Property information is information about real estate properties. Property information is information about individual properties, such as the property name, address, floor number (if it is an apartment or condominium), and room number, and is information that uniquely identifies the real estate property. Floor plans are image data of floor plans in a predetermined format generated from drawings acquired as part of the real estate property information. Quality-NG drawings are drawings stored for use by the feature point matching unit 138, described below, when it is determined that the floor plan in a predetermined format generated from a drawing is not of a certain quality. Other information includes information about the real estate property other than information that uniquely identifies the property, such as the selling price, rent, and map.

[0019] 31 in Figure 3 is data related to a real estate property when a floor plan in a prescribed format has been generated and the floor plan is of a certain quality, and the property information and floor plan are stored. 32 in Figure 3 is data related to a real estate property when a floor plan in a prescribed format has been generated but the floor plan is not of a certain quality, and the property information and a quality NG drawing are stored. 33 in Figure 3 is data related to a real estate property when a floor plan in a prescribed format could not be generated, and the property information is stored.

[0020] The NG drawing (floor plan) memory unit 122 stores a drawing that is determined not to be a floor plan when it is determined that the drawing acquired by the real estate property information acquisition unit 131 is not a floor plan. The NG drawing (unnecessary element) memory unit 123 stores a drawing that is determined to include unnecessary elements when the unnecessary element determination unit 133 determines that the drawing contains unnecessary elements. The NG drawing (quality) memory unit 124 stores a drawing that is determined not to be of a certain quality when the quality determination unit 135 determines that the quality is not certain. The drawings stored in the NG drawing (floor plan) memory unit 122, the NG drawing (unnecessary element) memory unit 123, and the NG drawing (quality) memory unit 124 (hereinafter referred to as NG drawings) are used when re-training the trained model 125.

[0021] The trained model 125 is a trained model that has been trained to determine whether image data of a drawing meets predetermined requirements and quality, and any model can be used. For example, deep learning using a neural network, a support vector machine, a decision tree, a naive Bayes method, a k-nearest neighbor method, or a combination of these can be used. The trained model 125 may include information such as parameters of the trained model. The trained model 125 includes trained models used by the floor plan determination unit 132, the unnecessary element determination unit 133, and the quality determination unit 135.

[0022] Trained model 125(a) is a trained model that has been trained to receive input image data of a drawing and output an OK or NG flag as an output result indicating whether the input image data is a floor plan. Trained model 125(b) is a trained model that has been trained to receive input image data of a drawing and output an OK or NG flag as an output result indicating whether the input image data contains unnecessary elements. Trained model 125(c) is a trained model that has been trained to receive input image data of a drawing and output an OK or NG flag as an output result indicating whether the input image data is of a certain quality or above as a floor plan. Note that trained model 125(c) may be a trained model in which explanatory variables include the vertical length of the image, the horizontal length of the image, the area of ​​the image, the variance of the lines, etc., and the objective variable is an OK or NG flag for quality.

[0023] FIG. 3(b) is an explanatory diagram showing an example of the data configuration of the standard floor plan storage unit 126. The standard floor plan storage unit 126 stores information related to the standard floor plan. The standard floor plan storage unit 126 stores feature point information of the standard floor plan, the standard floor plan, and other information in association with each other. The standard floor plan is a floor plan of a certain quality, and the feature point information of the standard floor plan is the coordinates of characteristic points in the standard floor plan.

[0024] Next, we will explain the control unit 130. The control unit 130 includes a real estate property information acquisition unit 131, a floor plan determination unit 132, an unnecessary element determination unit 133, a floor plan generation unit 134, a quality determination unit 135, a learning unit 136, a floor plan update unit 137, and a feature point matching unit 138.

[0025] The real estate property information acquisition unit 131 periodically causes the communication unit 110 to execute a crawling process to acquire real estate information to be made public on the web server 200. The real estate property information acquisition unit 131 also acquires real estate property information provided from other systems or other databases. The real estate property information includes at least property information and image data of drawings.

[0026] Floor plan determination unit 132 determines whether image data of a drawing acquired by real estate property information acquisition unit 131 is a floor plan. More specifically, floor plan determination unit 132 inputs the image data of the drawing into trained model 125(a) and determines whether it is a floor plan based on the output result output from trained model 125(a). Floor plan determination unit 132 stores the image data of a drawing determined not to be a floor plan in NG drawing (floor plan) storage unit 122, and stores property information in real estate floor plan information storage unit 121.

[0027] The unnecessary element determination unit 133 determines whether unnecessary elements are included in the image data of the drawing acquired by the real estate property information acquisition unit 131. More specifically, the unnecessary element determination unit 133 inputs the image data of the drawing into the trained model 125(b) and determines whether unnecessary elements are included based on the output result output from the trained model 125(b). The unnecessary element determination unit 133 stores the image data of the drawing determined to include unnecessary elements in the NG drawing (unnecessary element) storage unit 123, and stores the property information in the real estate floor plan information storage unit 121. Unnecessary elements include, for example, the name or logo of a real estate sales or rental brokerage company, plants, furniture, etc., and the trained model 125(b) may be generated by limiting or adding unnecessary elements depending on the drawing to be collected.

[0028] The floor plan generation unit 134 generates a floor plan in a predetermined format from the image data of the drawing that has been determined to be acceptable by the floor plan determination unit 132 and the unnecessary element determination unit 133. As an example, the floor plan generation unit 134 converts the image data of the drawing into monochrome and generates a floor plan by converting the outline of black dots into shapes such as straight lines and arcs, characters, etc.

[0029] The quality determination unit 135 determines whether the floor plan drawing in a predetermined format generated by the floor plan generation unit 134 is of a certain quality. More specifically, the quality determination unit 135 inputs image data of the drawing generated by the floor plan generation unit 134 into the trained model 125(c) and makes a determination based on the output result indicating whether the floor plan is of a certain quality output from the trained model 125(c). If the quality determination unit 135 determines that the floor plan is of a certain quality, it associates the property information with the image data of the floor plan determined to be of a certain quality and stores them in the real estate floor plan information storage unit 121. If the quality determination unit 135 determines that the floor plan is not of a certain quality, it stores the property information in the real estate floor plan information storage unit 121, and stores the image data of the drawing determined to be of a certain quality in the NG drawing (quality) storage unit 124 and the real estate floor plan information storage unit 121.

[0030] The learning unit 136 generates a trained model 125(a) that has been trained to output an output result indicating whether or not a drawing is a floor plan using drawings that are floor plans and drawings that are not floor plans as training data. The learning unit 136 generates a trained model 125(b) that has been trained to output an output result indicating whether or not a drawing contains unnecessary elements using drawings that include unnecessary elements and drawings that do not include unnecessary elements as training data. The learning unit 136 generates a trained model 125(c) that has been trained to output an output result indicating whether or not a drawing is of a certain quality using drawings that are of a certain quality and drawings that are not of a certain quality as training data.

[0031] The learning unit 136 re-learns the trained model 125(a) using it for drawings that have been determined to be NG drawings by the floor plan determination unit 132 and are stored in the NG drawing (floor plan) memory unit 122. The learning unit 136 re-learns the trained model 125(b) using it for drawings that have been determined to be NG drawings by the unnecessary element determination unit 133 and are stored in the NG drawing (unnecessary element) memory unit 123. The learning unit 136 re-learns the trained model 125(c) using it for drawings that have been determined to be NG drawings by the quality determination unit 135 and are stored in the NG drawing (quality) memory unit 124.

[0032] The drawing acquisition unit 137 acquires a drawing corresponding to property information for which no floor plan is stored in the real estate floor plan information storage unit 121. If a floor plan of a certain quality and in a predetermined format can be generated from the drawing acquired by the drawing acquisition unit 137, the generated floor plan is associated with the property information and stored in the real estate floor plan information storage unit 121. The process of generating a floor plan from the drawing acquired by the drawing acquisition unit 137 is performed by executing the processes of the floor plan determination unit 132, the unnecessary element determination unit 133, the floor plan generation unit 134, and the quality determination unit 135 described above.

[0033] If the quality determination unit 135 determines that the drawing is not of a certain quality, the feature point matching unit 138 matches the feature points of the standard floor plan with the feature points of the drawing, and if the feature points of the standard floor plan match the feature points of the drawing, stores the standard floor plan in the real estate floor plan information storage unit 121 in association with the property information.

[0034] The operation display unit 140 accepts operations, etc. by the user and displays the results of the operations, etc. More specifically, the operation display unit 140 displays NG drawings on the display screen and accepts input of the judgment results for the displayed NG drawings.

[0035] The floor plan generation process executed by the floor plan generation system 10 configured as described above will be described below. FIG. 4 is a flowchart showing the floor plan generation process steps executed by the floor plan generation server 100.

[0036] The real estate property information acquisition unit 131 of the floor plan generation server 100 acquires real estate property information (step S401). More specifically, the real estate property information acquisition unit 131 acquires real estate property information by periodically crawling the web server 200 via the communication unit 110. Alternatively, the real estate property information acquisition unit 131 receives and acquires real estate information transmitted from the web server 200 periodically.

[0037] The floor plan determination unit 110 determines whether the drawing included in the real estate property information is a floor plan (step S402). More specifically, the floor plan determination unit 132 inputs image data of the drawing and determines whether the drawing is a floor plan using a trained model 124(a) that has been trained to output an OK or NG flag, which is an output result indicating whether the drawing is a floor plan. If the floor plan determination unit 132 determines that the drawing is not a floor plan, it stores the drawing in the NG drawing memory unit (floor plan) 122, stores the property information in the real estate floor plan information memory unit 121, and ends the process.

[0038] If it is determined that the drawing is a floor plan, the unnecessary element determination unit 133 determines whether the drawing included in the real estate property information includes unnecessary elements (step S403). More specifically, the unnecessary element determination unit 133 inputs image data of the drawing and determines whether the drawing includes unnecessary elements using a trained model 124(b) that has been trained to output an OK or NG flag as an output result indicating whether unnecessary elements are included. If the unnecessary element determination unit 133 determines that the drawing includes unnecessary elements, it stores the drawing in the NG drawing memory unit (unnecessary elements) 123, stores the property information in the real estate floor plan information memory unit 121, and ends the process.

[0039] If it is determined that the drawing does not contain unnecessary elements, the floor plan generation unit 134 generates a floor plan in a predetermined format from the drawing (step S404). The quality determination unit 135 determines whether the floor plan drawing generated by the floor plan generation unit 134 is of a certain quality (step S405). More specifically, the quality determination unit 135 inputs image data of the drawing generated by the floor plan generation unit 134 and determines whether the quality is certain using a trained model 124(c) that has been trained to output an OK or NG flag as an output result indicating whether the quality is certain. If the quality determination unit 135 determines that the drawing is not of a certain quality, it stores the quality NG drawing in the NG drawing memory unit (quality) 124, stores the property information and the quality NG drawing in the real estate floor plan information memory unit 121, and ends the process.

[0040] If it is determined that the drawing is of a certain quality, the quality determining unit 135 associates the property information with the drawing, which is the floor plan generated by the floor plan generating unit 134, and stores them in the real estate floor plan information storage unit 121 (step S406).

[0041] In this way, even if drawings have been created in various formats by different plan creators, floor plans in a specified format can be easily generated by making multiple judgments and selecting the appropriate drawing.Even if it is difficult to generate a floor plan in a specified format from drawings in various formats using generative AI, by dividing the selection of the original drawing and the generation of the floor plan into multiple processes, it is possible to generate floor plans in a specified format that are of consistent quality and easy to compare, with a unified format.

[0042] Furthermore, by periodically executing the above-described process, it is possible to generate floor plans in a predetermined format from drawings included in newly published real estate property information that has become the subject of sale or rental. Furthermore, by determining whether there is a difference between the floor plan generated from the newly acquired drawings and the floor plan stored in the real estate floor plan information storage unit 121, it is possible to obtain the floor plan after renovation even if the floor plan has been changed due to renovation or the like.

[0043] In another embodiment, the quality determination unit 135 may determine whether or not a drawing that the unnecessary element determination unit 133 has determined not to contain unnecessary elements is of a certain quality before the floor plan generation unit 134 generates a floor plan in a predetermined format from the drawing, and if it is determined that the drawing is of a certain quality, the floor plan generation unit 134 may generate a floor plan in the predetermined format from the drawing.

[0044] Next, we will explain the trained model re-learning process executed by the floor plan generation system 10. Figure 5 is a flowchart showing the trained model re-learning process procedure executed by the floor plan generation server 100. We will explain the case where trained model 125(a), which determines whether or not a floor plan is present, is re-learned using a drawing stored in the NG drawing (floor plan) storage unit 122.

[0045] The learning unit 136 of the floor plan generation server 100 acquires an NG drawing from the NG drawing (floor plan) storage unit 122 (step S501). The operation display unit 140 displays the NG drawing acquired from the NG drawing (floor plan) storage unit 122 on the display screen (step S502). The operation display unit 140 accepts input of a determination result as to whether the NG drawing is a floor plan or not (step S503). Note that the processing of the operation display unit 140 may be executed not by the floor plan generation server 100 but by an information terminal device connected via the network N. The learning unit 136 inputs the drawing and the determination result into the trained model 125(a) and causes it to re-train (step S504).

[0046] In this way, for drawings that have been determined to be NG by floor plan determination unit 132, it is determined again whether the drawing is a floor plan or not, and trained model 125(a) is retrained, thereby improving the judgment accuracy of trained model 125(a). Note that in addition to drawings that have been determined to be NG by floor plan determination unit 132, it may also be determined again whether the drawing is a floor plan or not for drawings that have been determined to be OK by floor plan determination unit 132, and trained model 125(a) may be retrained.

[0047] For trained model 125(b), similar to the process described above, it is determined again whether or not the NG drawings stored in NG drawing (unnecessary element) memory unit 123 contain unnecessary elements, and trained model 125(b) is retrained. For trained model 125(c), it is determined again whether or not the NG drawings stored in NG drawing (quality) memory unit 124 are of a certain quality, and trained model 125(c) is retrained.

[0048] In the above-described embodiment, trained model 125(a), trained model 125(b), and trained model 125(b) are described as separate trained models, but, for example, two trained models may be generated as one trained model, or three trained models may be generated as one trained model. Furthermore, each trained model may be further generated as multiple trained models.

[0049] Next, a description will be given of the floor plan generation process for a property for which any of the determinations in the above-described floor plan generation process is NG and no floor plan is stored in the real estate floor plan information storage unit 121. Fig. 6 is a flowchart showing the floor plan generation process procedure executed by the floor plan generation server 100.

[0050] The floor plan acquisition unit 137 acquires property information for which no floor plan is stored in the real estate floor plan information storage unit 121 (step S601). The floor plan acquisition unit 137 acquires a floor plan corresponding to the property information (step S602). More specifically, the floor plan acquisition unit 137 acquires the floor plan from the web server 200 using the property information as a search condition.

[0051] The floor plan determination unit 132 determines whether the drawing is a floor plan (step S603). More specifically, the floor plan determination unit 132 inputs image data of the drawing and determines whether the drawing is a floor plan using the trained model 125(a) that has been trained to output an OK or NG flag, which is an output result indicating whether the drawing is a floor plan. If the floor plan determination unit 132 determines that the drawing is not a floor plan, it stores the drawing in the NG drawing storage unit (floor plan) 122 and ends the process.

[0052] If it is determined that the drawing is a floor plan, the unnecessary element determination unit 133 determines whether the drawing includes unnecessary elements (step S604). More specifically, the unnecessary element determination unit 133 inputs image data of the drawing and determines whether the drawing includes unnecessary elements using a trained model 125(b) that has been trained to output an OK or NG flag as an output result indicating whether the drawing includes unnecessary elements. If the unnecessary element determination unit 133 determines that the drawing includes unnecessary elements, it stores the drawing in the NG drawing storage unit (unnecessary elements) 123 and ends the process.

[0053] If it is determined that the drawing does not contain unnecessary elements, the floor plan generation unit 134 generates a floor plan in a predetermined format from the drawing (step S605). The quality determination unit 135 determines whether the floor plan generated by the floor plan generation unit 134 is of a certain quality (step S606). More specifically, the quality determination unit 135 inputs image data of the floor plan drawing generated by the floor plan generation unit 134 and determines whether the quality is certain using a trained model 125(c) that has been trained to output an OK or NG flag as an output result indicating whether the quality is certain. If the quality determination unit 135 determines that the standard floor plan is not of a certain quality, it stores the drawing in the NG drawing memory unit (quality) 124 and stores the drawing in the real estate floor plan information memory unit 121 in association with the property information, and ends the process.

[0054] If it is determined that the floor plan is of a certain quality, the quality determining unit 135 stores the floor plan in the real estate floor plan information storage unit 121 in association with the property information (step S607).

[0055] In this way, even if a floor plan cannot be generated from a drawing included in the real estate property information initially acquired from the web server 200, etc., the property information included in the real estate property information is stored in the real estate floor plan information storage unit 121, and by searching for the drawing using the property information as needed, it is possible to acquire drawings published on other websites or drawings updated after a certain period of time on the website from which the initial drawing was acquired. Furthermore, if a floor plan of a predetermined format can be generated with a certain level of quality from the drawing acquired by the drawing acquisition unit 137, the newly generated floor plan is stored in the real estate floor plan information storage unit 121. In this way, by subsequently generating a floor plan of a predetermined format with a certain level of quality and storing the floor plan corresponding to the property information in the real estate floor plan information storage unit 121, it is possible to obtain floor plans of a predetermined format with a certain level of quality over time.

[0056] Next, a floor plan generation process for property information in which the floor plan has been rejected due to quality will be described as an example different from the floor plan generation process in Fig. 6. Fig. 7 is a flowchart showing the floor plan generation process procedure executed by the floor plan generation server 100.

[0057] The feature point matching unit 138 acquires quality-NG drawings stored in the real estate floor plan information storage unit 121 (step S701). The feature point matching unit 138 calculates feature points from the quality-NG drawings (step S702). The feature points here refer to characteristic coordinates such as the positions of lines indicating the exterior walls of the floor plan or the walls separating rooms, and all or part of the coordinates of the intersections of lines included in the drawing may be used as feature points.

[0058] The feature point matching unit 138 matches the feature points of the NG quality drawing with the feature points of the standard floor plan (step S703). More specifically, the feature point matching unit 138 acquires the feature points of the standard floor plan from the standard floor plan storage unit 126. The feature point matching unit 138 matches the feature points of the NG quality drawing with the feature points of the standard floor plan, and determines whether the feature points of the NG quality drawing and the standard floor plan match. The feature point matching unit 138 acquires from the standard floor plan storage unit 126 a standard floor plan corresponding to the feature points of the standard floor plan that match the feature points of the NG quality drawing. If there is no standard floor plan that matches the coordinates of the feature points of the NG quality drawing, it may acquire a standard floor plan whose coordinates of the feature points match at least a predetermined percentage, or a standard floor plan whose coordinates of the feature points most match among the standard floor plans.

[0059] The feature point matching unit 138 generates a copy of the standard floor plan (step S704). The feature point matching unit 138 stores the copy of the standard floor plan in the real estate floor plan information storage unit 121 in association with the property information (step S705). Note that the feature point matching unit 138 may store the copy of the standard floor plan as a floor plan in the real estate floor plan information storage unit 121, or may store the copy as the standard floor plan separately from the floor plan.

[0060] As described above, since many real estate property floor plans are standard, even if the floor plan generated by the floor plan generation unit 134 is determined to be a quality-unacceptable drawing, a floor plan of a certain quality can be stored in the real estate floor plan information storage unit 121 by matching the quality-unacceptable drawing with the standard floor plan and replacing it with a standard floor plan with matching feature points. Also, even if the feature points of the quality-unacceptable drawing and the standard floor plan do not match, a floor plan of a certain quality close to the original drawing can be obtained by acquiring a standard floor plan with a certain percentage of matching feature points or the most matching standard floor plan and storing it in the real estate floor plan information storage unit 121. Also, by adding or modifying a floor plan of a certain quality close to the original drawing, a more complete floor plan can be created.

[0061] In the above process, the coordinates of the intersections of lines included in the drawing are used as an example of feature points, but coordinates of openings, etc. may also be added or substituted. Furthermore, when matching feature points of a quality-NG drawing with feature points of a standard floor plan, a reference point of the floor plan (e.g., the upper left corner of the floor plan or the center of gravity of the floor plan) may be aligned, and the scale of the quality-NG drawing may be matched to the scale of the standard floor plan. Furthermore, for drawings that are judged to be floor plans of a certain quality and in a predetermined format, feature point information for the drawing may be calculated, and the feature point information and the drawing may be stored as a standard floor plan in the characteristic floor plan storage unit 126.

[0062] By executing the above-described process, the floor plan generation server 100 generates floor plans in a predetermined format corresponding to publicly available property information and stores the property information and the floor plans in the predetermined format in the real estate floor plan information storage unit 121 in association with each other, thereby improving the real estate floor plan information storage unit 121. By providing such a real estate floor plan information storage unit 121, the information terminal device 300 operated by a user who wishes to purchase or rent a real estate property can access the real estate floor plan information storage unit 121 from a real estate property search page and display property information and floor plans in a predetermined format that match the search criteria by accepting input of search criteria. This allows the information terminal device 300 to display property information and floor plans in a consistent format that match the search criteria, allowing users to easily compare real estate properties. This also facilitates analysis based on floor plans for users analyzing real estate information. Furthermore, because the drawing formats are consistent, floor plans and 3D images in different formats can be easily generated from floor plans.

[0063] The hardware configuration of the floor plan generation server 100, web server 200, and information terminal device 300 according to the above-described embodiments is a typical computer including one or more processors such as a central processing unit (CPU), a microprocessing unit (MPU), or a graphics processing unit (GPU), and also including external storage devices such as read-only memory (ROM), random access memory (RAM), hard disk drives (HDDs), flash memory, and solid state drives (SSDs), a communication control device, and input / output devices. The CPU or other components read and run programs stored in the ROM, RAM, HDD, and the like, thereby realizing the above-described configuration and functions. The functions of each device may also be realized by electronic circuits such as application-specific integrated circuits (ASICs) and programmable logic devices (PLDs).

[0064] The programs that run on the floor plan generation server 100, the web server 200, and the information terminal device 300 may be stored on a computer connected to a network such as the Internet and provided by downloading via the network, or may be recorded in an installable or executable file format on a computer-readable recording medium such as a CD-ROM, DVD, USB memory, or SD card. Furthermore, the programs that realize the above-described functions and processes may be provided in the form of an API (Application Programming Interface), SaaS (Software as a Service), or cloud computing.

[0065] The present invention is not limited to the above-described embodiments, and does not necessarily have to be physically configured as illustrated. Furthermore, the present invention can be configured by functionally or physically dividing, integrating, replacing, modifying, or deleting all or part of the components described in the embodiments in any unit depending on various loads, usage conditions, etc. [Explanation of symbols]

[0066] 100... Floor plan generation server, 110... Communication unit, 120... Memory unit, 121... Real estate floor plan information memory unit, 122... NG drawing (floor plan) memory unit, 123... NG drawing (unnecessary element) memory unit, 124... NG drawing (quality) memory unit, 125... Trained model, 126... Standard floor plan memory unit, 130... Control unit, 131... Real estate property information acquisition unit, 132... Floor plan judgment unit, 133... Unnecessary element judgment unit, 134... Floor plan generation unit, 135... Quality judgment unit, 136... Learning unit, 137... Drawing acquisition unit, 138... Feature point matching unit, 200... Web server, 300... Information terminal device

Claims

1. A floor plan generation server, a real estate property information acquisition means for acquiring real estate property information in which property information for each real estate property is associated with a drawing relating to the real estate property; a floor plan determination means for determining whether the drawing is a floor plan; a floor plan generating means for generating a floor plan in a predetermined format from the drawing when the drawing is determined to be a floor plan; a quality determination means for determining whether the floor plan drawing in the predetermined format is of a certain quality; a storage means for storing the property information and the floor plan in the predetermined format in a real estate floor plan information storage unit in association with each other when the floor plan is determined to be of a certain quality; A floor plan generation server comprising:

2. an unnecessary element determining means for determining whether or not the drawing includes unnecessary elements; 2. The floor plan generation server according to claim 1, wherein said floor plan generation means generates a floor plan in a predetermined format from said drawing when it determines that said drawing is a floor plan and that said drawing does not contain unnecessary elements.

3. 3. The floor plan generating server according to claim 1, wherein said real estate property information acquiring means acquires said real estate property information by periodically executing a crawling process on a web server connected to a network.

4. the floor plan determination means receives a drawing as input and determines whether the drawing is a floor plan using a trained model that outputs an OK or NG flag as an output result indicating whether the drawing is a floor plan; The floor plan generation server according to claim 1, wherein the quality determination means determines whether the drawing is of a predetermined quality using a trained model that takes a drawing as input and outputs an OK or NG flag as an output result indicating whether the drawing is of a predetermined quality.

5. The floor plan generation server of claim 2, wherein the unnecessary element determination means determines whether the drawing contains unnecessary elements using a trained model that takes a drawing as input and outputs an OK or NG flag as an output result indicating whether the drawing contains unnecessary elements.

6. 6. The floor plan generation server according to claim 4 or claim 5, further comprising: a learning means for accepting input of a judgment result for a drawing for which an NG flag has been output, and re-training the trained model using the received judgment result and the drawing.

7. The storage means stores, when the trained model outputs an NG flag, the property information corresponding to the drawing for which the NG flag has been output in the real estate floor plan information storage unit; 6. The floor plan generating server according to claim 4, wherein said plan acquisition means acquires a plan corresponding to said property information by using said property information stored in said real estate floor plan information storage unit.

8. When the quality determination means determines that the floor plan drawing in the predetermined format is not of a certain quality, the storage means stores the property information and the drawing in the real estate floor plan information storage unit; 3. The floor plan generation server according to claim 1, further comprising a feature point matching means for matching feature points of the standard floor plan with feature points of the drawing, and if the feature points of the standard floor plan match the feature points of the drawing, storing the standard floor plan in the real estate floor plan information storage unit in association with the property information.

9. 1. A computer-implemented method comprising: a real estate information acquisition step of acquiring real estate property information in which property information for each real estate property is associated with a drawing relating to the real estate property; a floor plan determination step of determining whether the drawing is a floor plan; a floor plan generating step of generating a floor plan in a predetermined format from the drawing when it is determined that the drawing is a floor plan; a quality determination step of determining whether the floor plan drawing in the predetermined format is of a certain quality; a storing step of storing the property information and the floor plan in the predetermined format in a real estate floor plan information storage unit in association with each other when the floor plan is determined to be of a certain quality; A floor plan generation method comprising:

10. A floor plan generating program that causes a computer to execute the floor plan generating method according to claim 9.

Images