CN112747357B - Heating control method based on deep learning - Google Patents
Heating control method based on deep learning Download PDFInfo
- Publication number
- CN112747357B CN112747357B CN202110142784.1A CN202110142784A CN112747357B CN 112747357 B CN112747357 B CN 112747357B CN 202110142784 A CN202110142784 A CN 202110142784A CN 112747357 B CN112747357 B CN 112747357B
- Authority
- CN
- China
- Prior art keywords
- heating
- pipeline
- shelter
- shielding
- image
- 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.)
- Expired - Fee Related
Links
- 238000010438 heat treatment Methods 0.000 title claims abstract description 248
- 238000013135 deep learning Methods 0.000 title claims abstract description 26
- 238000000034 method Methods 0.000 title claims abstract description 24
- 230000011218 segmentation Effects 0.000 claims abstract description 22
- 230000008859 change Effects 0.000 claims abstract description 9
- 230000009471 action Effects 0.000 claims abstract description 8
- 238000012545 processing Methods 0.000 claims abstract description 6
- 238000013527 convolutional neural network Methods 0.000 claims description 7
- 238000003709 image segmentation Methods 0.000 claims description 6
- 238000000605 extraction Methods 0.000 claims description 5
- 230000008569 process Effects 0.000 claims description 5
- 238000009432 framing Methods 0.000 claims description 2
- 238000010586 diagram Methods 0.000 description 6
- 230000005855 radiation Effects 0.000 description 6
- 230000017525 heat dissipation Effects 0.000 description 5
- 238000012546 transfer Methods 0.000 description 5
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 description 5
- 238000012549 training Methods 0.000 description 4
- 238000004458 analytical method Methods 0.000 description 3
- 230000008901 benefit Effects 0.000 description 3
- 238000004364 calculation method Methods 0.000 description 3
- 238000010276 construction Methods 0.000 description 3
- 238000005034 decoration Methods 0.000 description 3
- 238000013461 design Methods 0.000 description 3
- 230000000694 effects Effects 0.000 description 3
- 238000013528 artificial neural network Methods 0.000 description 2
- 239000000463 material Substances 0.000 description 2
- 239000011159 matrix material Substances 0.000 description 2
- 238000011176 pooling Methods 0.000 description 2
- 208000006440 Open Bite Diseases 0.000 description 1
- 238000013459 approach Methods 0.000 description 1
- 230000004888 barrier function Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000005265 energy consumption Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 230000006870 function Effects 0.000 description 1
- 239000008236 heating water Substances 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 230000037081 physical activity Effects 0.000 description 1
- 230000002035 prolonged effect Effects 0.000 description 1
- 238000013468 resource allocation Methods 0.000 description 1
Images
Classifications
-
- F—MECHANICAL ENGINEERING; LIGHTING; HEATING; WEAPONS; BLASTING
- F24—HEATING; RANGES; VENTILATING
- F24D—DOMESTIC- OR SPACE-HEATING SYSTEMS, e.g. CENTRAL HEATING SYSTEMS; DOMESTIC HOT-WATER SUPPLY SYSTEMS; ELEMENTS OR COMPONENTS THEREFOR
- F24D19/00—Details
- F24D19/10—Arrangement or mounting of control or safety devices
- F24D19/1006—Arrangement or mounting of control or safety devices for water heating systems
- F24D19/1009—Arrangement or mounting of control or safety devices for water heating systems for central heating
- F24D19/1015—Arrangement or mounting of control or safety devices for water heating systems for central heating using a valve or valves
-
- 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
- G06N3/08—Learning methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/11—Region-based segmentation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10016—Video; Image sequence
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20081—Training; Learning
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20084—Artificial neural networks [ANN]
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02B—CLIMATE CHANGE MITIGATION TECHNOLOGIES RELATED TO BUILDINGS, e.g. HOUSING, HOUSE APPLIANCES OR RELATED END-USER APPLICATIONS
- Y02B30/00—Energy efficient heating, ventilation or air conditioning [HVAC]
Abstract
The invention discloses a heating control method based on deep learning, which belongs to the technical field of heating and comprises the following steps: s1: collecting image data of a floor radiant heating area in real time through a camera device; s2: processing image data acquired by a camera device through an image semantic segmentation model based on deep learning, and identifying the position, the area, the shielding degree and the change condition information of a heating shielding object; s3: the heating pipeline control system calculates the areas needing to be reduced or stopped for heating by utilizing the position, the area, the shielding degree and the variation condition information of the heating shielding object identified by the image semantic segmentation model based on deep learning through a decision algorithm model, makes a corresponding heating decision, controls the action of the heating pipeline system, and reduces or cuts off the heating of the corresponding position of the heating shielding object; s4: and keeping the current heating decision, and periodically and circularly executing the steps to update the heating decision.
Description
Technical Field
The invention belongs to the technical field of heating, and particularly relates to a heating control method based on deep learning.
Background
The Floor Heating is short for Floor radiation Heating, and is called radiation Floor Heating, the whole Floor is uniformly heated by using a heat medium in a Floor radiation layer as a radiator, and heat is supplied to the indoor space by using the Floor in a radiation and convection heat transfer mode, so that the purpose of comfortable Heating is achieved. The water floor heating system is divided into a water floor heating system and an electric floor heating system according to different heat transfer media, wherein the water floor heating system adopts hot water with the temperature not higher than 60 ℃ as a heat transfer medium, the hot water circularly flows in a heating pipe embedded in a filling layer under the ground to heat the whole floor, and the heat is supplied to the indoor space through the ground in a heat transfer mode of radiation and convection.
Because the floor heating applied in large scale in the current market heats all indoor areas in a radiation and convection heat transfer mode, effective, inefficient and ineffective heating areas are not distinguished, and huge resources are wasted during the heating period every year. In the building interior, the form and style of the sheltered heating area and the effective heating area are changeable, and the regularity is low. If the original machine-learned approach of artificially designing image features is used, different features must be designed for different types of occlusions. For example, in designing a method for detecting a semi-shielding type shield such as an indoor table and chair, it is not necessarily suitable for a full-shielding type shield such as a cabinet; the method for detecting the shielding object with the place with dense human body activities as the background is not necessarily suitable for detecting the shielding object with the place with the freight warehouse as the main place as the background. Under the complex and changeable application scene, the traditional machine learning method is not free from worry and high in construction cost.
Disclosure of Invention
In order to solve the defects in the prior art, the invention provides a heating control method based on deep learning, which is used for judging the position condition of effective heating in a building according to the actual environment, performing targeted heating resource allocation, reducing the energy consumption of actual heating and reducing the heating cost on the premise of ensuring that the heating effect is not changed.
The technical scheme adopted by the invention for solving the technical problem is as follows: a heating control method based on deep learning comprises the following steps:
s1: collecting image data of a floor radiant heating area in real time through a camera device;
s2: processing image data acquired by a camera device through an image semantic segmentation model based on deep learning, and identifying the position, the area, the shielding degree and the change condition information of a heating shielding object;
s3: the heating pipeline control system calculates the areas needing to be reduced or stopped for heating by utilizing the information of the position, the area, the shielding degree and the change condition of the heating shielding object identified by the image semantic segmentation model based on deep learning through a decision algorithm model, makes a corresponding heating decision, controls the action of the heating pipeline system, and reduces or cuts off the heating of the corresponding position of the heating shielding object;
s4: and keeping the current heating decision, and periodically and circularly executing the steps to update the heating decision.
Furthermore, the heating shelters are divided into full shelters, half shelters and no shelters according to the sheltering degree processed by the image semantic segmentation model based on deep learning.
Further, discernment heating shelter from thing position, size, shelters from degree and change situation information, include: the method comprises the steps of framing an image acquired by a camera device, extracting the image, dividing the extracted image into a plurality of sub-images, sending the sub-images into a convolutional neural network, performing semantic division on the sub-images, judging whether a heating shelter exists in the sub-images, if so, identifying the corresponding shelter degree of the heating shelter, and finally correspondingly listing the identified heating shelter into full shelter information statistics and half shelter information statistics according to the type of a label according to the judgment result; and after all the subimages are judged, combining the identification results to obtain the complete heating shelter information of the identification area.
Further, the image extraction includes: reading in an original video stream of the camera device, creating a corresponding directory for storing fixed frame images, acquiring a video frame rate of the video, reading video frames, and performing storage operation every fixed frame.
Further, the dividing the extracted image into a plurality of sub-images comprises: reading the fixed frame image in storage, calculating the size of the original image, segmenting and extracting the area with the fixed pixel size, and performing storage operation until the original image is completely extracted, thereby completing image segmentation.
Furthermore, the heating pipeline system comprises two layers of heating pipelines, wherein one layer of heating pipeline is transversely arranged, the other layer of heating pipeline is longitudinally arranged, and the two layers of heating pipelines are vertically arranged and jointly cover a heating area; each layer of heat supply pipeline consists of a plurality of single heat supply pipelines which are sequentially arranged at intervals, and each single heat supply pipeline of each layer of heat supply pipeline is controlled to be switched on and switched off by an electric control valve which is connected with a heat source; the intersection of the horizontal and vertical heat supply pipelines is set as a heating node, and the heating node performs on-off action on the heat supply pipeline along with the corresponding network node control signal.
Furthermore, each heating node is numbered, and the control of the heating pipeline control system on the heating nodes comprises the following steps: the method comprises the steps that a heating pipeline control system obtains position information and shielding degree information of a heating shielding object through a decision algorithm model, processes and calculates the obtained information, and finally converts the information into actual position coordinates, shielding range size and shielding degree parameter information of the heating shielding object, positions heating node numbers of the positions of the heating shielding object, and sends control signals to heating nodes related to the positions of the heating shielding object corresponding to an electric control valve of a heating pipeline according to the parameter information so as to control the specific heating position and heating amount of the heating pipeline system;
judging whether the heating shelter is full shelter or not, and shutting down the network nodes at the corresponding positions when the heating shelter is full shelter, so as to shut down the two single pipeline electric control valves corresponding to the heating node numbers;
when the total shielding is judged to be not full shielding, calculating the weights of the two single pipeline shields corresponding to the heating node number, and judging whether the weights of the two single pipeline shields are the same; if the number of the shutdown numbers of the pipelines of the layer to which the pipeline belongs is the same, shutting down the single pipeline with a small shutdown number, and if the number of the shutdown numbers of the pipelines of the layer to which the pipeline belongs is the same, shutting down one single pipeline randomly; if the weights of the two single pipeline shelters are not equal, shutting down one single pipeline with the higher weight of the shelter;
and when the network nodes are judged to be not shielded, the network nodes at the corresponding positions normally supply heat, and the two electric control valves of the heating pipelines corresponding to the network nodes are kept in a normally open state.
Further, the weight Y of the obstruction is calculated according to the following formula:
wherein X represents the number of heating shelters passing through the same pipeline, A represents the sum of the sheltering degrees of all the heating shelters along the way, wherein each full shelter counts 0.5, and each half shelter counts 0.2.
The invention has the beneficial effects that: the position, the size and the shielding degree of a ground heating shielding object in a building are judged according to the actual environment, and a matched heating system is controlled according to the position, the size and the shielding degree to further control the heating area and the flow, heating resources are allocated in a targeted manner, so that a novel efficient and energy-saving heating mode is realized, the actual heating energy loss is reduced on the premise of ensuring that the heating effect is unchanged, the heating cost is reduced, and the heating economic benefit is improved.
Drawings
FIG. 1 is a flow chart of the present invention for identifying and detecting a heating covering;
FIG. 2 is a flow chart of image extraction according to the present invention;
FIG. 3 is a sub-image segmentation flow diagram of the present invention;
FIG. 4 is a schematic diagram of the full shade level of the present invention;
FIG. 5 is a schematic view of the half occlusion degree of the present invention;
FIG. 6 is a schematic view of the present invention showing the degree of non-occlusion;
FIG. 7 is a schematic diagram of the image semantic segmentation model design of the present invention;
FIG. 8 is a schematic diagram of a heating pipeline control system decision algorithm design according to the present invention;
FIG. 9 is a block diagram of the heating duct system of the present invention;
FIG. 10 is a diagram illustrating the shutdown of the heat supply pipeline according to the present invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention will be described in detail with reference to the accompanying drawings and specific embodiments.
Example 1
A heating control method based on deep learning is mainly applied to large-scale office places, a large amount of furniture and office decorations need to be erected, and a large amount of scenes which cannot effectively heat areas are generated. The present invention is intended to determine an optimal heating plan and implement the same.
The whole technical scheme is as follows: the existing camera device is deployed or utilized in a building, the image information of the position of a ground heating shelter such as furniture in the building is collected, an image is analyzed by using an image semantic segmentation model based on deep learning, the position and the shielding degree of the shelter are judged, and a control system controls the flow and the on-off of a heating pipeline according to the analysis result of the image semantic segmentation model based on the deep learning.
The method specifically comprises the following steps: firstly, an image semantic segmentation model based on deep learning: and identifying the position, size and shielding degree of the ground heating shielding object in the building. The main types of semantic segmentation inside the building are a shielded heating area and an effective heating area, wherein the shielded heating area is divided into a half shielded area and a full shielded area, the heat cycle efficiency of the shielded heating area is comprehensively judged according to the actual state of the shielded object, such as the actual conditions of the shielded area, such as half opening, and the like, so that more detailed classification is carried out.
1. The method for detecting the heating shelter in the building comprises the following steps:
the deep learning does not need to artificially design picture characteristics, and only needs to build a corresponding neural network, collect a sufficient training sample set and train the neural network, so that a better classification and segmentation effect can be achieved. In the method for detecting the heating shelters in the building, the feature extraction and classification capability of the convolutional neural network based on deep learning is key. Referring to fig. 1, frames are cut for images collected by a camera device, the images are extracted, the extracted images are divided into a plurality of sub-images, the sub-images are sent to a convolutional neural network, semantic division is carried out on the sub-images, whether heating shelters exist in the sub-images or not is judged, if yes, the sheltering degree corresponding to the heating shelters is identified, and finally the identified heating shelters are correspondingly listed into full shelter information statistics and half shelter information statistics according to the types of labels according to the judgment result; and after all the subimages are judged, combining the identification results to obtain the complete heating shelter information of the identification area.
1.1, image extraction:
the video is essentially the stacking of one frame of image, if the video is processed, each frame of the video does not need to be processed, the front frame and the rear frame have time sequence information, and some redundancy exists. For the practical situation of identifying the ground heating shelter in the building, most of the indoor shelters in the building are large in movement cost and are fixed and unchangeable for a long time, so that the standard video data does not need to be detected completely in thirty frames per second, and only specific frame images need to be extracted at regular time for analysis. Referring to fig. 2, a specific implementation flow is as follows: reading in an original video stream of the camera device, creating a corresponding directory for storing fixed frame images, acquiring a video frame rate of the video, reading video frames, and performing storage operation every fixed frame.
1.2 image segmentation:
because the space in different buildings is different in size, and the original video frame obtained from the original image is too large, the network is too wide when the original video frame is directly sent into the convolutional neural network for operation, and the calculated amount caused by the width is increased by a square number, the calculation time can be greatly prolonged when the original image is directly calculated, and even the final recognition result cannot be obtained. Therefore, the obtained original image is divided into a plurality of 200 × 200 RGB three-color channel sub-images, and then the sub-images are sequentially sent into a convolutional neural network to judge whether a heating shelter exists. Referring to fig. 3, the image segmentation process: reading the fixed frame image in storage, calculating the size of the original image, segmenting and extracting the area with the fixed pixel size, and performing storage operation until the original image is completely extracted, thereby completing image segmentation.
2. And (3) judging a heating shelter:
2.1, heating shelter classification:
the heating shelter that defines in this application contains all objects that can shelter from the ground heating region in the broad sense to according to sheltering from the different three types that divide into of degree: full occlusion, half occlusion, no occlusion. Wherein, the definition of the full shading is that the full shading is a ground-attached chassis, a heat dissipation space without air circulation or a heat dissipation space approximate to the heat dissipation space without air circulation, such as a storage bed cabinet, a storage bed and the like with the bottom directly attached to the ground; the definition of the semi-shielding is that the semi-shielding is fixed on the ground, but the fixed frame occupies a small area, and the chassis has a certain air circulation heat dissipation space away from the ground, such as a bed with an overhead bottom; the definition of no shielding is that the bottom fixing frame is very small, and the bottom surface has a higher space from the ground, such as a desk and a chair with a suspended bottom. Examples of classifications are shown in fig. 4-6.
2.2, judging the type of the heating shelter:
the image semantic segmentation model trained by a large number of sample data sets can accurately identify all sample types in the internal images of the building, and after the types of backgrounds such as the ground, walls and the like are removed, the analysis and classification of the shielding objects can be carried out on other objects. Accurately identifying all articles in the picture through a segmentation model of a convolutional neural network, accurately delineating the boundary of the articles in the image, judging the size of a gap between a chassis and the ground according to the intersection degree of the border of the articles in the image, the ground background and the wall background, and classifying. Such as storage beds, wardrobes, bedside cabinets and the like, the chassis is directly contacted with the ground, and no full-shielding type article with a certain air circulation heat dissipation space exists. And identifying three key information of the position, the size and the shielding degree of the shielding object.
The semantic segmentation model of the training image selects a DeepLab v3+ network, the DeepLab v3+ network is based on the DeepLab v3, the original structure is used as an encoder, and a decoder is additionally added, so that the multi-scale feature capturing capability of the spatial pyramid pooling structure is combined with the capability of the encoder-decoder structure for processing the boundary details. The depth separable convolution in the Xceptance is applied to a cavity space pyramid pooling (ASPP) and decoding module, and the speed and the precision are further improved. The algorithm can control the feature resolution by adjusting the hole convolution, so that the trade-off between the precision and the speed is flexible, and the principle is shown in FIG. 7.
Heating system
The heating pipeline control system calculates the areas needing to be reduced or stopped for heating by using the information of the position, the size, the shielding degree and the change condition of the heating shielding object identified by the image semantic segmentation model based on deep learning through a decision algorithm model, makes a corresponding heating decision, controls the action of the heating pipeline system, and reduces or cuts off the heating of the corresponding position of the heating shielding object.
1. Heating decision of heating pipeline control system
After the indoor image of the building is processed by the image semantic segmentation model, the position, the size, the shielding degree and other information of the ground heating shielding object in the building can be accurately identified, the information is sent to a heating pipeline control system for processing and calculation, and finally the information is converted into the actual position coordinate, the shielding range size and the shielding degree information of the ground heating shielding object in the building. And the heating pipeline control system can make a corresponding heating adjustment decision according to the identified result, control the action of the heating pipeline system and change the heating strategy.
2. Heat supply pipeline system
The heat supply pipeline is designed into a matrix with two intersecting directions, and the heating power or flow of each area is intelligently controlled by using the principle of matrix element positioning, so that the performance requirement of intelligent heating is met. As shown in fig. 9, the heating pipe system includes two layers of heating pipelines, wherein one layer of heating pipeline is disposed horizontally, the other layer of heating pipeline is disposed longitudinally, and the two layers of heating pipelines are disposed up and down to cover the heating area together; each layer of heat supply pipeline consists of a plurality of single heat supply pipelines which are sequentially arranged at intervals, and each single heat supply pipeline of each layer of heat supply pipeline is controlled to be switched on and switched off by an electric control valve which is connected with a heat source; the intersection of the horizontal and vertical heat supply pipelines is set as a heating node, and the heating node performs on-off action on the heat supply pipeline along with the corresponding network node control signal.
Each heating node is provided with an upper single heating pipeline and a lower single heating pipeline, namely two heating elements. Therefore, the heating quantity of the corresponding area can be controlled only by controlling the on-off of the electric control valves of the two single heating pipelines which are intersected at the corresponding positions, and the full on, the half on and the full off of the heating power of the area are correspondingly adjusted by fully opening, opening one switch and fully closing the two heating pipelines transversely and longitudinally. For example, as shown in fig. 10, when there is a full-shielding article in the element X1 area, all heating power needs to be shut down, and at this time, an electrically controlled valve is needed to shut down two pipes X1 and Y5; half of shielding articles exist in the element X2 area, heating power needs to be reduced, at the moment, one of the Y2 and the X6 is turned off, half of the heating power of the area where the element X2 is located can be reduced, and other heating pipelines normally heat.
| Region numbering | Types of barriers | Need to implement function | Policy | Detailed description of the invention |
| Region of element X1 | Full-shielding article | Shut down of total heating power | All closing | Electric control valve closing X1 and Y5 two pipelines |
| Region of element X2 | Semi-shielding article | Reducing heating power | Half-open | Shutting down any one of two pipelines Y2 and X6 |
| Region of elements | Article without shelter | Normal heating | Full open | Keeping two pipeline electric control valves in normal open state |
Numbering each heating node, as shown in fig. 8, the control of the heating node by the heating pipe control system comprises the following steps: the heating pipeline control system acquires position information and shielding degree information of a heating shielding object through a decision algorithm model, processes and calculates the acquired information, finally converts the information into actual position coordinates, shielding range size and shielding degree parameter information of the heating shielding object, positions a heating node number where the position of the heating shielding object is located, and sends a control signal to a heating pipeline electric control valve corresponding to a heating node related to the position of the heating shielding object according to the parameter information so as to control the specific heating position and heating amount of the heating pipeline system;
judging whether the heating shelter is full shelter or not, and shutting down the network nodes at the corresponding positions when the heating shelter is full shelter, so as to shut down the two single pipeline electric control valves corresponding to the heating node numbers;
when the total shielding is judged to be not full shielding, calculating the weights of the two single pipeline shields corresponding to the heating node number, and judging whether the weights of the two single pipeline shields are the same; if the number of the shutdown numbers of the pipelines of the layer to which the pipeline belongs is the same, shutting down the single pipeline with a small shutdown number, and if the number of the shutdown numbers of the pipelines of the layer to which the pipeline belongs is the same, shutting down one single pipeline randomly; and if the weights of the two single pipeline shelters are not equal, shutting down the single pipeline with the higher weight of the shelter.
And when the network nodes are judged to be not shielded, the network nodes at the corresponding positions normally supply heat, and the two electric control valves of the heating pipelines corresponding to the network nodes are kept in a normally open state.
And the heating nodes needing to reduce heating power only need to close one of the two heating pipelines. How to select which pipeline to shut down depends on the number of the shelters in the area passed by the two pipelines and the judgment of the calculation result of the weighting coefficient of the sheltering degree. The specific formula for calculating the weight Y of the shielding object is as follows:
wherein X represents the number of heating shelters passing through the same pipeline, A represents the sum of the sheltering degrees of all the heating shelters along the way, wherein each full shelter counts 0.5, and each half shelter counts 0.2.
In the floor heating laying process, two layers are required to be laid in the transverse and longitudinal directions when pipelines are erected. An electric control valve is arranged on the floor heating water separator. Other floor heating construction modes, flows and materials are the same as those of the traditional floor heating construction mode, flow and materials.
The technical route is summarized as follows:
1. collecting a training image semantic segmentation model to identify a data set and a sample set (such as furniture, office appliances, indoor decorations and the like) of various shielding articles;
2. training an image semantic segmentation model based on deep learning;
3. and (3) processing the effective and ineffective heating areas judged based on the image semantic segmentation model through a decision algorithm model to obtain a specific heating decision, sending the specific heating decision into a heating pipeline system, and automatically adjusting the heating areas and the power.
Example 2
Business model and expected benefits:
the system is operated in cooperation with companies needing large-area or multi-room centralized heating such as office buildings and large office places, and all heating services are carried by adopting the technical scheme of the embodiment 1 from the laying of the heating pipelines.
The prospective market is: except individual users, almost all large-scale office places have to be provided with a large amount of furniture and office decorations, a large number of areas which cannot be heated effectively are generated, the current heating charges are collected by flow counting instead of actual consumption, and a large number of room heating systems can be modified.
And (3) prospect: due to technical and cost limitations, the technical scheme can be used for cooperative operation of companies needing large-area or multi-room central heating in office buildings, large office places and the like at present, and is expected to become standard configuration of future resident buildings along with large-scale mass production and mature technology.
The above description is only for the preferred embodiment of the present invention, but the scope of the present invention is not limited thereto, and any person skilled in the art should be able to cover the technical solutions and the inventive concepts of the present invention within the technical scope of the present invention.
Claims (4)
1. A heating control method based on deep learning is characterized by comprising the following steps:
s1: collecting image data of a floor radiant heating area in real time through a camera device;
s2: processing image data acquired by a camera device through an image semantic segmentation model based on deep learning, and identifying the position, the area, the shielding degree and the change condition information of a heating shielding object;
the heating sheltering object is divided into full shelter, half shelter and no shelter according to the shelter degree processed by the image semantic segmentation model based on deep learning;
discernment heating shelters from thing position, size, shelters from degree and change condition information, include: the method comprises the steps of framing an image acquired by a camera device, extracting the image, dividing the extracted image into a plurality of sub-images, sending the sub-images into a convolutional neural network, performing semantic division on the sub-images, judging whether a heating shelter exists in the sub-images, if so, identifying the corresponding shelter degree of the heating shelter, and finally correspondingly listing the identified heating shelter into full shelter information statistics and half shelter information statistics according to the type of a label according to the judgment result; after all the subimages are judged, combining the identification results to obtain complete heating shelter information of the identification area;
s3: the heating pipeline control system calculates the areas needing to be reduced or stopped for heating by utilizing the information of the position, the area, the shielding degree and the change condition of the heating shielding object identified by the image semantic segmentation model based on deep learning through a decision algorithm model, makes a corresponding heating decision, controls the action of the heating pipeline system, and reduces or cuts off the heating of the corresponding position of the heating shielding object;
the heating pipeline system comprises two layers of heating pipelines, wherein one layer of heating pipeline is transversely arranged, the other layer of heating pipeline is longitudinally arranged, and the two layers of heating pipelines are vertically arranged and jointly cover a heating area; each layer of heat supply pipeline consists of a plurality of single heat supply pipelines which are sequentially arranged at intervals, and each single heat supply pipeline of each layer of heat supply pipeline is controlled to be switched on and switched off by an electric control valve which is connected with a heat source; the intersection of the horizontal and vertical layers of heat supply pipelines is set as a heating node, and the heating node performs on-off action on the heat supply pipeline along with a corresponding network node control signal;
numbering each heating node, wherein the control of the heating pipeline control system on the heating nodes comprises the following steps: the heating pipeline control system acquires position information and shielding degree information of a heating shielding object through a decision algorithm model, processes and calculates the acquired information, finally converts the information into actual position coordinates, shielding range size and shielding degree parameter information of the heating shielding object, positions a heating node number where the position of the heating shielding object is located, and sends a control signal to a heating pipeline electric control valve corresponding to a heating node related to the position of the heating shielding object according to the parameter information so as to control the specific heating position and heating amount of the heating pipeline system;
judging whether the heating shelter is full shelter or not, and shutting down the network nodes at the corresponding positions when the heating shelter is full shelter, so as to shut down the two single pipeline electric control valves corresponding to the heating node numbers;
when the total shielding is judged to be not full shielding, calculating the weights of the two single pipeline shields corresponding to the heating node number, and judging whether the weights of the two single pipeline shields are the same; if the number of the shutdown numbers of the pipelines of the layer to which the pipeline belongs is the same, shutting down the single pipeline with a small shutdown number, and if the number of the shutdown numbers of the pipelines of the layer to which the pipeline belongs is the same, shutting down one single pipeline randomly; if the weights of the two single pipeline shelters are not equal, shutting down one single pipeline with the higher weight of the shelter;
when the network nodes are judged to be not shielded, the network nodes at the corresponding positions normally supply heat, and the two heating pipeline electric control valves corresponding to the network nodes are kept in a normally open state;
s4: and keeping the current heating decision, and periodically and circularly executing the steps to update the heating decision.
2. The deep learning-based heating control method according to claim 1, wherein the image extraction comprises: reading in an original video stream of the camera device, creating a corresponding directory for storing fixed frame images, acquiring a video frame rate of the video, reading video frames, and performing storage operation every fixed frame.
3. The deep learning-based heating control method according to claim 1, wherein the dividing the extracted image into a plurality of sub-images comprises: reading the fixed frame image in storage, calculating the size of the original image, segmenting and extracting the area with the fixed pixel size, and performing storage operation until the original image is completely extracted, thereby completing image segmentation.
4. The deep learning-based heating control method according to claim 1, wherein the weight Y of the shelter is calculated according to the following formula:
wherein X represents the number of heating shelters passing through the same pipeline, A represents the sum of the sheltering degrees of all the heating shelters along the way, wherein each full shelter counts 0.5, and each half shelter counts 0.2.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110142784.1A CN112747357B (en) | 2021-02-02 | 2021-02-02 | Heating control method based on deep learning |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110142784.1A CN112747357B (en) | 2021-02-02 | 2021-02-02 | Heating control method based on deep learning |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN112747357A CN112747357A (en) | 2021-05-04 |
| CN112747357B true CN112747357B (en) | 2022-05-20 |
Family
ID=75653537
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110142784.1A Expired - Fee Related CN112747357B (en) | 2021-02-02 | 2021-02-02 | Heating control method based on deep learning |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112747357B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115127132B (en) * | 2022-05-16 | 2023-10-24 | 北京正弘奥能环保科技有限公司 | Control method and device of radiant heating host and computer equipment |
Family Cites Families (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105805822B (en) * | 2016-03-24 | 2018-11-13 | 常州英集动力科技有限公司 | Heating energy-saving control method based on neural network prediction and system |
| CN108240679B (en) * | 2018-02-22 | 2019-07-02 | 烟台科创捷能机电工程有限公司 | A kind of heat supply method based on building heating load prediction, device and system |
| CN109724137A (en) * | 2018-07-23 | 2019-05-07 | 永康市蜂蚁科技有限公司 | Multifunctional aluminium-copper composite heat supply pipe |
| CN109915931B (en) * | 2018-07-25 | 2021-02-26 | 临海市小核桃工业设计服务部 | Expanded steel-aluminum composite heating pipe |
| CN110836415A (en) * | 2019-12-12 | 2020-02-25 | 无锡热能在线科技有限公司 | Intelligent heating control system and method based on artificial intelligence and Internet of things big data |
-
2021
- 2021-02-02 CN CN202110142784.1A patent/CN112747357B/en not_active Expired - Fee Related
Also Published As
| Publication number | Publication date |
|---|---|
| CN112747357A (en) | 2021-05-04 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Benezeth et al. | Towards a sensor for detecting human presence and characterizing activity | |
| CN112747357B (en) | Heating control method based on deep learning | |
| CN107622244A (en) | A kind of indoor scene based on depth map becomes more meticulous analytic method | |
| CN106250895B (en) | A kind of remote sensing image region of interest area detecting method | |
| CN107895080B (en) | Dynamic load estimation and fresh air control method for large-space building air conditioner integrating image information | |
| CN103617632B (en) | A kind of moving target detecting method of combination neighbor frame difference method and mixed Gauss model | |
| CN109063723A (en) | The Weakly supervised image, semantic dividing method of object common trait is excavated based on iteration | |
| CN104899883B (en) | A kind of room objects cube body detecting method of depth image scene | |
| CN109874198B (en) | Commercial hotel guest room illumination control device based on scene automatic identification | |
| CN106152408A (en) | Intelligent air conditioner controller, control method and air-conditioner | |
| CN107256386A (en) | Human behavior analysis method based on deep learning | |
| CN113509104B (en) | Cleaning method, storage medium and cleaning robot | |
| CN101572803A (en) | Customizable automatic tracking system based on video monitoring | |
| CN206948686U (en) | Illumination control apparatus and system | |
| CN102509346A (en) | Object illumination migration method based on edge retaining | |
| CN109508675A (en) | A kind of pedestrian detection method for complex scene | |
| CN107135595A (en) | Illumination control apparatus and system | |
| CN104346801A (en) | Image-composition evaluating device, information processing device and method thereof | |
| CN102831472A (en) | People counting method based on video flowing image processing | |
| CN107152756A (en) | The control method and device of air conditioner | |
| CN106545949B (en) | Site museum tourist corridor local environment regulation and control system | |
| CN110096793A (en) | A kind of house energy consumption prediction technique based on residential subscribers' activity pattern | |
| CN105740814B (en) | A method of determining solid waste dangerous waste storage configuration using video analysis | |
| CN102622581B (en) | Face detection method and face detection system | |
| CN108921593A (en) | A kind of open air large-size screen monitors intelligent management system |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20220520 |
|
| CF01 | Termination of patent right due to non-payment of annual fee |