Three-dimensional modeling method, device and electronic equipment applied to power distribution room

Abstract

Embodiments of the present disclosure disclose a three-dimensional modeling method, device and electronic equipment applied to a power distribution room. A specific embodiment of the method comprises: generating a set of image quality improved images according to a set of target images; performing image stitching on each image quality improved image in the set of image quality improved images; generating an initial power distribution room three-dimensional model and model description information according to a stitching image and a pre-trained initial model; for model element description information, performing the following processing steps: determining whether there is a model element matching the model element description information in a pre-constructed model element library according to an element type and an element feature vector included in the model element description information; in response to the existence, determining the model element matching the model element description information as a candidate model element; and performing model element updating on the initial power distribution room three-dimensional model. The embodiment shortens the modeling cycle and reduces the consumption of computing resources.

Description

Technical Field

[0001] Embodiments of this disclosure relate to the field of computer technology, and more specifically to a three-dimensional modeling method, apparatus, and electronic device for use in power distribution rooms. Background Technology

[0002] As a crucial facility for power supply, the power distribution room can be transformed into a detailed monitoring system through 3D modeling. This effectively establishes the correspondence between the power distribution room and the 3D model, facilitating subsequent control and management of the power distribution room based on the 3D model. Currently, common 3D modeling methods include using techniques such as point cloud modeling.

[0003] However, when using the above method, the following technical problem often occurs:

[0004] Point cloud modeling involves a huge amount of data processing, which increases the modeling cycle and the consumption of computing resources.

[0005] Furthermore, when performing 3D modeling based on images, the following technical problem arises:

[0006] Image-based modeling methods often result in 3D models with poor accuracy.

[0007] The information disclosed in this background section is only intended to enhance the understanding of the background of the inventive concept, and therefore may contain information that does not form prior art known to those skilled in the art. Summary of the Invention

[0008] The summary portion of this disclosure is intended to provide a brief overview of the concepts, which will be described in detail in the detailed description portion. This summary portion is not intended to identify key or essential features of the claimed technical solutions, nor is it intended to limit the scope of the claimed technical solutions.

[0009] Some embodiments of this disclosure propose a three-dimensional modeling method, apparatus, and electronic device for use in power distribution rooms to solve one or more of the technical problems mentioned in the background section above.

[0010] In a first aspect, some embodiments of this disclosure provide a 3D modeling method for a power distribution room. The method includes: acquiring a target image set, wherein the target image set is a pre-acquired wide-angle image of the power distribution room to be 3D modeled; generating an image set with improved image quality based on the target image set, wherein the improved image quality is an image obtained by locally improving the image quality of at least one target image; stitching together the improved image quality images in the image set to generate a stitched image; and generating an initial 3D model of the power distribution room and model description information based on the stitched image and a pre-trained initial 3D model, wherein the model description information includes: a set of model element description information, a model... The element description information includes: element type, element feature vector, and element position information. The element position information represents the position of the model element corresponding to the model element description information in the initial 3D model of the power distribution room. For each model element description information in the above model element description information set, the following processing steps are performed: Based on the element type and element feature vector included in the above model element description information, determine whether there is a model element in the pre-built model element library that matches the above model element description information; In response to the existence, determine the model element that matches the above model element description information as a candidate model element; Based on the obtained candidate model element set, update the model elements of the above initial 3D model of the power distribution room to generate an updated 3D model of the power distribution room.

[0011] Secondly, some embodiments of this disclosure provide a 3D modeling device for a power distribution room. The device includes: an acquisition unit configured to acquire a target image set, wherein the target image set is a pre-acquired wide-angle image of the power distribution room to be 3D modeled; a first generation unit configured to generate an image set with improved image quality based on the target image set, wherein the image set with improved image quality is an image obtained by locally improving the image quality of at least one target image; an image stitching unit configured to stitch together the images in the image set with improved image quality to generate a stitched image; and a second generation unit configured to generate an initial 3D model of the power distribution room and model description information based on the stitched image and a pre-trained initial 3D model, wherein the model description information includes: model elements. The model element description information set includes: element type, element feature vector, and element position information. The element position information represents the position of the model element corresponding to the model element description information in the initial 3D model of the power distribution room. The execution unit is configured to perform the following processing steps for each model element description information in the above model element description information set: determine whether there is a model element matching the above model element description information in the pre-built model element library based on the element type and element feature vector included in the above model element description information; in response to the existence, determine the model element matching the above model element description information as a candidate model element; and the model element update unit is configured to update the model elements of the above initial 3D model of the power distribution room based on the obtained candidate model element set to generate an updated 3D model of the power distribution room.

[0012] Thirdly, some embodiments of this disclosure provide an electronic device, including: one or more processors; and a storage device having one or more programs stored thereon, wherein when the one or more programs are executed by the one or more processors, the one or more processors implement the method described in any implementation of the first aspect above.

[0013] Fourthly, some embodiments of this disclosure provide a computer-readable medium having a computer program stored thereon, wherein the program, when executed by a processor, implements the method described in any of the implementations of the first aspect above.

[0014] The various embodiments of this disclosure have the following beneficial effects: the 3D modeling method for power distribution rooms according to some embodiments of this disclosure shortens the modeling cycle and reduces the consumption of computing resources. Specifically, the reason for the long modeling cycle is that the point cloud modeling method involves a huge amount of data processing. Based on this, the 3D modeling method for power distribution rooms according to some embodiments of this disclosure first acquires a target image set, wherein the target image set is a pre-acquired wide-angle image of the power distribution room to be 3D modeled. Compared with the point cloud modeling method, the image-based 3D modeling method has the advantage of convenient image acquisition. Secondly, based on the target image set, an image set with improved image quality is generated, wherein the image with improved image quality is an image obtained by locally improving the image quality of at least one target image. The image quality improvement method ensures the accuracy of the subsequently generated 3D model. Next, the images in the image set with improved image quality are stitched together to generate a stitched image. Furthermore, based on the stitched images and the pre-trained initial 3D model, an initial 3D model of the power distribution room and model description information are generated. The model description information includes a set of model element description information, which includes element type, element feature vector, and element position information. The element position information represents the position of the model element corresponding to the model element description information within the initial 3D model of the power distribution room. This, combined with the image, generates an initial 3D model of the power distribution room to be 3D modeled. In addition, for each model element description information in the set, the following processing steps are performed: First, based on the element type and element feature vector included in the model element description information, it is determined whether a model element matching the model element description information exists in the pre-constructed model element library. Second, in response to the existence of a matching model element description information, the model element matching the model element description information is identified as a candidate model element. Finally, based on the obtained candidate model element set, the initial 3D model of the power distribution room is updated to generate an updated 3D model of the power distribution room. By using matching model element replacement, the model accuracy of the updated 3D model of the power distribution room is further improved. This method generates 3D models using only images, which shortens the modeling cycle and reduces computational resource consumption compared to point cloud modeling. Attached Figure Description

[0015] The above and other features, advantages, and aspects of the embodiments of this disclosure will become more apparent from the accompanying drawings and the following detailed description. Throughout the drawings, the same or similar reference numerals denote the same or similar elements. It should be understood that the drawings are schematic, and elements are not necessarily drawn to scale.

[0016] Figure 1This is a flowchart of some embodiments of the three-dimensional modeling method for power distribution rooms according to the present disclosure;

[0017] Figure 2 These are schematic diagrams illustrating the structure of some embodiments of the three-dimensional modeling apparatus for use in power distribution rooms according to the present disclosure;

[0018] Figure 3 This is a schematic diagram of the structure of an electronic device suitable for implementing some embodiments of the present disclosure. Detailed Implementation

[0019] Embodiments of this disclosure will now be described in more detail with reference to the accompanying drawings. While some embodiments of this disclosure are shown in the drawings, it should be understood that this disclosure can be implemented in various forms and should not be construed as limited to the embodiments set forth herein. Rather, these embodiments are provided to provide a more thorough and complete understanding of this disclosure. It should be understood that the accompanying drawings and embodiments of this disclosure are for illustrative purposes only and are not intended to limit the scope of protection of this disclosure.

[0020] It should also be noted that, for ease of description, only the parts relevant to the invention are shown in the accompanying drawings. Unless otherwise specified, the embodiments and features described in this disclosure can be combined with each other.

[0021] It should be noted that the concepts of "first" and "second" mentioned in this disclosure are used only to distinguish different devices, modules or units, and are not used to limit the order of functions performed by these devices, modules or units or their interdependencies.

[0022] It should be noted that the terms "a" and "a plurality of" used in this disclosure are illustrative rather than restrictive, and those skilled in the art should understand that, unless otherwise expressly indicated in the context, they should be understood as "one or more".

[0023] The names of messages or information exchanged between multiple devices in the embodiments of this disclosure are for illustrative purposes only and are not intended to limit the scope of such messages or information.

[0024] This disclosure will now be described in detail with reference to the accompanying drawings and embodiments.

[0025] refer to Figure 1 The diagram illustrates a flow 100 of some embodiments of a three-dimensional modeling method for a power distribution room according to the present disclosure. This three-dimensional modeling method for a power distribution room includes the following steps:

[0026] Step 101: Obtain the target image set.

[0027] In some embodiments, the entity executing the 3D modeling method for a power distribution room (e.g., a computing device) can acquire the aforementioned target image set via a wired or wireless connection. The target image set consists of pre-acquired wide-angle images of the power distribution room to be 3D modeled. Specifically, the target image set is a collection of images captured by a camera with a wide-angle lens, containing a panoramic view of the power distribution room to be 3D modeled. The power distribution room to be 3D modeled can be any power distribution room from which a corresponding 3D model is generated.

[0028] It should be noted that the aforementioned wireless connection methods may include, but are not limited to, 3G / 4G / 5G connections, WiFi connections, Bluetooth connections, WiMAX connections, Zigbee connections, UWB (ultra wideband) connections, and other currently known or future wireless connection methods.

[0029] It should be noted that the aforementioned computing devices can be either hardware or software. When the computing device is hardware, it can be implemented as a distributed cluster consisting of multiple servers or terminal devices, or as a single server or terminal device. When the computing device is software, it can be installed within the hardware devices listed above. It can be implemented as, for example, multiple software programs or software modules used to provide distributed services, or as a single software program or software module. No specific limitations are made here. It should be understood that the number of computing devices can be arbitrary, depending on the implementation requirements.

[0030] Step 102: Generate an image set with improved image quality based on the target image set.

[0031] In some embodiments, the execution entity can generate a set of images with improved image quality based on a set of target images. The improved images are obtained by locally improving the image quality of at least one target image. In practice, when acquiring target images, there are often overlapping areas between the acquired target images; therefore, a set of images with improved image quality can be generated based on the set of target images. Specifically, the execution entity can generate the improved images by image stacking.

[0032] As an example, the target image set may include: target image A, target image B, and target image C. Target image A and target image B have an overlapping region A. Target image B and target image C also have an overlapping region B. Therefore, for target image A and target image B, the aforementioned execution entity can stitch target image A and target image B together, and average the pixel values ​​of the pixels in target image A and target image B within the overlapping region A to obtain image A with improved image quality. Similarly, the aforementioned execution entity can stitch target image B and target image C together, and average the pixel values ​​of the pixels in target image B and target image C within the overlapping region B to obtain image B with improved image quality.

[0033] In some optional implementations of certain embodiments, the process by which the execution entity generates an enhanced image set based on the target image set may include the following steps:

[0034] The first step is to perform the following image preprocessing steps for each target image in the above target image set:

[0035] The first sub-step involves performing distortion calibration on the target image to generate a distortion-corrected image.

[0036] In practice, since the target image is a wide-angle image, there is a certain degree of distortion. Therefore, the aforementioned execution subject can use geometric correction to calibrate the distortion of the target image in order to generate a distortion-corrected image.

[0037] The second sub-step involves performing target detection on the above-mentioned distortion-corrected image to generate a set of target object description information.

[0038] The target object description information set includes a set of object feature points. In practice, the target object, as an important part of the image, is crucial for subsequent 3D modeling. For example, the target object could be a transformer; the accuracy of transformer modeling affects the usability of the subsequently generated updated 3D model of the power distribution room. Furthermore, the image alignment method based on the target object is faster than the method of comparing all pixels. In addition, considering that target detection is only one sub-step in image preprocessing, to improve target detection efficiency, this disclosure uses the NanoDet model to perform target detection on the above-mentioned distortion-corrected image to generate the target object description information set.

[0039] The second step is to perform image similarity matching on the target images in the above target image set based on the target object description information set corresponding to the target image, so as to generate a target image group set.

[0040] In practice, the aforementioned execution entity can use feature point matching to determine the feature point matching degree of the object feature point set included in the target object description information corresponding to the target image, thereby using the matched target images as target image groups to obtain a set of target image groups.

[0041] As an example, the target object description information set corresponding to target image A includes: target object description information A1 and target object description information A2. Specifically, target object description information A1 includes the set of object feature points A1. 11 and object feature point A 12 The target object description information A2 includes the set of object feature points, which includes: object feature points A. 21 and object feature point A 22 The target object description information set corresponding to target image B includes: target object description information B1 and target object description information B2. Among them, the target object description information B1 includes the set of object feature points B1. 11 and object feature point B 12 The target object description information B2 includes the set of object feature points, which includes: object feature points B. 21 and object feature point B 22 Among them, object feature point A 11 and object feature point A 12 Each and the object feature point B 11 and object feature point B 12 Matching. Object feature point A 21 and object feature point A 22 Each and the object feature point B 21 and object feature point B 22 Matching. Therefore, target image A and target image B constitute the target image group.

[0042] Third, for each target image group in the above target image group set, perform the following image quality enhancement steps:

[0043] The first sub-step involves randomly selecting one target image from the aforementioned target image group as the base image.

[0044] As an example, the target image group may include target image A and target image B. The aforementioned execution entity may select target image A as the base image.

[0045] The second sub-step involves performing local image quality enhancement on the base image based on the target images in the target image group other than the base image, in order to generate the image quality enhancement images in the image quality enhancement image set.

[0046] In practice, the aforementioned execution entity may simply overlay the local regions containing the object feature points of the target images (excluding the base image) in the target image group onto the base image and perform pixel value averaging to generate an image with improved image quality.

[0047] Step 103: Perform image stitching on each image in the image set after image quality enhancement to generate a stitched image.

[0048] In some embodiments, the execution entity can stitch together the individual image-enhanced images in the image-enhanced image set to generate a stitched image. In practice, the execution entity can stitch together the image-enhanced images sequentially according to the image acquisition order to obtain the stitched image.

[0049] In some optional implementations of certain embodiments, the execution entity performs image stitching on the individual image-upgraded images in the image-upgraded image set to generate a stitched image, which may include the following steps:

[0050] The first step is to extract feature points from the horizontal edge regions of each image in the above image enhancement set to generate edge region features.

[0051] The horizontal edge region refers to the rectangular areas on both sides of the image in the image after image quality enhancement. In practice, due to the acquisition order of the target image, global feature point extraction involves a large amount of data processing and does not significantly improve the accuracy of image stitching. Therefore, this disclosure only extracts feature points from the horizontal edge region of the image after image quality enhancement to generate edge region features. In practice, the aforementioned execution entity can use the SIFT (Scale-Invariant Feature Transform) algorithm to extract feature points from the horizontal edge region of the image after image quality enhancement to generate edge region features.

[0052] The second step involves performing image matching on the enhanced images in the aforementioned image enhancement image set based on the edge region features corresponding to the enhanced images, thereby generating a stitching order index array. This stitching order index array represents the image stitching order of each enhanced image in the aforementioned image enhancement image set. In practice, the executing entity can perform image matching on the enhanced images in the aforementioned image enhancement image set using feature point matching.

[0053] As an example, the set of images after image quality enhancement may include: image A, image B, image C, and image D. Image A corresponds to index A. Image B corresponds to index B. Image C corresponds to index C. Image D corresponds to index D. The initial stitching order index array can be [index A, index B, index C, index D]. First, for image A, the edge region features corresponding to image A are sequentially matched with the edge region features corresponding to image B, image C, and image D. For example, matching with the edge region features corresponding to image C. Therefore, the initial stitching order index array after the first adjustment can be [index A, index C, index B, index D]. Next, the edge region features corresponding to the image C after image quality enhancement are matched with the edge region features corresponding to the image B and the image D after image quality enhancement, respectively. For example, the edge region features corresponding to the image C after image quality enhancement are matched. Therefore, the initial stitching order index array after the second adjustment can be [index A, index C, index D, index B]. Then, the edge region features corresponding to the image D after image quality enhancement are matched with the edge region features corresponding to the image B after image quality enhancement. For example, the edge region features corresponding to the image D after image quality enhancement are matched with the edge region features corresponding to the image B after image quality enhancement. Therefore, [index A, index C, index D, index B] is the stitching order index array.

[0054] The third step involves stitching the images in the image quality enhancement set according to the stitching order index array to generate the stitched image.

[0055] As an example, the aforementioned execution entity can stitch the image A corresponding to index A to the left of the image C corresponding to index C, stitch the image C corresponding to index C to the left of the image D corresponding to index D, and stitch the image D corresponding to index D to the left of the image B corresponding to index B.

[0056] Step 104: Generate a model based on the stitched image and the pre-trained initial 3D model, generating an initial 3D model of the power distribution room and model description information.

[0057] In some embodiments, the aforementioned execution entity can generate a model based on the stitched image and a pre-trained initial 3D model, generating an initial 3D model of the power distribution room and model description information. The model description information includes a set of model element description information. The model element description information includes: element type, element feature vector, and element position information. The element position information represents the position of the model element corresponding to the model element description information in the initial 3D model of the power distribution room. The element type represents the element category of the model element. The element feature vector can be a one-dimensional feature vector for the model element.

[0058] Optionally, the initial 3D model generation model includes: an image encoder, an image decoder, a feature refinement model, and a threshold adjustment model. The image encoder includes: a first 2D convolutional layer, a first residual convolutional block, a second residual convolutional block, a third residual convolutional block, a fourth residual convolutional block, a fifth residual convolutional block, a sixth residual convolutional block, a seventh residual convolutional block, a second 2D convolutional layer, a third 2D convolutional layer, and a fourth 2D convolutional layer. A max-pooling layer is placed between the first 2D convolutional layer and the first residual convolutional block. A max-pooling layer is placed between the second 2D convolutional layer and the third 2D convolutional layer. The number of channels in the first residual convolutional block is (64, 64, 256). The number of channels in the second residual convolutional block is (64, 64, 256). The number of channels in the third residual convolutional block is (64, 64, 256). The fourth residual convolutional block has (128, 128, 256) channels. The fifth residual convolutional block has (128, 128, 256) channels. The sixth residual convolutional block has (128, 128, 256) channels. The seventh residual convolutional block has (128, 128, 256) channels. The decoder includes five serially connected 3D deconvolutional layers. Specifically, the five serially connected 3D deconvolutional layers include: a first 3D deconvolutional layer, a second 3D deconvolutional layer, a third 3D deconvolutional layer, a fourth 3D deconvolutional layer, and a fifth 3D deconvolutional layer. The kernel size of the first, second, third, and fourth 3D deconvolutional layers is 4. 3 The kernel size of the fifth 3D deconvolution layer is 1. 3 The first 3D deconvolution layer has 512 channels. The second 3D deconvolution layer has 128 channels. The third 3D deconvolution layer has 32 channels. The fourth 3D deconvolution layer has 8 channels. The fifth 3D deconvolution layer has 1 channel. The feature refinement model includes: a first 3D convolution layer, a second 3D convolution layer, a third 3D convolution layer, a first fully connected layer, a second fully connected layer, a first attention block, a second attention block, and a third attention block. The kernel size of the first, second, and third 3D convolution layers is 4. 3The first 3D convolutional layer has 32 channels. The second 3D convolutional layer has 64 channels. The third 3D convolutional layer has 128 channels. Each of the first, second, and third attention blocks includes an attention mechanism layer and a 3D deconvolutional layer. Specifically, the first attention block includes a 3D deconvolutional layer with 64 channels and a kernel size of 4. 3 The second attention block includes a 3D deconvolution layer with 32 channels and a kernel size of 4. 3 The third attention block includes a 3D deconvolution layer with 1 channel and a kernel size of 4. 3 The threshold adjustment model can use YOLO as the backbone network. Specifically, the threshold adjustment model extracts image features from the stitched image, converts them into feature maps with three different feature dimensions, and superimposes them. It also predicts the class probability of the target object bounded by the anchor box and adjusts the threshold in an adaptive manner to selectively adjust the threshold during 3D modeling.

[0059] In some optional implementations of certain embodiments, the execution entity generates an initial 3D model of the power distribution room and model description information based on the stitched image and the pre-trained initial 3D model, which may include the following steps:

[0060] The first step is to generate image-encoded features based on the image encoder and the stitched image described above.

[0061] The second step is to generate image decoded features based on the image decoder and the image encoded features described above.

[0062] The third step is to generate refined image features based on the feature refinement model and the image decoded features described above.

[0063] The fourth step is to generate threshold adjustment parameters based on the threshold adjustment model and the stitched image described above.

[0064] The fifth step involves adjusting the parameters based on the aforementioned thresholds and refining the image features to generate the initial 3D model of the power distribution room and the model description information.

[0065] Specifically, the element type can be the category of the target object bounded by the predicted anchor box in the threshold adjustment model. The element feature vector can be the one-dimensional feature vector corresponding to the target object bounded by the predicted anchor box. The element position information can characterize the position of the target object after 3D reconstruction in the initial 3D model of the power distribution room. Specifically, the element position information can be determined by projection.

[0066] The aforementioned phrases "optionally" and "in some optional implementations of some embodiments," as an inventive point of this disclosure, address the second technical problem mentioned in the background art, namely, "the accuracy of the 3D model obtained by image-based modeling is poor." Based on this, firstly, this disclosure adds a feature refinement model to the encoder-decoder structure to better extract the shape, texture, and spatial relationships of the 3D object. Simultaneously, an attention mechanism is introduced into the feature refinement model to improve reconstruction quality and detail restoration capabilities. Next, a threshold adjustment model is introduced so that 3D reconstruction can be selectively performed by controlling the threshold during reconstruction, thereby significantly improving the accuracy of the 3D model obtained through modeling.

[0067] Step 105: For each model element description in the model element description information set, perform the following processing steps:

[0068] Step 1051: Based on the element type and element feature vector included in the model element description information, determine whether there are model elements in the pre-built model element library that match the model element description information.

[0069] In some embodiments, the execution entity can determine whether a model element matching the model element description information exists in a pre-built model element library based on the element type and element feature vector included in the model element description information. The model element library can be a database storing pre-built and updated model elements. The model element can be a 3D model of a power equipment. In practice, the execution entity can determine whether a model element matching the model element description information exists in the pre-built model element library by querying the database based on the element type and element feature vector included in the model element description information.

[0070] Optionally, each model element in the model element library has a corresponding hash identifier, which is a 256-bit hash string. Specifically, the hash identifier can be obtained by concatenating the element type and the element feature vector of the model element and then performing hash processing.

[0071] In some optional implementations of certain embodiments, determining whether a model element matching the model element description information exists in the pre-built model element library based on the element type and element feature vector included in the model element description information may include the following steps:

[0072] The first step is to perform one-hot encoding on the element types included in the above model element description information to generate encoded element types.

[0073] In practice, since the number of element types is fixed, the length of the element type vector after one-hot encoding is fixed.

[0074] The second step is to concatenate the encoded element type with the element feature vector included in the model element description information to generate a concatenated vector.

[0075] The third step is to perform hashing on the concatenated vector to generate a hashed vector.

[0076] The hashed vector is a 256-bit hash string.

[0077] The fourth step is to match the hashed vector with the hash identifiers corresponding to the model elements in the model element library to determine whether there are model elements in the model element library that match the model element description information.

[0078] Specifically, the aforementioned execution entity can traverse the model element library to determine whether there are model elements in the model element library that have the same hash identifier and hashed vector, thereby determining whether there are model elements in the model element library that match the aforementioned model element description information.

[0079] Step 1052: In response to existence, model elements that match the model element description information are identified as candidate model elements.

[0080] In some embodiments, in response to the existence of a model element, the execution entity determines the model element that matches the model element description information as a candidate model element.

[0081] Step 106: Based on the obtained candidate model element set, update the model elements of the initial power distribution room 3D model to generate the updated power distribution room 3D model.

[0082] In some embodiments, the aforementioned execution entity updates the model elements of the initial three-dimensional power distribution room model based on the obtained candidate model element set to generate an updated three-dimensional power distribution room model.

[0083] In some optional implementations of certain embodiments, updating the model elements of the initial 3D power distribution room model based on the obtained candidate model element set to generate an updated 3D power distribution room model may include the following steps:

[0084] For each candidate model element in the above candidate model element set, replace the model element in the above initial power distribution room 3D model that corresponds to the above candidate model element with the above candidate model element.

[0085] Optionally, the above method further includes:

[0086] The first step is to associate the real-time operating parameters of the power distribution room equipment in the power distribution room to be 3D modeled with the updated 3D model of the power distribution room.

[0087] In practice, in response to the successful modeling of the updated 3D power distribution room, the aforementioned execution entity can monitor the real-time operating parameters of the power distribution room equipment through automatic data point embedding, and synchronize the real-time operating parameters with the updated 3D power distribution room model in real time to achieve parameter association.

[0088] The second step, in response to the successful parameter association, is to display the equipment operation status of the power distribution room equipment in the power distribution room to be 3D modeled in real time through the updated 3D model of the power distribution room.

[0089] The various embodiments of this disclosure have the following beneficial effects: the 3D modeling method for power distribution rooms according to some embodiments of this disclosure shortens the modeling cycle and reduces the consumption of computing resources. Specifically, the reason for the long modeling cycle is that the point cloud modeling method involves a huge amount of data processing. Based on this, the 3D modeling method for power distribution rooms according to some embodiments of this disclosure first acquires a target image set, wherein the target image set is a pre-acquired wide-angle image of the power distribution room to be 3D modeled. Compared with the point cloud modeling method, the image-based 3D modeling method has the advantage of convenient image acquisition. Secondly, based on the target image set, an image set with improved image quality is generated, wherein the image with improved image quality is an image obtained by locally improving the image quality of at least one target image. The image quality improvement method ensures the accuracy of the subsequently generated 3D model. Next, the images in the image set with improved image quality are stitched together to generate a stitched image. Furthermore, based on the stitched images and the pre-trained initial 3D model, an initial 3D model of the power distribution room and model description information are generated. The model description information includes a set of model element description information, which includes element type, element feature vector, and element position information. The element position information represents the position of the model element corresponding to the model element description information within the initial 3D model of the power distribution room. This, combined with the image, generates an initial 3D model of the power distribution room to be 3D modeled. In addition, for each model element description information in the set, the following processing steps are performed: First, based on the element type and element feature vector included in the model element description information, it is determined whether a model element matching the model element description information exists in the pre-constructed model element library. Second, in response to the existence of a matching model element description information, the model element matching the model element description information is identified as a candidate model element. Finally, based on the obtained candidate model element set, the initial 3D model of the power distribution room is updated to generate an updated 3D model of the power distribution room. By using matching model element replacement, the model accuracy of the updated 3D model of the power distribution room is further improved. This method generates 3D models using only images, which shortens the modeling cycle and reduces computational resource consumption compared to point cloud modeling.

[0090] Further reference Figure 2 As an implementation of the methods shown in the above figures, this disclosure provides some embodiments of a three-dimensional modeling device applied to a power distribution room. These device embodiments are similar to... Figure 1 Corresponding to the method embodiments shown, this 3D modeling device for power distribution rooms can be specifically applied to various electronic devices.

[0091] like Figure 2As shown, a 3D modeling device 200 for a power distribution room in some embodiments includes: an acquisition unit 201, a first generation unit 202, an image stitching unit 203, a second generation unit 204, an execution unit 205, and a model element update unit 206. The acquisition unit 201 is configured to acquire a set of target images, wherein the set of target images is a pre-acquired wide-angle image of the power distribution room to be 3D modeled; the first generation unit 202 is configured to generate a set of image quality-enhanced images based on the set of target images, wherein the image quality-enhanced images are images obtained by locally enhancing the image quality of at least one target image; the image stitching unit 203 is configured to stitch together the image quality-enhanced images in the set of image quality-enhanced images to generate a stitched image; the second generation unit 204 is configured to generate an initial 3D model of the power distribution room and model description information based on the stitched image and a pre-trained initial 3D model, wherein the model description information includes: a set of model element description information, model element description information... The information includes: element type, element feature vector, and element position information. The element position information represents the position of the model element corresponding to the model element description information in the initial three-dimensional model of the power distribution room. The execution unit 205 is configured to perform the following processing steps for each model element description information in the above model element description information set: determine whether there is a model element matching the above model element description information in the pre-built model element library according to the element type and element feature vector included in the above model element description information; in response to the existence, determine the model element matching the above model element description information as a candidate model element; the model element update unit 206 is configured to update the model elements of the above initial three-dimensional model of the power distribution room according to the obtained candidate model element set to generate an updated three-dimensional model of the power distribution room.

[0092] It is understandable that the units described in the 3D modeling device 200 applied to the power distribution room are similar to the reference units. Figure 1 The steps in the described method correspond to each other. Therefore, the operations, features, and beneficial effects described above for the method are also applicable to the three-dimensional modeling device 200 and the units contained therein applied to the power distribution room, and will not be repeated here.

[0093] The following is for reference. Figure 3 It shows a schematic diagram of the structure of an electronic device (e.g., a computing device) 300 suitable for implementing some embodiments of the present disclosure. Figure 3 The electronic device shown is merely an example and should not be construed as limiting the functionality and scope of the embodiments of this disclosure.

[0094] like Figure 3As shown, the electronic device 300 may include a processing unit (e.g., a central processing unit, a graphics processing unit, etc.) 301, which can perform various appropriate actions and processes according to a program stored in a read-only memory 302 or a program loaded from a storage device 308 into a random access memory 303. The random access memory 303 also stores various programs and data required for the operation of the electronic device 300. The processing unit 301, the read-only memory 302, and the random access memory 303 are interconnected via a bus 304. An input / output interface 305 is also connected to the bus 304.

[0095] Typically, the following devices can be connected to I / O interface 305: input devices 306 including, for example, touchscreens, touchpads, keyboards, mice, cameras, microphones, accelerometers, gyroscopes, etc.; output devices 307 including, for example, liquid crystal displays (LCDs), speakers, vibrators, etc.; storage devices 308 including, for example, magnetic tapes, hard disks, etc.; and communication devices 309. Communication device 309 allows electronic device 300 to communicate wirelessly or wiredly with other devices to exchange data. Although Figure 3 An electronic device 300 with various devices is shown; however, it should be understood that it is not required to implement or possess all of the devices shown. More or fewer devices may be implemented or possessed alternatively. Figure 3 Each box shown can represent a device or multiple devices as needed.

[0096] In particular, according to some embodiments of this disclosure, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, some embodiments of this disclosure include a computer program product comprising a computer program carried on a computer-readable medium, the computer program containing program code for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via a communication device 309, or installed from a storage device 308, or installed from a read-only memory 302. When the computer program is executed by the processing device 301, it performs the functions defined in the methods of some embodiments of this disclosure.

[0097] It should be noted that, in some embodiments of this disclosure, the computer-readable medium may be a computer-readable signal medium or a computer-readable storage medium, or any combination thereof. A computer-readable storage medium may be, for example,—but not limited to—an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples of a computer-readable storage medium may include, but are not limited to: an electrical connection having one or more wires, a portable computer disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage device, magnetic storage device, or any suitable combination thereof. In some embodiments of this disclosure, a computer-readable storage medium may be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, apparatus, or device. In some embodiments of this disclosure, a computer-readable signal medium may include a data signal propagated in baseband or as part of a carrier wave, carrying computer-readable program code. Such propagated data signals may take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. A computer-readable signal medium can be any computer-readable medium other than a computer-readable storage medium, which can send, propagate, or transmit a program for use by or in connection with an instruction execution system, apparatus, or device. The program code contained on the computer-readable medium can be transmitted using any suitable medium, including but not limited to: wires, optical fibers, RF (radio frequency), etc., or any suitable combination thereof.

[0098] In some implementations, clients and servers can communicate using any currently known or future-developed network protocol such as HTTP (Hypertext Transfer Protocol) and can interconnect with digital data communication (e.g., communication networks) of any form or medium. Examples of communication networks include local area networks (“LANs”), wide area networks (“WANs”), the Internet (e.g., the Internet of Things), and peer-to-peer networks (e.g., ad hoc peer-to-peer networks), as well as any currently known or future-developed networks.

[0099] The aforementioned computer-readable medium may be included in the aforementioned electronic device; or it may exist independently without being assembled into the electronic device. The aforementioned computer-readable medium carries one or more programs that, when executed by the electronic device, cause the electronic device to: acquire a target image set, wherein the target image set is a pre-acquired wide-angle image of the power distribution room to be 3D modeled; generate an image set with improved image quality based on the target image set, wherein the image set with improved image quality is an image obtained by locally improving the image quality of at least one target image; stitch together the images in the image set with improved image quality to generate a stitched image; and generate an initial 3D model of the power distribution room and model description information based on the stitched image and a pre-trained initial 3D model generation model, wherein the model description information includes: model element description information. The set of model element description information includes: element type, element feature vector, and element position information. The element position information represents the position of the model element corresponding to the model element description information in the initial 3D model of the power distribution room. For each model element description information in the set of model element description information, the following processing steps are performed: Based on the element type and element feature vector included in the model element description information, determine whether there is a model element in the pre-constructed model element library that matches the model element description information; In response to the existence, determine the model element that matches the model element description information as a candidate model element; Based on the obtained candidate model element set, update the model elements of the initial 3D model of the power distribution room to generate an updated 3D model of the power distribution room.

[0100] Computer program code for performing operations of some embodiments of this disclosure can be written in one or more programming languages ​​or a combination thereof, including object-oriented programming languages ​​such as Java, Smalltalk, and C++, and conventional procedural programming languages ​​such as the "C" language or similar programming languages. The program code can be executed entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving remote computers, the remote computer can be connected to the user's computer via any type of network—including a local area network (LAN) or a wide area network (WAN)—or can be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0101] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of this disclosure. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, can be implemented using a dedicated hardware-based system that performs the specified function or operation, or using a combination of dedicated hardware and computer instructions.

[0102] The units described in some embodiments of this disclosure can be implemented in software or hardware. The described units can also be housed in a processor; for example, a processor may be described as including an acquisition unit, a first generation unit, an image stitching unit, a second generation unit, an execution unit, and a model element update unit. The names of these units do not necessarily limit the unit itself; for example, the execution unit may be described as "a unit that, for each model element description information in the aforementioned model element description information set, performs the following processing steps: determining whether a model element matching the aforementioned model element description information exists in a pre-built model element library based on the element type and element feature vector included in the aforementioned model element description information; and, in response to the existence of a matching model element, determining the matching model element as a candidate model element."

[0103] The functions described above in this document can be performed at least in part by one or more hardware logic components. For example, exemplary types of hardware logic components that can be used, without limitation, include: field-programmable gate arrays (FPGAs), application-specific integrated circuits (ASICs), application-specific standard products (ASSPs), system-on-a-chip (SoCs), complex programmable logic devices (CPLDs), and so on.

[0104] The above description is merely a selection of preferred embodiments of this disclosure and an explanation of the technical principles employed. Those skilled in the art should understand that the scope of the invention involved in the embodiments of this disclosure is not limited to technical solutions formed by specific combinations of the above-described technical features, but should also cover other technical solutions formed by arbitrary combinations of the above-described technical features or their equivalents without departing from the above-described inventive concept. For example, technical solutions formed by substituting the above-described features with (but not limited to) technical features with similar functions disclosed in the embodiments of this disclosure.

Claims

1. A three-dimensional modeling method for power distribution rooms, comprising: Acquire a target image set, wherein the target image set is a pre-acquired wide-angle image of the power distribution room to be 3D modeled; Based on the target image set, a set of images with improved image quality is generated, wherein the images with improved image quality are images obtained by locally improving the image quality of at least one target image. The images in the image set after image quality enhancement are stitched together to generate a stitched image. Based on the stitched image and the pre-trained initial 3D model, an initial 3D model of the power distribution room and model description information are generated. The model description information includes a set of model element description information, which includes element type, element feature vector, and element position information. The element position information represents the position of the model element corresponding to the model element description information in the initial 3D model of the power distribution room. For each model element description in the set of model element description information, perform the following processing steps: Based on the element type and element feature vector included in the model element description information, it is determined whether a model element matching the model element description information exists in a pre-built model element library. The model elements in the model element library correspond to hash identifiers, which are 256-bit hash strings. The process includes: performing one-hot encoding on the element type included in the model element description information to generate an encoded element type; concatenating the encoded element type with the element feature vector included in the model element description information to generate a concatenated vector; hashing the concatenated vector to generate a hashed vector, where the hashed vector is a 256-bit hash string; and matching the hashed vector with the hash identifier corresponding to the model element in the model element library to determine whether a model element matching the model element description information exists in the model element library. In response to existence, model elements that match the model element description information are identified as candidate model elements; Based on the obtained candidate model element set, the initial 3D model of the power distribution room is updated to generate an updated 3D model of the power distribution room.

2. The method according to claim 1, wherein, The method further includes: The real-time operating parameters of the power distribution room equipment in the power distribution room to be 3D modeled are associated with the updated 3D model of the power distribution room. In response to successful parameter association, the updated 3D model of the power distribution room displays the equipment operation status of the power distribution room equipment in the power distribution room to be 3D modeled in real time.

3. The method according to claim 2, wherein, The step of generating an enhanced image set based on the target image set includes: For each target image in the target image set, perform the following image preprocessing steps: The target image is subjected to distortion calibration to generate a distortion-corrected image; Target detection is performed on the distortion-corrected image to generate a target object description information set, wherein the target object description information set includes: a set of object feature points; Based on the target object description information set corresponding to the target image, image similarity matching is performed on the target images in the target image set to generate a target image group set; For each target image group in the target image group set, perform the following image quality enhancement steps: Randomly select one target image from the target image group as the base image; Based on the target images in the target image group other than the base image, the base image is locally enhanced to generate the enhanced images in the enhanced image set.

4. The method according to claim 3, wherein, The step of stitching together the individual enhanced images in the enhanced image set to generate a stitched image includes: For each image in the image enhancement set, feature points are extracted from the horizontal edge regions of the image enhancement to generate edge region features. Based on the edge region features corresponding to the image after image quality enhancement, image matching is performed on the image after image quality enhancement in the image after image quality enhancement set to generate a stitching order index array, wherein the stitching order index array represents the image stitching order of each image after image quality enhancement in the image after image quality enhancement set. Based on the stitching order index array, each image in the image quality enhancement set is stitched together to generate the stitched image.

5. The method according to claim 4, wherein, The step of updating the initial 3D model of the power distribution room based on the obtained candidate model element set to generate an updated 3D model of the power distribution room includes: For each candidate model element in the candidate model element set, the model element in the initial 3D model of the power distribution room that corresponds to the candidate model element is replaced with the candidate model element.

6. The method according to claim 5, wherein, The initial 3D model generation model includes: an image encoder, an image decoder, a feature refinement model, and a threshold adjustment model; and The step of generating an initial 3D model of the power distribution room and model description information based on the stitched image and the pre-trained initial 3D model includes: Based on the image encoder and the stitched image, generate image-encoded features; Based on the image decoder and the image encoded features, generate image decoded features; Based on the feature refinement model and the decoded image features, the refined image features are generated; Based on the threshold adjustment model and the stitched image, threshold adjustment parameters are generated; Based on the threshold adjustment parameters and the refined features of the image, the initial 3D model of the power distribution room and the model description information are generated.

7. A three-dimensional modeling device for use in a power distribution room, comprising: The acquisition unit is configured to acquire a target image set, wherein the target image set is a pre-acquired wide-angle image of the power distribution room to be 3D modeled; The first generation unit is configured to generate a set of images with improved image quality based on the target image set, wherein the images with improved image quality are images obtained by locally improving the image quality of at least one target image; The image stitching unit is configured to stitch together each of the image-enhanced images in the image-enhanced image set to generate a stitched image. The second generation unit is configured to generate a model based on the stitched image and the pre-trained initial 3D model, generating an initial 3D model of the power distribution room and model description information. The model description information includes a set of model element description information, which includes element type, element feature vector, and element position information. The element position information represents the position of the model element corresponding to the model element description information in the initial 3D model of the power distribution room. An execution unit is configured to perform the following processing steps for each model element description in the model element description information set: Based on the element type and element feature vector included in the model element description information, determine whether a model element matching the model element description information exists in a pre-built model element library, wherein the model elements in the model element library correspond to hash identifiers, and the hash identifiers are 256-bit hash strings. This includes: performing one-hot encoding on the element type included in the model element description information to generate an encoded element type; concatenating the encoded element type with the element feature vector included in the model element description information to generate a concatenated vector; performing hash processing on the concatenated vector to generate a hashed vector, wherein the hashed vector is a 256-bit hash string; matching the hashed vector with the hash identifier corresponding to the model element in the model element library to determine whether a model element matching the model element description information exists in the model element library; and, in response to the existence of a matching model element, identifying the model element matching the model element description information as a candidate model element. The model element update unit is configured to update the model elements of the initial three-dimensional model of the power distribution room based on the obtained set of candidate model elements, so as to generate an updated three-dimensional model of the power distribution room.

8. An electronic device, comprising: One or more processors; A storage device on which one or more programs are stored; When the one or more programs are executed by the one or more processors, the one or more processors implement the method as described in any one of claims 1 to 6.

9. A computer-readable medium having a computer program stored thereon, wherein, When the computer program is executed by a processor, it implements the method as described in any one of claims 1 to 6.

Images