Photovoltaic base construction site construction progress monitoring method based on YOLOv8

Abstract

The invention discloses a construction progress monitoring method for a photovoltaic base construction site based on YOLOv8, and the method comprises the steps: dividing the total construction amount into a plurality of construction stages in order, and collecting aerial images of key components of each construction stage to form a data set; a Context Aggregation module is introduced to optimize a target detection algorithm of the YOLOv8, and an improved model used for increasing the detection precision is formed; using the data set to train the improved model; carrying out aerial photography on the construction site by using an unmanned aerial vehicle; analyzing the aerial video of the construction site by using the trained improved model, and counting the types and the number of key components in the aerial video; the current construction stage is judged based on the counted types of the key components and the sequential relation of all the construction stages, and the construction percentage of the current construction stage is determined based on the counted number of the key components of the current construction stage and the designed number of the key components of the current construction stage; through the monitoring method, the monitoring precision is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of building construction, and in particular to a construction progress monitoring method for a photovoltaic base construction site based on YOLOv8. Background Art

[0002] In the context of vigorously developing a new power system with new energy as the main energy source, photovoltaic power generation, as an important part of clean energy, has received strong support and encouragement. The scale of photovoltaic power generation base stations will further expand, and the number of construction will continue to increase. Construction progress control has become more and more complicated and difficult, including many construction parties, complex and changeable environment, large amount of data and other problems, resulting in inefficient construction progress management, prone to delays and quality problems.

[0003] Traditional construction progress monitoring requires the construction party to manually report the daily construction progress to the system, which is a time-consuming, costly, and low-automation labor-intensive task. There is also an existing technology that monitors the construction progress through drones, but the monitoring accuracy of traditional drone monitoring technology is poor. In addition, since photovoltaic power base stations are generally built in a sand environment, the components and the environment are easily confused in this environment, and traditional drone monitoring technology may result in misjudgment and missed judgments. Summary of the invention

[0004] The purpose of the present invention is to overcome the defects of the prior art and provide a construction progress monitoring method for a photovoltaic base construction site based on YOLOv8, which can monitor the construction accuracy of photovoltaic power generation base stations in a complex Shago environment with high automation and high monitoring accuracy.

[0005] To achieve the above purpose, the technical solution adopted by the present invention is a construction progress monitoring method for a photovoltaic base construction site based on YOLOv8, comprising the steps of:

[0006] Step 1: Divide the total construction volume into multiple construction stages in sequence, determine the key components based on the characteristics of each construction stage, and collect aerial images of the key components in each construction stage to form a data set;

[0007] Step 2: Introduce the Context Aggregation module to optimize the YOLOv8 target detection algorithm to form an improved model for increasing detection accuracy;

[0008] Step 3: Using the data set to train the improved model;

[0009] Step 4: Use a drone to take aerial photos of the construction site;

[0010] Step 5: Analyze the aerial video of the construction site using the trained improved model, and count the types and quantities of key components in the aerial video;

[0011] Step 6: Determine the current construction stage based on the statistical types of key components and the chronological order of each construction stage, and determine the construction percentage of the current construction stage based on the statistical number of key components in the current construction stage and the designed number of key components in the current construction stage.

[0012] The construction progress monitoring method for the photovoltaic base construction site based on YOLOv8 of the present invention is further improved in that, when executing the above step 1, the construction stage includes the pile foundation driving stage, the photovoltaic bracket construction stage and the photovoltaic panel laying stage, the key component of the pile foundation driving stage is the pile foundation structure, the key component of the photovoltaic bracket construction stage is the photovoltaic bracket, and the key component of the photovoltaic panel laying stage is the photovoltaic panel.

[0013] A further improvement of the construction progress monitoring method for a photovoltaic base construction site based on YOLOv8 of the present invention is that, when executing the above step 1, the Polygen tool in the Labelme software is used to annotate key components in the aerial image to form the data set.

[0014] The construction progress monitoring method of the photovoltaic base construction site based on YOLOv8 of the present invention is further improved in that the Context Aggregation module includes:

[0015] A first unit for enhancing detection accuracy of a key component, the first unit comprising:

[0016] A first branch for suppressing environmental background interference, wherein the first branch generates an attention map through an attention mechanism;

[0017] A second branch for capturing shape information of key components in an image, performing a convolution operation on an input image through a convolutional neural network to output a feature map, and then fusing the feature map with the attention map to form an optimized feature map, wherein the second branch is connected to the first branch;

[0018] A second unit is used to associate context information related to the key component to form a context graph and then fuse the context graph with the optimized feature graph to form a context refinement graph, the second unit being connected to the first unit.

[0019] A further improvement of the construction progress monitoring method for a photovoltaic base construction site based on YOLOv8 of the present invention is that, when executing the above step 3, the data set is divided into a training set for training the improved model, a validation set for evaluating the training effect, and a test set for evaluating the performance of the improved model in a ratio of 7:2:1.

[0020] The construction progress monitoring method of the photovoltaic base construction site based on YOLOv8 of the present invention is further improved in that a first algorithm for judging the current construction stage is written into the improved model, and when executing the above step 6:

[0021] When the key component type in the statistical result is only the pile foundation structure, the first algorithm determines that the current construction stage is the pile foundation driving stage;

[0022] When the key component types in the statistical results include pile foundation structures and photovoltaic brackets, the first algorithm determines that the current construction stage is the photovoltaic bracket construction stage;

[0023] When the key component types in the statistical results include photovoltaic brackets and photovoltaic panels, the first algorithm determines that the current construction stage is the photovoltaic panel laying stage.

[0024] A further improvement of the construction progress monitoring method for a photovoltaic base construction site based on YOLOv8 of the present invention is that a second algorithm for calculating the construction progress percentage of the current construction stage is written into the improved model. When executing the above step 6, the second algorithm divides the actual number of key components in the current construction stage in the statistical results by the designed number of key components in the current construction stage.

[0025] Compared with the prior art, the advantages of the present invention are:

[0026] 1. By introducing the Context Aggregation module to optimize the target detection algorithm of YOLOv8, it can suppress background interference in complex environments and improve monitoring accuracy.

[0027] 2. Through the first algorithm and the second algorithm, the current construction stage and the construction progress percentage of the current construction stage can be automatically confirmed, with a high degree of automation. BRIEF DESCRIPTION OF THE DRAWINGS

[0028] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0029] Figure 1 It is a flow chart of the present invention.

[0030] Figure 2 It is a working schematic diagram of the Context Aggregation module of the present invention.

[0031] Figure 3This is a performance evaluation comparison table of the improved model of the present invention and the existing detection model. DETAILED DESCRIPTION

[0032] The following describes the embodiments of the present invention through specific examples, and those skilled in the art can easily understand other advantages and effects of the present invention from the contents disclosed in this specification. The present invention can also be implemented or applied through other different specific embodiments, and the details in this specification can also be modified or changed in various ways based on different viewpoints and applications without departing from the spirit of the present invention.

[0033] The following is a further detailed description of the construction progress monitoring method for a photovoltaic base construction site based on YOLOv8 in conjunction with the accompanying drawings and specific embodiments.

[0034] See also Figure 1 to Figure 3 As shown, a construction progress monitoring method for a photovoltaic base construction site based on YOLOv8 (a target detection and image segmentation model developed by Ultralytics) includes the following steps:

[0035] Step 1: Divide the total construction volume into multiple construction stages in sequence, determine the key components based on the characteristics of each construction stage, and collect aerial images of the key components in each construction stage to form a data set;

[0036] Step 2: Introduce the Context Aggregation module to optimize the YOLOv8 target detection algorithm and form an improved model for increasing detection accuracy;

[0037] Step 3: Use the data set to train the improved model;

[0038] Step 4: Use a drone to take aerial photos of the construction site;

[0039] Step 5: Use the trained improved model to analyze the aerial video of the construction site and count the types and quantities of key components in the aerial video;

[0040] Step 6: Determine the current construction stage based on the statistical types of key components and the chronological order of each construction stage, and determine the construction percentage of the current construction stage based on the statistical number of key components in the current construction stage and the designed number of key components in the current construction stage.

[0041] Specifically, the Context Aggregation module is added to the YOLOv8 image feature pyramid SPPF module.

[0042] Preferably, when executing the above step 1, the construction stage includes the pile foundation driving stage, the photovoltaic bracket construction stage and the photovoltaic panel laying stage. The key component of the pile foundation driving stage is the pile foundation structure, the key component of the photovoltaic bracket construction stage is the photovoltaic bracket, and the key component of the photovoltaic panel laying stage is the photovoltaic panel.

[0043] Preferably, when executing the above step 1, the key components in the aerial image are annotated using the Polygen tool in Labelme software (a graphical image annotation software) to form the data set.

[0044] Preferably, the Context Aggregation module includes:

[0045] A first unit for enhancing detection accuracy of a key component, the first unit comprising:

[0046] A first branch for suppressing environmental background interference, wherein the first branch generates an attention map through an attention mechanism;

[0047] A second branch for capturing shape information of key components in an image, performing a convolution operation on the input image through a convolutional neural network to output a feature map, and then fusing the feature map with the attention map to form an optimized feature map, wherein the second branch is connected to the first branch;

[0048] A second unit is used for associating context information related to the key component to form a context graph and then fusing the context graph with the optimized feature graph to form a context refinement graph, the second unit being connected to the first unit.

[0049] The attention map indicates which areas in the image are most important for the current target detection task. In scenes such as deserts, the features of the background will be similar to the texture, shape and other features of the components, and the attention map can help the improved model focus on specific areas of the image, thereby suppressing background interference.

[0050] The feature map contains local features in the image (such as edges, corners, textures, etc.) and is an information representation of the image from low to high levels.

[0051] The context graph refers to additional information related to the target or area of ​​interest. It not only involves the features directly presented in the image, but also includes the spatial and semantic relationships between the target object and other elements. This information helps the improved model understand the layout of multiple key components and the overall scene, thereby improving the recognition accuracy of the target.

[0052] Specifically, the feature map and the attention map are integrated through a matrix multiplication operation to form the optimized feature map, and the optimized feature map and the context map are combined through an element-level addition operation to form the context refinement map.

[0053] The Context Aggregation module accurately evaluates and optimizes the information content of image pixels, effectively balancing the retention of global context information and local features. This unique processing strategy significantly reduces the dependence on feature aggregation at other spatial locations in the image, providing a more efficient and accurate target recognition method for image analysis.

[0054] Preferably, when executing the above step 3, the data set is divided into a training set for training the improved model, a validation set for evaluating the training effect, and a test set for evaluating the performance of the improved model in a ratio of 7:2:1.

[0055] Specifically, when using the validation set for performance evaluation, multiple existing detection models are evaluated together with the improved model under the same conditions.

[0056] Specifically, the improved model is named Photovoltaic Net.

[0057] like Figure 3 As shown in the figure, the performance scores of four detection models including the improved model are compared on two important evaluation indicators: mAP@50 and mAP@50:95.

[0058] The improved model, Photovoltaic Net, scored 0.826 and 0.389 in mAP@50 and mAP@50:95 respectively. The score of the improved model in mAP@50 is 1.6% higher than that of YOLOv8 without the introduction of the Context Aggregation module, and the score in mAP@50:95 is the highest among the four detection models.

[0059] By setting the mAP@50 and mAP@50:95 to jointly measure the accuracy of the detection model, a quantitative basis is provided for model optimization.

[0060] Preferably, the improved model has a first algorithm for determining the current construction stage written therein, and when executing the above step 6:

[0061] When the key component type in the statistical result is only the pile foundation structure, the first algorithm determines that the current construction stage is the pile foundation driving stage;

[0062] When the key component types in the statistical results include pile foundation structures and photovoltaic brackets, the first algorithm determines that the current construction stage is the photovoltaic bracket construction stage;

[0063] When the key component types in the statistical results include photovoltaic brackets and photovoltaic panels, the first algorithm determines that the current construction stage is the photovoltaic panel laying stage.

[0064] By setting the first algorithm, it is convenient to automatically determine the current construction stage.

[0065] Preferably, a second algorithm for calculating the construction progress percentage of the current construction phase is written into the improved model. When executing the above step 6, the second algorithm divides the actual number of key components of the current construction phase in the statistical result by the designed number of key components of the current construction phase.

[0066] By setting the second algorithm, it is convenient to calculate the construction progress of the current construction stage.

[0067] It should be noted that the structures, proportions, sizes, etc. illustrated in the drawings of this specification are only used to match the contents disclosed in the specification for people familiar with this technology to understand and read, and are not used to limit the limiting conditions for the implementation of the present invention, so they have no substantial technical significance. Any modification of the structure, change of the proportion relationship or adjustment of the size should still fall within the scope of the technical content disclosed by the present invention without affecting the effects and purposes that can be achieved by the present invention. At the same time, the terms such as "upper", "lower", "left", "right", "middle" and "one" quoted in this specification are only for the convenience of description, and are not used to limit the scope of the implementation of the present invention. The change or adjustment of their relative relationship should also be regarded as the scope of the implementation of the present invention without substantially changing the technical content.

[0068] The above is only a preferred embodiment of the present invention, and does not limit the present invention in any form. Although the present invention has been disclosed as a preferred embodiment as above, it is not used to limit the present invention. Any technician familiar with this profession can make some changes or modifications to equivalent embodiments of equivalent changes using the technical contents disclosed above without departing from the scope of the technical solution of the present invention. However, any simple modification, equivalent change and modification made to the above embodiments based on the technical essence of the present invention without departing from the content of the technical solution of the present invention still fall within the scope of the technical solution of the present invention.

Claims

1. A construction progress monitoring method for a photovoltaic base construction site based on YOLOv8, characterized in that: Includes steps: Step 1: Divide the total construction volume into multiple construction stages in sequence, determine the key components based on the characteristics of each construction stage, and collect aerial images of the key components in each construction stage to form a data set; Step 2: Introduce the Context Aggregation module to optimize the YOLOv8 target detection algorithm to form an improved model for increasing detection accuracy; Step 3: Using the data set to train the improved model; Step 4: Use a drone to take aerial photos of the construction site; Step 5: Analyze the aerial video of the construction site using the trained improved model, and count the types and quantities of key components in the aerial video; Step 6: Determine the current construction stage based on the statistical types of key components and the chronological order of each construction stage, and determine the construction percentage of the current construction stage based on the statistical number of key components in the current construction stage and the designed number of key components in the current construction stage.

2. The construction progress monitoring method for a photovoltaic base construction site based on YOLOv8 as claimed in claim 1, characterized in that: When executing the above step 1, the construction stage includes the pile foundation driving stage, the photovoltaic bracket construction stage and the photovoltaic panel laying stage. The key component of the pile foundation driving stage is the pile foundation structure, the key component of the photovoltaic bracket construction stage is the photovoltaic bracket, and the key component of the photovoltaic panel laying stage is the photovoltaic panel.

3. The construction progress monitoring method for a photovoltaic base construction site based on YOLOv8 as claimed in claim 1, characterized in that: When executing the above step 1, the key components in the aerial image are annotated using the Polygen tool in the Labelme software to form the data set.

4. The construction progress monitoring method for a photovoltaic base construction site based on YOLOv8 as claimed in claim 1, characterized in that: The Context Aggregation module includes: A first unit for enhancing detection accuracy of a key component, the first unit comprising: A first branch for suppressing environmental background interference, wherein the first branch generates an attention map through an attention mechanism; A second branch for capturing shape information of key components in an image, performing a convolution operation on an input image through a convolutional neural network to output a feature map, and then fusing the feature map with the attention map to form an optimized feature map, wherein the second branch is connected to the first branch; A second unit is used for associating context information related to the key component to form a context graph, and then fusing the context graph with the optimized feature graph to form a context refinement graph, wherein the second unit is connected to the first unit.

5. The construction progress monitoring method for a photovoltaic base construction site based on YOLOv8 as claimed in claim 1, characterized in that: When executing the above step 3, the data set is divided into a training set for training the improved model, a validation set for evaluating the training effect, and a test set for evaluating the performance of the improved model in a ratio of 7:2:

1.

6. The construction progress monitoring method for a photovoltaic base construction site based on YOLOv8 as claimed in claim 2, characterized in that: The improved model has a first algorithm for determining the current construction stage written therein. When executing step 6 above: When the key component type in the statistical result is only the pile foundation structure, the first algorithm determines that the current construction stage is the pile foundation driving stage; When the key component types in the statistical results include pile foundation structures and photovoltaic brackets, the first algorithm determines that the current construction stage is the photovoltaic bracket construction stage; When the key component types in the statistical results include photovoltaic brackets and photovoltaic panels, the first algorithm determines that the current construction stage is the photovoltaic panel laying stage.

7. The construction progress monitoring method for a photovoltaic base construction site based on YOLOv8 as claimed in claim 1, characterized in that: The improved model contains a second algorithm for calculating the construction progress percentage of the current construction phase. When executing step 6 above, the second algorithm divides the actual number of key components of the current construction phase in the statistical results by the designed number of key components of the current construction phase.

Images