Intelligent building construction monitoring method and device based on unmanned aerial vehicle remote sensing

Abstract

The invention discloses an intelligent building construction monitoring method and device based on unmanned aerial vehicle remote sensing, and relates to the technical field of construction monitoring. The method comprises the steps of determining a target monitoring position matched with a target monitoring area based on the target monitoring area of a currently executed construction task; acquiring a first image of the target monitoring area at the target monitoring position by using an unmanned aerial vehicle; performing image enhancement processing on the first image by using an MSR algorithm to obtain a target image; performing feature extraction on the target image to obtain target feature information; the target feature information is used for determining the construction condition of the target monitoring area. According to the technical scheme provided by the invention, the construction monitoring efficiency and the monitoring result accuracy can be improved.

Description

Technical Field

[0001] This application relates to the technical field of construction monitoring, and specifically provides an intelligent building construction monitoring method and device based on UAV remote sensing. Background Technique

[0002] Building construction is a complex and time-consuming process involving multiple links and numerous participants. Precise and reliable construction monitoring technology is one of the key factors to ensure the smooth implementation of project management activities such as project quality, construction cost control, and personnel safety. In the prior art, construction monitoring often relies on manual inspection methods, that is, the construction progress and material consumption are obtained through manual statistics, or the personnel and materials on the construction site are obtained through manual inspections. This method not only requires additional manpower, but also has low efficiency and accuracy, greatly affecting the construction progress of buildings.

[0003] Therefore, how to overcome the above-mentioned technical problems and defects has become a key problem to be solved. Summary of the Invention

[0004] The purpose of this application is to provide an intelligent building construction monitoring method and device based on UAV remote sensing to solve the problems raised in the above background technique and improve the efficiency and accuracy of construction monitoring.

[0005] The technical solution of the embodiment of this application is implemented as follows:

[0006] The embodiment of this application provides an intelligent building construction monitoring method based on UAV remote sensing. The method includes:

[0007] Based on the target monitoring area where the current construction task is being executed, determine the target monitoring position that matches the target monitoring area;

[0008] Use a UAV to obtain a first image of the target monitoring area at the target monitoring position;

[0009] Use the Multi-Scale Retinex (MSR) algorithm to perform image enhancement processing on the first image to obtain a target image;

[0010] Extract features from the target image to obtain target feature information; the target feature information is used to determine the construction situation of the target monitoring area.

[0011] In the above solution, the use of the MSR algorithm to perform image enhancement processing on the first image to obtain a target image includes:

[0012] Perform discrete wavelet transform (DWT) processing on the first image to obtain a preliminary denoised image;

[0013] Use the MSR algorithm to perform image enhancement processing on the preliminary denoised image to obtain an enhanced image;

[0014] Perform non-local means (NLM) denoising processing on the enhanced image to obtain a target image.

[0015] In the above solution, the use of the MSR algorithm to perform image enhancement processing on the preliminary denoised image to obtain an enhanced image includes:

[0016] Perform Gaussian blur processing on the preliminary denoised image at different scales to obtain average brightness images at multiple scales;

[0017] Based on the difference between the average brightness image at each scale and the preliminary denoised image, obtain the reflected light image at each scale;

[0018] Perform image fusion processing on the reflected light images at all scales to obtain a target reflected light image;

[0019] Perform power-law transformation on the target reflected light image to obtain an enhanced image.

[0020] In the above solution, the obtaining of the reflected light image at each scale based on the difference between the average brightness image at each scale and the preliminary denoised image includes:

[0021] For the average brightness image at each scale, divide the pixel value of each pixel in the preliminary denoised image by the pixel value of the corresponding pixel in the average brightness image to obtain the relative pixel value of each pixel;

[0022] Generate the reflected light image at the corresponding scale based on the relative pixel value of each pixel.

[0023] In the above solution, the determination of the target monitoring position matching the target monitoring area based on the target monitoring area where the current construction task is being performed includes:

[0024] Determine the area attribute of the target monitoring area where the current construction task is being performed; different area attributes correspond to different spatial features;

[0025] Based on the area attribute and the performance parameters of the drone, determine the flight altitude of the drone when collecting the target monitoring area and the distance from the target monitoring area to obtain the target monitoring position matching the target monitoring area.

[0026] In the above solution, obtaining the first image of the target monitoring area at the target monitoring position by using a drone includes:

[0027] Determine the image acquisition frequency of the target monitoring area based on the priority of the target monitoring area;

[0028] Based on the image acquisition frequency, use a drone to obtain the first image of the target monitoring area at the target monitoring position.

[0029] In the above solution, the method further includes determining the priority of the target monitoring area; determining the priority of the target monitoring area includes:

[0030] Based on at least one preset evaluation index, construct an evaluation system for the priority of the construction area to obtain a first evaluation system;

[0031] Use the first evaluation system to evaluate the priority of the target monitoring area to obtain the priority of the target monitoring area; wherein,

[0032] The evaluation indexes include safety risk, construction progress, regional attributes, and environmental impact.

[0033] In the above solution, the evaluation index may further include: data fluctuation index; the data fluctuation index characterizes the fluctuation of the sensor data at the construction site.

[0034] In the above solution, the method may further include determining the data fluctuation index; determining the data fluctuation index includes:

[0035] Obtain the sensor data of the target monitoring area within the first time period along the time axis to obtain the first data;

[0036] Perform time series analysis on the first data to determine the distribution of the sensor data of the target monitoring along the time axis to obtain the data fluctuation index of the target monitoring area.

[0037] In the above solution, the first data includes multiple groups of second data, and different second data correspond to different sensors; performing time series analysis on the first data to determine the distribution of the sensor data of the target monitoring along the time axis to obtain the data fluctuation index of the target monitoring area includes:

[0038] Perform standardization processing on each group of second data in the first data to obtain multiple groups of third data;

[0039] Perform time series analysis on each group of third data to obtain the single-group data fluctuation degree index of each group of third data;

[0040] Perform a weighted average process on the single - group data fluctuation index of all third - party data to obtain the data fluctuation degree of the target monitoring area;

[0041] Based on the data fluctuation degree of the target monitoring area and the score assigned to the data fluctuation index factor in the first evaluation system, obtain the data fluctuation index of the target monitoring area.

[0042] An embodiment of the present application also provides an intelligent building construction monitoring device based on UAV remote sensing. The device includes:

[0043] A first processing unit, configured to determine a target monitoring position matching the target monitoring area based on the target monitoring area where the current construction task is being executed;

[0044] An acquisition unit, configured to use a UAV to acquire a first image of the target monitoring area at the target monitoring position;

[0045] A second processing unit, configured to perform image enhancement processing on the first image using the MSR algorithm to obtain a target image;

[0046] A third processing unit, configured to extract features from the target image to obtain target feature information; the target feature information is used to determine the construction situation of the target monitoring area.

[0047] The intelligent building construction monitoring method and device based on UAV remote sensing provided by the embodiments of the present application can reduce the human and resource consumption brought by personnel statistics and inspections by using UAVs to collect images at fixed points for construction monitoring. At the same time, the method of replacing humans with machines also greatly improves the processing efficiency of monitoring data and further improves the accuracy of monitoring results. Further, by performing image enhancement processing on the images collected by the UAV, it is possible to reduce the impact of environmental complexity caused by factors such as light changes, obstacles, and dust at the construction site on the image acquisition results, improve the clarity of the image acquisition results, so that when performing feature recognition on the collected images, the accuracy of the feature extraction results can be improved, and thus the accuracy of the monitoring results can be improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] Figure 1 It is a schematic flowchart of an intelligent building construction monitoring method based on UAV remote sensing provided by an embodiment of the present application;

[0049] Figure 2 It is a schematic flowchart of S102 in the intelligent building construction monitoring method based on UAV remote sensing according to an embodiment of the present application;

[0050] Figure 3 It is a schematic flowchart of S2 in the intelligent building construction monitoring method based on UAV remote sensing according to an embodiment of the present application;

[0051] Figure 4 This is a schematic diagram of the first image in the intelligent building construction monitoring method based on UAV remote sensing according to the embodiment of the present application;

[0052] Figure 5 It is a schematic diagram of the target image obtained by performing MSR processing on Figure 4 using the intelligent building construction monitoring method based on UAV remote sensing provided by the embodiment of the present application;

[0053] Figure 6 This is a schematic structural diagram of an intelligent building construction monitoring device based on UAV remote sensing provided by the embodiment of the present application. Detailed implementation manners

[0054] Next, the technical solutions in the embodiments of the present application will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present application.

[0055] In the prior art, construction monitoring can be carried out by using the method of collecting images by UAVs to improve the monitoring efficiency and accuracy and reduce the human consumption; UAVs can quickly obtain high-definition image data over a large range and can also perform tasks in dangerous environments, greatly improving the efficiency and safety of construction monitoring. However, the construction site environment is complex and changeable, and there are many interference factors that affect the clarity of image collection, such as dust, insufficient light, interlaced shadows, and occlusion by construction machinery and materials; among them, flying dust will seriously affect the air transparency, resulting in a decrease in image contrast and loss of details, while insufficient light or overexposure will lead to uneven brightness distribution of the image, affecting the visual recognition effect; these interference factors will have a great impact on the extraction of image features, thus greatly reducing the accuracy of the monitoring results.

[0056] Based on this, in various embodiments of the present application, by performing image enhancement processing on the images collected by UAVs and adopting the method of using UAVs to collect images at fixed points instead of manual inspections for construction monitoring, the human consumption and resource consumption brought by personnel statistics and inspections can be reduced. At the same time, the method of replacing humans with machines also greatly improves the processing efficiency of monitoring data and further improves the accuracy of monitoring results; further, by performing image enhancement processing on the images collected by UAVs, the influence of environmental complexity caused by factors such as light changes, occlusions, and dust on the image collection results at the construction site can be reduced, and the clarity of the image collection results can be improved. Therefore, when performing feature recognition on the collected images, the accuracy of feature extraction results can be improved, and thus the accuracy of monitoring results can be improved.

[0057] An embodiment of the present application provides an intelligent building construction monitoring method based on UAV remote sensing, which is applied to an electronic device, specifically, it can be applied to electronic devices such as personal computers, industrial computers, and mobile terminals. As Figure 1 shown, the method may include S101 to S104. The following will detail S101 to S104 in combination with specific embodiments.

[0058] S101: Based on the target monitoring area where the current construction task is being executed, determine the target monitoring position that matches the target monitoring area.

[0059] In practical applications, the target monitoring area can be understood as the area associated with the current construction task. Specifically, the target monitoring area can be the building under construction during the current construction period, the material stacking area for the materials required for the current construction task, the working area of the construction machinery executing the current construction task, the personnel activity area for the personnel executing the current construction task, etc.

[0060] Here, considering that in different task cycles, the areas associated with the construction task will also be different, that is, the target monitoring areas will also be different. Therefore, each area in the construction area that needs to be monitored can be divided according to the area attributes to obtain multiple monitoring areas, and each monitoring area is numbered; this process can also be called the process of pre-configuring multiple monitoring areas; within each task cycle, according to the building under construction and its associated areas, the corresponding target monitoring area is selected from the multiple pre-configured monitoring areas.

[0061] In practical applications, the area attribute can be understood as the use of the corresponding target monitoring area. Specifically, the area attribute can be construction, material preparation, machinery, or personnel, which respectively represent that the target monitoring area is the building area under construction during the current construction period, the material stacking area for the materials required for the current construction task, the working area of the construction machinery executing the current construction task, or the personnel activity area for the personnel executing the current construction task.

[0062] In practical applications, before S101 is executed, the target monitoring area of the current construction period can be determined first; specifically, according to the construction progress, the building under construction and other functional areas associated with the construction building can be determined.

[0063] Based on this, in one embodiment, determining the target monitoring area of the current construction period may include:

[0064] Determine the building where the construction task is being executed during the current construction period to obtain the target building area;

[0065] Based on each target building, determine the area associated with the construction of the target building to obtain the associated area; the associated area includes a material stacking area and / or a working area of construction machinery.

[0066] Take the target building area and the associated area as the target monitoring area.

[0067] In practical applications, within the same construction period, there can be at least one target monitoring area within the construction area; since the attributes of different target monitoring areas are different, specifically, attributes such as height, shape, structural complexity, and safety requirements are different, therefore, the postures of the unmanned aerial vehicle (UAV) when collecting images of different target monitoring areas are also different; thus, the posture of the UAV when performing the image collection task for different target monitoring areas can be determined according to the area attributes of the different target monitoring areas.

[0068] Based on this, in one embodiment, determining the target monitoring position matching the target monitoring area based on the target monitoring area where the current construction task is being executed, that is, S101, may include:

[0069] Determine the area attributes of the target monitoring area where the current construction task is being executed; different area attributes correspond to different spatial characteristics;

[0070] Based on the area attributes and the performance parameters of the UAV, determine the flight height of the UAV when collecting the target monitoring area and the distance from the target monitoring area, so as to obtain the target monitoring position matching the target monitoring area.

[0071] In practical applications, target monitoring areas with different area attributes may also have different spatial characteristics such as height, shape, and structural complexity. Exemplarily, Table 1 shows the differences in the subspace characteristics of the construction building and the material stacking area:

[0072] Table 1

[0073]

[0074]

[0075] It can be seen from Table 1 that target monitoring areas with different area attributes also have different requirements for the spatial coordinates and quantity of the collection positions of the UAV. Therefore, the UAV attitude configuration strategy (such as the UAV attitude requirements in Table 1) matching the area attributes can be called according to the area attributes of the target monitoring area, and thus, according to the called UAV attitude configuration strategy and the actual construction progress (such as the construction height of the building), and based on the performance parameters of the UAV, the target monitoring position corresponding to the target monitoring area can be configured; here, the UAV attitude configuration strategy can be pre-configured according to the actual application scenario, and as for how to specifically configure it, the embodiments of the present application do not make any limitations in this regard.

[0076] S102: Use the UAV to obtain the first image of the target monitoring area at the target monitoring position.

[0077] In practical applications, the first image refers to the initial image without any processing or modification, which may contain information under various lighting conditions, including direct lighting, reflected light, shadows, etc.; the first image can also be referred to as the original image, and the embodiments of this application do not limit this, as long as its function can be achieved.

[0078] In practical applications, the monitoring importance of different target monitoring areas is different. Therefore, in order to reduce the amount of monitoring data processing while ensuring the comprehensiveness and reliability of the monitoring results, the corresponding sampling frequency can be configured according to the priority of each target monitoring area.

[0079] Based on this, in one embodiment, as Figure 2 shown, the step of using the unmanned aerial vehicle to obtain the first image of the target monitoring area at the target monitoring position, that is, S102, may include:

[0080] S201: Determine the image acquisition frequency of the target monitoring area based on the priority of the target monitoring area;

[0081] S202: Based on the image acquisition frequency, use the unmanned aerial vehicle to obtain the first image of the target monitoring area at the target monitoring position.

[0082] In practical applications, the importance of different target monitoring areas is different within the same construction period, and the importance of the same target monitoring area is also different in different construction periods; therefore, the priority of each corresponding target monitoring area can be evaluated within each construction period, and then the corresponding image acquisition frequency can be configured according to the priority.

[0083] Based on this, in one embodiment, the method may further include:

[0084] Determine the priority of the target monitoring area.

[0085] In practical applications, the priority of the corresponding target monitoring area can be determined before the start of each construction period.

[0086] In practical applications, the priority can be determined by constructing an evaluation system.

[0087] Based on this, in one embodiment, the step of determining the priority of the target monitoring area may include:

[0088] Based on at least one preset evaluation index, construct an evaluation system for the priority of the construction area to obtain a first evaluation system;

[0089] Use the first evaluation system to evaluate the priority of the target monitoring area to obtain the priority of the target monitoring area; wherein,

[0090] The evaluation indicators include safety risks, construction progress, regional attributes, and environmental impacts.

[0091] In practical applications, when using the first evaluation system to evaluate the priority of the target monitoring area and obtaining the priority of the target monitoring area, the values of each evaluation indicator in the first evaluation system can be determined according to the construction data of the target monitoring area, and then the scores of each evaluation indicator are weighted to obtain the final score of the first evaluation system. Then, according to the corresponding relationship between the evaluation system score and the priority, the priority of the corresponding target monitoring area is determined, and then according to the corresponding relationship between the priority and the sampling frequency, the corresponding sampling frequency is determined; of course, a corresponding relationship between the evaluation system score, priority, and sampling frequency can also be established, or a direct corresponding relationship between the evaluation system score and the sampling frequency can be established, so as to directly configure the sampling frequency of the corresponding monitoring area according to the evaluation system score; here, the corresponding relationship can also be called a mapping relationship, which is not limited in this embodiment of the present application as long as its function can be realized.

[0092] In practical applications, the higher the score of the scoring system, the higher the priority and the higher the sampling frequency.

[0093] Exemplarily, the mapping relationship between the evaluation system score, priority, and sampling frequency is shown in Table 2:

[0094] Table 2

[0095] Priority Evaluation system score Sampling frequency First priority > 80 points Twice per hour Second priority 61 - 80 points Once per hour Third priority 30 - 60 points Once per day Fourth priority < 30 points Once per month

[0096] Here, by introducing the priority, the importance of each target monitoring area can be more intuitively seen, so that while configuring the sampling frequency, data support and reference can be provided for the subsequent construction monitoring process; for example, when subsequent operations such as auditing and analyzing the construction site situation are required, the evaluation system scores of each monitoring area in different construction periods can be calculated, then the historical average score of each monitoring area can be calculated, and then all monitoring areas can be sorted according to the historical average score, so as to select the monitoring areas with higher scores for auditing or data analysis according to the sorting results of the monitoring areas.

[0097] In practical applications, the volatility of sensor data in the monitoring area can be introduced into the evaluation system to improve the reliability of the evaluation system scoring results.

[0098] Based on this, in one embodiment, the evaluation indicator may further include a data fluctuation index; the data fluctuation index characterizes the fluctuation of sensor data at the construction site.

[0099] In practical applications, the first evaluation system can be constructed by means of quantitative assignment.

[0100] Exemplarily, the assignment of each evaluation index in the first evaluation system is shown in Table 3 as follows:

[0101] Table 3

[0102]

[0103]

[0104] In practical applications, after obtaining the scores of each evaluation index in the first evaluation system, the weights of the corresponding evaluation indexes can be configured according to the importance of each evaluation index, and then the scores of each evaluation index are weighted according to the configured weights to obtain the final score of the evaluation system; exemplarily, the weights of the five evaluation indexes of safety risk, construction progress, regional attributes, environmental impact, and data fluctuation index shown in Table 3 are 0.3, 0.25, 0.2, 0.15, and 0.1 in sequence.

[0105] In practical applications, before constructing the first evaluation system, the data fluctuation index can be calculated first.

[0106] Based on this, in one embodiment, the method may further include determining the data fluctuation index.

[0107] In practical applications, for each monitoring area among multiple monitoring areas, the sensor data in the monitoring area can be analyzed in time series along the time axis, so as to calculate the data fluctuation index.

[0108] Based on this, in one embodiment, the determining the data fluctuation index may include:

[0109] Obtain the sensor data of the target monitoring area within the first time period along the time axis to obtain the first data;

[0110] Perform time series analysis on the first data to determine the distribution of the sensor data of the target monitoring along the time axis, and obtain the data fluctuation index of the target monitoring area.

[0111] In practical applications, the first time period can be greater than the duration of the previous construction cycle; specifically, it can include the previous N construction cycles, where N is an integer greater than 1; of course, the specific value of the first time period can also be set according to the actual application scenario as long as it is greater than the duration of the previous construction cycle, and the embodiments of the present application do not make any limitations.

[0112] In practical applications, the first data can also be referred to as the first sensor data, and can also be referred to as the historical sensor data. The embodiments of the present application do not make any limitations on this, as long as its function can be realized.

[0113] In practical applications, the sensor data can be the detection data of some sensors in the monitoring area. Specifically, the data of sensors used to evaluate the monitoring target can be selected. For example, noise sensors, vibration sensors, gas sensors, etc. for evaluating safety, and dust sensors, wind speed sensors, etc. for evaluating environmental impacts.

[0114] In practical applications, performing time series analysis on the first data can be achieved by calculating the standard deviation or variance of the first data and normalizing the calculation results.

[0115] In practical applications, after obtaining the sensor data of multiple sensors, the data of each sensor can be integrated to calculate the final data fluctuation index.

[0116] Based on this, in one embodiment, the first data includes multiple groups of second data, and different second data correspond to different sensors; the performing time series analysis on the first data to determine the distribution of the sensor data of the target monitoring along the time axis and obtaining the data fluctuation index of the target monitoring area may include:

[0117] Performing standardization processing on each group of second data in the first data to obtain multiple groups of third data;

[0118] Performing time series analysis on each group of third data to obtain the single - group data fluctuation index of each group of third data;

[0119] Performing weighted average processing on the single - group data fluctuation indices of all third data to obtain the data fluctuation degree of the target monitoring area;

[0120] Based on the data fluctuation degree of the target monitoring area and the score assigned to the data fluctuation index factor in the first evaluation system, obtaining the data fluctuation index of the target monitoring area.

[0121] In practical applications, by performing standardization processing on each group of second data, it can have the same scale, thus ensuring the accuracy of the final data processing result.

[0122] In practical applications, performing time series analysis on each group of third data to obtain the single - group data fluctuation index of each group of third data can be to calculate the standard deviation of each group of third data, and then normalize the standard deviation of each group of third data to obtain the data fluctuation degree of each group of third data.

[0123] Exemplarily, the standard deviation calculation formula of the third data is as follows:

[0124]

[0125] where σ mrepresents the standard deviation of the sensor numbered m, N represents the number of sampling points in the third data, and x i represents the data of sensor m at sampling point i, represents the average value of all sampling point data in the third data.

[0126] In practical applications, based on the data fluctuation degree of the target monitoring area and the score assigned to the data fluctuation index factor in the first evaluation system, the data fluctuation index of the target monitoring area is obtained. It can be understood that according to the assignment of the data fluctuation index in the first evaluation system, the data fluctuation degree is converted into the data fluctuation index. Specifically, it can be multiplying the data fluctuation degree by the assignment of the data fluctuation index in the first evaluation system. For example, if the data fluctuation index score in Table 3 is 20 points, then the actual score of the data fluctuation index is the calculated data fluctuation degree multiplied by 20.

[0127] In practical applications, after determining the score of the evaluation system for the target monitoring area, the sampling frequency corresponding to the score of the evaluation system can be determined according to the preset mapping relationship (such as the mapping relationship between the score of the evaluation system, the priority, and the sampling frequency), and used as the sampling frequency of the target monitoring area during the current construction period.

[0128] S103: Perform image enhancement processing on the first image using the MSR algorithm to obtain a target image.

[0129] In practical applications, DWT and NLM can be introduced into the traditional MSR network architecture to improve the image quality.

[0130] Based on this, in one embodiment, the performing image enhancement processing on the first image using the MSR algorithm to obtain a target image, that is, S103, may include:

[0131] S1: Perform DWT processing on the first image to obtain a preliminary denoised image;

[0132] S2: Perform image enhancement processing on the preliminary denoised image using the MSR algorithm to obtain an enhanced image;

[0133] S3: Perform NLM denoising processing on the enhanced image to obtain a target image.

[0134] In one embodiment, the performing DWT processing on the first image to obtain a preliminary denoised image, that is, S1, may include:

[0135] Perform two-dimensional DWT processing on the first image to obtain multiple detail coefficients;

[0136] Perform denoising processing on each detail coefficient using a soft threshold function;

[0137] Inverse transform the denoised detail coefficients to obtain a preliminary denoised image.

[0138] Here, multiple detail coefficients may include an approximation coefficient (LL), a horizontal detail coefficient (LH), a vertical detail coefficient (HL), and a diagonal detail coefficient (HH).

[0139] In practical applications, denoising each detail coefficient using a soft threshold function can be expressed as:

[0140]

[0141] where C j represents the detail coefficient of the j-th layer, represents the denoising result of the detail coefficient of the j-th layer, T j represents the threshold, and S(×) represents the soft threshold function.

[0142] Here, by introducing DWT, it is possible to effectively remove high-frequency noise while retaining the main structural information of the image, thereby improving the accuracy of the image feature extraction results in low-illumination scenarios (such as dust, shadow occlusion, etc.).

[0143] In one embodiment, the NLM denoising process on the enhanced image to obtain the target image, i.e., S3, may include:

[0144] For each pixel in the enhanced image, use the pixel as a reference block;

[0145] Determine the similar blocks of the reference block within the neighborhood of the reference block;

[0146] Perform weighted averaging on the pixel values of all similar blocks to obtain the updated pixel value of the reference block;

[0147] Based on the updated pixel values of each pixel in the enhanced image, obtain the target image.

[0148] In practical applications, for each reference block, a similarity measurement method can be used to calculate its similarity with other blocks in the image, and the blocks with similarity higher than a preset threshold are used as the corresponding similar blocks; among them, the threshold can be configured according to the specific application scenario and noise level, and this application does not make restrictions.

[0149] In practical applications, similarity measurement methods such as Euclidean distance, weighted Euclidean distance, or normalized cross-correlation (NCC) can be used.

[0150] Here, by introducing NLM, the relationship between similar blocks in the image can be introduced during the image enhancement process, so that noise can be effectively removed while retaining details.

[0151] Here, by adding DWT and NLM processing in image enhancement, not only the effect of the traditional MSR algorithm is enhanced, but also significant progress has been made in noise suppression and detail retention, which is applicable to image processing tasks in complex environments.

[0152] In one embodiment, as Figure 3 shown, using the MSR algorithm to perform image enhancement processing on the preliminary denoised image to obtain an enhanced image, that is, S2, specifically may include:

[0153] S301: Perform Gaussian blur processing on the preliminary denoised image at different scales to obtain average brightness images at multiple scales;

[0154] S302: Based on the difference between the average brightness image at each scale and the preliminary denoised image, obtain the reflected light image at each scale;

[0155] S303: Perform image fusion processing on the reflected light images at all scales to obtain a target reflected light image;

[0156] S304: Perform power-law transformation on the target reflected light image to obtain an enhanced image.

[0157] In practical applications, when performing Gaussian blur processing on the preliminary denoised image at different scales, Gaussian blur processing can be performed on the preliminary denoised image at different spatial scales or resolutions, so as to achieve brightness averaging processing on the image at each scale and obtain the average brightness image at each scale; in this process, for each pixel in the preliminary denoised image, the average brightness of the local area centered on this pixel can be calculated, and this average brightness can be used as the reference brightness of this pixel, and then according to the reference brightness of all pixels, an image representing the average brightness at this scale can be obtained.

[0158] Exemplarily, the MSR algorithm can be implemented using the following formula:

[0159]

[0160] where, R(x, y) represents the enhanced image, I(x, y) represents the original image, F i represents the Gaussian filter of the i-th scale, * represents the convolution operation, ω i represents the weight of the i-th scale, ω i satisfies

[0161] Here, the MSR algorithm can also be called the multi-scale Retinex algorithm, which is an image enhancement algorithm derived from the visual perception mechanism of the human eye and can effectively handle the problem of uneven illumination. When the human eye observes a scene, it automatically adjusts its sensitivity to light according to the brightness of the surrounding environment, so as to perceive the true brightness and color of the scene. The MSR algorithm simulates this process by calculating the average brightness of the area around each pixel in the image and using it as the reference brightness of the pixel, thereby obtaining an enhanced version of the image. Here, by introducing the MSR algorithm into the processing of UAV remote sensing images, not only the enhancement of image contrast and color saturation, as well as exposure correction, are realized, reducing the impact of uneven illumination at the construction site on image feature extraction, but also image dehazing can be achieved, reducing the problem of low image clarity caused by dust or fog at the construction site. Therefore, by introducing the MSR algorithm to process the collected images, the clarity of the object contours in the images can be improved, thereby improving the accuracy of the feature extraction results.

[0162] In practical applications, based on the difference between the average brightness image of each scale and the preliminary denoised image, when obtaining the reflected light image of each scale, the difference between the same pixel position in the preliminary denoised image and the average brightness image can be determined pixel by pixel, and the reflected light image corresponding to the scale can be obtained according to the differences of all pixel positions.

[0163] Based on this, in one embodiment, the obtaining the reflected light image of each scale based on the difference between the average brightness image of each scale and the preliminary denoised image may include:

[0164] For the average brightness image of each scale, divide the pixel value of each pixel in the preliminary denoised image by the pixel value of the corresponding pixel in the average brightness image to obtain the relative pixel value of each pixel;

[0165] Generate the reflected light image corresponding to the scale based on the relative pixel value of each pixel.

[0166] In practical applications, considering that the intensity of the reflected light is usually related to the average brightness of the background or environment, therefore, by dividing the pixel value of each pixel in the preliminary denoised image by the pixel value of the corresponding pixel in the average brightness image at each scale, the normalization processing of the original image is realized, so as to eliminate or reduce the impact of brightness changes at different scales on the image, and further significantly highlight the reflected light information in the image. At the same time, since the above process is carried out at multiple scales, a series of images reflecting the reflected light information at different scales will be finally obtained, so that the details or features that are not easily noticed in the original image can be presented by using these images, thereby improving the image enhancement effect and the accuracy of the image recognition result. Figure 5 Shows the use of the method described in the embodiments of the present application for Figure 4The results of the processing, such as Figure 4 and Figure 5 shown, compared with Figure 4 , Figure 5 the buildings and the construction site ground area in

[0167] S104: Extract features from the target image to obtain target feature information; the target feature information is used to determine the construction situation of the target monitoring area.

[0168] In practical applications, the method may further include determining the construction situation of the target monitoring area based on the target features.

[0169] In practical applications, algorithms such as Scale-invariant feature transform (SIFT) and Histogram of Oriented Gradient (HOG) can be used to implement feature extraction of the target image, which is not limited in the embodiments of the present application.

[0170] In summary, the intelligent building construction monitoring method based on UAV remote sensing provided by the embodiments of the present application can reduce the labor consumption and resource consumption brought by personnel statistics and inspections by using UAVs to collect images at fixed points instead of manual inspections. At the same time, the method of replacing manual labor with machines also greatly improves the processing efficiency of monitoring data and further improves the accuracy of monitoring results. Further, by performing image enhancement processing on the images collected by the UAV, the impact of environmental complexity caused by factors such as light changes, occlusions, and dust in the construction site on the image collection results can be reduced, and the clarity of the image collection results can be improved. Therefore, when performing feature recognition on the collected images, the accuracy of the feature extraction results can be improved, and thus the accuracy of the monitoring results can be improved.

[0171] To implement the intelligent building construction monitoring method based on UAV remote sensing of the present application, the embodiments of the present application further provide an intelligent building construction monitoring device based on UAV remote sensing, which is set on an electronic device, such as Figure 6 shown, and the device may include:

[0172] The first processing unit 601 is configured to determine a target monitoring position matching the target monitoring area based on the target monitoring area where the current construction task is being performed;

[0173] The acquisition unit 602 is configured to use a UAV to acquire a first image of the target monitoring area at the target monitoring position;

[0174] The second processing unit 603 is configured to perform image enhancement processing on the first image by using the MSR algorithm to obtain a target image;

[0175] The third processing unit 604 is configured to perform feature extraction on the target image to obtain target feature information; the target feature information is used to determine the construction condition of the target monitoring area.

[0176] In one embodiment, the second processing unit 603 may specifically be configured to:

[0177] Perform discrete wavelet transform (DWT) processing on the first image to obtain a preliminary denoised image;

[0178] Perform image enhancement processing on the preliminary denoised image by using the MSR algorithm to obtain an enhanced image;

[0179] Perform non-local means (NLM) denoising processing on the enhanced image to obtain a target image.

[0180] In one embodiment, the performing image enhancement processing on the preliminary denoised image by using the MSR algorithm to obtain an enhanced image may include:

[0181] Perform Gaussian blur processing on the preliminary denoised image at different scales to obtain average brightness images at multiple scales;

[0182] Based on the difference between the average brightness image at each scale and the preliminary denoised image, obtain a reflected light image at each scale;

[0183] Perform image fusion processing on the reflected light images at all scales to obtain a target reflected light image;

[0184] Perform power-law transformation on the target reflected light image to obtain an enhanced image.

[0185] In one embodiment, the based on the difference between the average brightness image at each scale and the preliminary denoised image, obtaining a reflected light image at each scale may specifically include:

[0186] For the average brightness image at each scale, divide the pixel value of each pixel in the preliminary denoised image by the pixel value of the corresponding pixel in the average brightness image to obtain the relative pixel value of each pixel;

[0187] Based on the relative pixel value of each pixel, generate a reflected light image at the corresponding scale.

[0188] In one embodiment, the first processing unit 601 may specifically be configured to:

[0189] Determine the regional attribute of the target monitoring area where the current construction task is being performed; different regional attributes correspond to different spatial features;

[0190] Based on the regional attributes and the performance parameters of the drone, determine the flight altitude of the drone when collecting the target monitoring area and the distance from the target monitoring area, so as to obtain a target monitoring position matching the target monitoring area.

[0191] In one embodiment, the acquisition unit 602 can specifically be used for:

[0192] Based on the priority of the target monitoring area, determine the image acquisition frequency of the target monitoring area;

[0193] Based on the image acquisition frequency, use the drone to obtain a first image of the target monitoring area at the target monitoring position.

[0194] In one embodiment, the first processing unit 601 can also be used to determine the priority of the target monitoring area; the determining the priority of the target monitoring area includes:

[0195] Based on at least one preset evaluation index, construct an evaluation system for the priority of the construction area to obtain a first evaluation system;

[0196] Use the first evaluation system to evaluate the priority of the target monitoring area to obtain the priority of the target monitoring area; wherein,

[0197] The evaluation indexes include safety risk, construction progress, regional attributes and environmental impact.

[0198] In one embodiment, the evaluation index can also include: data fluctuation index; the data fluctuation index characterizes the fluctuation of the sensor data at the construction site.

[0199] In one embodiment, the first processing unit 601 can also be used to determine the data fluctuation index; the determining the data fluctuation index includes:

[0200] Acquire the sensor data of the target monitoring area within the first time period along the time axis to obtain first data;

[0201] Perform time series analysis on the first data to determine the distribution of the sensor data of the target monitoring along the time axis, so as to obtain the data fluctuation index of the target monitoring area.

[0202] In one embodiment, the first data includes multiple groups of second data, and different second data correspond to different sensors; the performing time series analysis on the first data to determine the distribution of the sensor data of the target monitoring along the time axis, so as to obtain the data fluctuation index of the target monitoring area includes:

[0203] Normalize each group of the second data in the first data to obtain multiple groups of third data;

[0204] Perform time series analysis on each group of the third data to obtain the single-group data fluctuation degree index of each group of the third data;

[0205] Perform weighted average processing on the single-group data fluctuation indexes of all the third data to obtain the data fluctuation degree of the target monitoring area;

[0206] Based on the data fluctuation degree of the target monitoring area and the score assigned to the data fluctuation index factor in the first evaluation system, obtain the data fluctuation index of the target monitoring area.

[0207] It should be noted that when the intelligent building construction monitoring device based on UAV remote sensing provided in the above embodiment performs intelligent building construction monitoring based on UAV remote sensing, only the above division of each program module is used for illustration. In practical applications, the above processing can be allocated to different program modules according to needs, that is, the internal structure of the device is divided into different program modules to complete all or part of the above-described processing. In addition, the intelligent building construction monitoring device based on UAV remote sensing provided in the above embodiment and the embodiment of the intelligent building construction monitoring method based on UAV remote sensing belong to the same concept. For the specific implementation process, please refer to the method embodiment, which will not be elaborated here.

[0208] It should be noted that "first", "second", etc. are used to distinguish similar objects and do not necessarily describe a specific order or sequence.

[0209] In addition, the technical solutions described in the embodiments of the present application can be arbitrarily combined without conflict.

[0210] The above is only a preferred embodiment of the present application and is not intended to limit the protection scope of the present application.

Claims

1. A method for monitoring intelligent building construction based on drone remote sensing, characterized in that: The method comprises: Based on the target monitoring area of ​​the current construction task being performed, determining a target monitoring position that matches the target monitoring area; Acquire a first image of the target monitoring area at the target monitoring position using a drone; Performing image enhancement processing on the first image using a multi-scale retinal enhancement (MSR) algorithm to obtain a target image; Feature extraction is performed on the target image to obtain target feature information; the target feature information is used to determine the construction status of the target monitoring area.

2. The method according to claim 1, characterized in that The using the MSR algorithm to perform image enhancement processing on the first image to obtain a target image includes: Performing wavelet denoising (DWT) processing on the first image to obtain a preliminary denoised image; Performing image enhancement processing on the preliminary denoised image using an MSR algorithm to obtain an enhanced image; The enhanced image is subjected to non-local mean denoising (NLM) processing to obtain a target image.

3. The method according to claim 2, characterized in that The method of performing image enhancement processing on the preliminary denoised image by using the MSR algorithm to obtain an enhanced image includes: Performing Gaussian blur processing of different scales on the preliminary denoised image to obtain average brightness images of multiple scales; Obtaining a reflected light image at each scale based on a difference between the average brightness image at each scale and the preliminary denoised image; Perform image fusion processing on reflected light images of all scales to obtain the target reflected light image; A power law transformation is performed on the target reflected light image to obtain an enhanced image.

4. The method according to claim 3, characterized in that The step of obtaining a reflected light image at each scale based on the difference between the average brightness image at each scale and the preliminary denoised image comprises: For the average brightness image of each scale, dividing the pixel value of each pixel in the preliminary denoised image by the pixel value of the corresponding pixel in the average brightness image to obtain a relative pixel value of each pixel; Based on the relative pixel value of each pixel, a reflected light image of the corresponding scale is generated.

5. The method according to claim 1, characterized in that The step of determining a target monitoring position matching the target monitoring area based on the target monitoring area currently performing the construction task comprises: Determine the regional attributes of the target monitoring area where the construction task is currently being performed; different regional attributes correspond to different spatial features; Based on the area attributes and the performance parameters of the drone, the flight altitude of the drone when collecting data from the target monitoring area and the distance from the target monitoring area are determined to obtain a target monitoring position that matches the target monitoring area.

6. The method according to claim 5, characterized in that The method of acquiring a first image of the target monitoring area at the target monitoring position by using a drone includes: Based on the priority of the target monitoring area, determining the image acquisition frequency of the target monitoring area; Based on the image acquisition frequency, a first image of the target monitoring area is acquired at the target monitoring position using a drone.

7. The method according to claim 6, characterized in that The method further includes determining the priority of the target monitoring area; the determining the priority of the target monitoring area includes: Based on at least one preset evaluation index, an evaluation system for construction area priority is constructed to obtain a first evaluation system; The priority of the target monitoring area is evaluated by using the first evaluation system to obtain the priority of the target monitoring area; wherein, The evaluation indicators include safety risks, construction progress, regional attributes and environmental impact.

8. The method according to claim 7, characterized in that The evaluation index may further include: a data fluctuation index; the data fluctuation index represents the fluctuation of the construction site sensor data; the method may further include determining the data fluctuation index; the determining the data fluctuation index includes: Acquire sensor data of the target monitoring area within a first time period along the time axis to obtain first data; A time series analysis is performed on the first data to determine the distribution of the sensor data of the target monitoring along the time axis, and obtain a data fluctuation index of the target monitoring area.

9. The method according to claim 8, characterized in that The first data includes multiple groups of second data, and different second data correspond to different sensors; the time series analysis of the first data is performed to determine the distribution of the sensor data of the target monitoring according to the time axis, and obtain the data fluctuation index of the target monitoring area, including: Performing standardization processing on each group of second data in the first data to obtain multiple groups of third data; Performing time series analysis on each set of third data to obtain a single set of data volatility index for each set of third data; Performing weighted average processing on the single group data fluctuation indexes of all third data to obtain the data fluctuation degree of the target monitoring area; Based on the data volatility of the target monitoring area and the score assigned to the data volatility index factor in the first evaluation system, the data volatility index of the target monitoring area is obtained.

10. An intelligent building construction monitoring device based on drone remote sensing, characterized in that: The device comprises: A first processing unit is used to determine a target monitoring position matching the target monitoring area based on the target monitoring area where the construction task is currently being performed; A collection unit, configured to acquire a first image of the target monitoring area at the target monitoring position using a drone; A second processing unit is used to perform image enhancement processing on the first image by using an MSR algorithm to obtain a target image; The third processing unit is used to extract features from the target image to obtain target feature information; the target feature information is used to determine the construction status of the target monitoring area.

Images