Building facade cleaning effect quantitative evaluation and pricing method

Abstract

The invention discloses a building facade cleaning effect quantitative evaluation and pricing method, and relates to the technical field of building engineering and cleaning service management. Comprising the steps that images and physical parameters of the building facade before and after cleaning are collected, the cleanliness index, the glossiness recovery rate and the detail stain residual rate are calculated respectively, and a comprehensive cleaning effect quantized value is generated through dynamic empowerment addition and fusion; based on the basic quotation of cleaning construction, mapping the comprehensive cleaning effect quantized value into an effect floating coefficient through a nonlinear function; a working condition adjustment coefficient is constructed according to the facade material, the stain type and the construction risk; the area of different materials of the building facade is identified, the effect floating coefficient and the working condition adjustment coefficient are combined, material-based reference unit prices are subjected to linkage correction and summarization, and a final valuation is generated; and binding the whole process data and a calculation result with spatio-temporal information, and generating a non-tampering traceable electronic file and a verification report through a block chain technology.

Description

Technical Field

[0001] This invention belongs to the field of building engineering and cleaning service management technology, specifically relating to a method for quantitative evaluation and pricing of building facade cleaning effects. Background Technology

[0002] Building facade cleaning, as an important building maintenance service, has always faced challenges in terms of effectiveness evaluation and reasonable pricing.

[0003] Traditional cleaning services suffer from the following technical problems: Cleaning effectiveness relies primarily on visual inspection and experience, lacking objective and quantifiable evaluation standards. Furthermore, different evaluators may offer significantly different assessments of the same cleaning result, easily leading to quality disputes between service providers and homeowners. Cleaning costs are typically estimated based on area and manual experience, lacking a direct correlation with cleaning effectiveness, construction difficulty, and material characteristics, resulting in a disconnect between price and service quality and failing to reflect the market principle of "high quality, high price." Key information such as the pre- and post-cleaning condition, construction process data, and evaluation results lacks systematic recording and reliable documentation, making it difficult to trace and verify in case of disputes. Finally, traditional methods often focus on the overall cleanliness of large areas while neglecting the cleaning effect on details such as seams, corners, and decorative components—areas that are crucial to demonstrating cleaning quality.

[0004] Therefore, there is an urgent need to develop a comprehensive technical solution that can objectively evaluate the cleaning effect, scientifically and reasonably price the product, and achieve reliable traceability throughout the entire process. Summary of the Invention

[0005] This invention aims to solve at least one of the technical problems existing in the prior art; to this end, this invention proposes a method for quantitative evaluation and pricing of building facade cleaning effects, to solve the following technical problem: The traditional reliance on manual visual inspection and experience-based judgment leads to subjective biases and incomplete assessments; the traditional pricing model is rigid and fails to reflect the true value and technical difficulty of the service; the service process and results lack credible evidence, making it difficult to trace and verify in case of disputes; and the use of a comprehensive unit price due to the huge differences in cleaning costs for different materials results in inaccurate and unfair pricing.

[0006] To address the above problems, this invention provides a method for quantitatively evaluating and pricing the cleaning effect of building facades, comprising the following steps: S1: Collect images and physical parameters of the building facade before and after cleaning, calculate the cleanliness index, gloss recovery rate and detail stain residue rate through preset algorithms, and generate a comprehensive cleaning effect quantification value by dynamically weighting and adding them together. S2: Based on the basic quotation for cleaning construction, the quantitative value of the comprehensive cleaning effect is mapped into an effect fluctuation coefficient through a non-linear function; S3: Adjustment coefficient for construction conditions based on facade material, stain type, and construction risk; S4: Identify the area of ​​different materials on the building facade, combine the effect fluctuation coefficient and working condition adjustment coefficient, and perform linkage correction and summary on the benchmark unit price based on the material to generate the final price; S5: Binds the data and calculation results of the entire process to spatiotemporal information, and generates tamper-proof and traceable electronic archives and verification reports through blockchain technology.

[0007] Preferably, in step S1, the cleanliness index includes the following steps: High-definition imaging equipment was used to collect images of the exterior facade of the same area before and after cleaning under the same lighting and shooting angle. Image processing technology was used to segment and count the total area of ​​stains before and after cleaning. The average gray value of the surface before cleaning and the average gray value of the surface after cleaning are extracted using a gray-level histogram algorithm. The cleanliness index is obtained by determining preset weight coefficients through the analytic hierarchy process (AHP) combined with expert experience, and then weighting and summing the results. Specifically: ,in, and are preset weighting coefficients, where is the total area of ​​the stains after cleaning. This represents the total area of ​​the stain before cleaning. This represents the average grayscale value of the surface after cleaning. This represents the average grayscale value of the surface before cleaning.

[0008] Preferably, in step S1, the gloss recovery rate includes the following steps: Obtain the standard gloss value of the exterior building materials in their brand-new state. Then, measure the gloss value before and after cleaning at the same testing point. Calculate the gloss recovery rate by the percentage difference. Specifically: in, Gloss recovery rate This is the gloss value after cleaning. This is the gloss value before cleaning. This refers to the standard gloss value of building materials in their brand-new state.

[0009] Preferably, in step S1, the detailed stain residue rate includes the following steps: Establish a standard image library of ideal clean-state details for various building facade details, including joints, edges, and decorative components of different materials; The detailed area images acquired after cleaning are registered and aligned with the corresponding standard images in the standard image library. The overall difference between the cleaned image and the standard image is calculated as a characterization value for stain residue in this type of detailed area. Specifically: ;in, These are the weighting coefficients. It is a structural similarity index. For the first The actual image after cleaning of the detailed areas. The corresponding image in the standard image library Standard clean images of detailed areas It is an L1 norm. The total number of pixels in a single image; Calculate the average degree of difference across all evaluated detail categories as the final detail stain retention rate. Specifically: ,in, This refers to the number of detailed area categories involved in this assessment.

[0010] Preferably, in step S1, generating a quantitative value for the overall cleaning effect through weighted summation includes the following steps: Based on the scene attributes and key requirements of the facade, the corresponding weight combination is dynamically selected from the pre-trained weight configuration pattern library; The weight configuration mode library is a classification model trained based on historical successful project data. Its output is a recommended weight coefficient combination of three factors: cleanliness index, gloss recovery rate, and detail stain residue rate. The cleanliness index, gloss recovery rate, and detail stain residue rate were normalized. The normalized values ​​were then weighted and added to the recommended weighting coefficients to obtain a quantitative value for the overall cleaning effect. Specifically: ,in, , and These are the normalized values ​​of cleanliness index, gloss recovery rate, and detail stain residue rate, respectively. , and Recommended weight coefficient combinations.

[0011] Preferably, in step S2, the nonlinear mapping function includes the following steps: Preset effect benchmark threshold It also defines two effect fluctuation ranges, including an incentive range and a penalty range; Based on the aforementioned effect fluctuation range, a nonlinear mapping function is constructed to calculate the effect fluctuation coefficient. Specifically: in, This represents the theoretical maximum value of the overall cleaning effect quantification. For the excitation intensity coefficient, The penalty intensity coefficient, and This is the curvature adjustment coefficient.

[0012] Preferably, step S3 includes the following steps: Obtain the static operating parameters of the building facade, quantify each parameter according to the preset classification rules, and form the corresponding static operating condition influence factors. Based on predefined coupling evaluation rules, the interaction between the static operating condition influencing factors is nonlinearly evaluated, and the coupling correction coefficient characterizing the complexity is calculated. The coupling evaluation rule is established by analyzing the correlation between multi-factor combinations and corresponding construction efficiency in historical working condition data; Access the real-time monitoring data stream of the construction process, identify risky operation periods based on predefined risk assessment rules, and calculate dynamic risk adjustment factors based on the proportion of risky operation time to total working hours; The static operating condition influence factor, coupling correction coefficient, and dynamic risk adjustment factor are comprehensively calculated to generate the final operating condition adjustment coefficient.

[0013] Preferably, the operating condition adjustment coefficient includes the following steps: in, This is the operating condition adjustment factor. This refers to the number of types of static operating condition parameters. For the first The weighting coefficients of static operating condition factors, For the first Quantitative influencing factors of quasi-static operating condition factors, For coupling correction coefficients, It serves as a dynamic risk adjustment factor.

[0014] Preferably, step S4 includes the following steps: A semantic segmentation model is used to automatically identify and segment areas of different materials on the overall image of the building facade. Based on the segmentation results and the spatial scale calibration information of the image, the cleaning area of ​​each material component is automatically calculated. ,in, Material type; Construct a benchmark unit price library, whereby the unit price library is for each material. A preset basic cleaning unit price is established, and the aforementioned working condition adjustment coefficient is used as a difficulty multiplier to initially adjust the basic unit price, resulting in a dynamic sub-item benchmark unit price. ,in, Basic cleaning unit price; Introducing the aforementioned effect fluctuation coefficient, calculate the adjusted price for each material component. Summarize all itemized prices to obtain the final total price. .

[0015] Preferably, step S5 includes the following steps: Based on the quantitative value of the overall cleaning effect and the residual rate of detailed stains, the overall cleaning achievement rate and the detailed cleaning rate are calculated, and a visual verification report containing the percentage indicators, key process data and image comparisons is automatically generated. Calculate the hash value of the verification report and key process data, and upload the hash value, key indicator triples, timestamp, and GPS location information together to form a notarized transaction to the blockchain network to obtain a notarized certificate containing a unique transaction hash TxID; wherein, the key indicator triples ,in, To achieve a comprehensive cleanliness achievement rate, For detailed cleaning rate; Generate an embedded transaction hash TxID and key indicators. Intelligent pricing sheet; A public verification interface based on the transaction hash TxID is provided, supporting the verification of data integrity and consistency through a blockchain explorer.

[0016] The beneficial effects of this invention are: This invention transforms the subjective "cleanliness" into a precise and measurable quantitative indicator through image recognition and multi-algorithm fusion, including cleanliness index, gloss recovery rate, and detail cleaning rate. This significantly reduces interference from human factors, provides the industry with a unified and replicable quality assessment standard, and greatly enhances the credibility and acceptability of the assessment results. This invention achieves "higher price for better quality, penalty for poor quality" by mapping the quantified value of comprehensive effect to an effect fluctuation coefficient through a piecewise nonlinear function. Secondly, it adopts a multi-level coupled decision tree model to nonlinearly couple static factors such as material, stains, and construction height, and introduces a dynamic risk adjustment factor based on real-time monitoring to jointly generate a working condition adjustment coefficient. Finally, it combines the area of ​​each material item automatically identified through semantic segmentation with the dynamic unit price for linked pricing. This achieves a precise and dynamic link between price and delivery effect and construction difficulty, while also allowing for more prudent assessment and management of construction risks. This invention generates a unique hash value (digital fingerprint) by calculating a complete verification report and key process data, and packages this hash value together with core business indicator triples and spatiotemporal information, and uploads it to the blockchain in the form of a notarized transaction to obtain a unique transaction hash TxID as a "digital ID card". This solves the long-standing problem of "verbal testimony is not enough" in the service industry and generates a legally valid electronic evidence package for each cleaning service. Attached Figure Description

[0017] Figure 1 This is a schematic diagram of the method flow of the present invention. Detailed Implementation

[0018] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0019] Please see Figure 1 As shown, this invention provides a method for quantitative evaluation and pricing of building facade cleaning effects, comprising the following steps: S1: Collect images and physical parameters of the building facade before and after cleaning, calculate the cleanliness index, gloss recovery rate and detail stain residue rate through preset algorithms, and generate a comprehensive cleaning effect quantification value by dynamically weighting and adding them together. S2: Based on the basic quotation for cleaning construction, the quantitative value of the comprehensive cleaning effect is mapped into an effect fluctuation coefficient through a non-linear function; S3: Adjustment coefficient for construction conditions based on facade material, stain type, and construction risk; S4: Identify the area of ​​different materials on the building facade, combine the effect fluctuation coefficient and working condition adjustment coefficient, and perform linkage correction and summary on the benchmark unit price based on the material to generate the final price; S5: Binds the data and calculation results of the entire process to spatiotemporal information, and generates tamper-proof and traceable electronic archives and verification reports through blockchain technology.

[0020] In one embodiment of the present invention, step S1, the cleanliness index, includes the following steps: High-definition imaging equipment was used to collect images of the exterior facade of the same area before and after cleaning under the same lighting and shooting angle. Image processing technology was used to segment and count the total area of ​​stains before and after cleaning. The average gray value of the surface before cleaning and the average gray value of the surface after cleaning are extracted using a gray-level histogram algorithm. The cleanliness index is obtained by determining preset weight coefficients through the analytic hierarchy process (AHP) combined with expert experience, and then weighting and summing the results. Specifically: ,in, and For preset weighting coefficients, This represents the total area of ​​the stain after cleaning. This represents the total area of ​​the stain before cleaning. This represents the average grayscale value of the surface after cleaning. This represents the average grayscale value of the surface before cleaning.

[0021] Specifically, before carrying out the cleaning operation, select a DSLR or mirrorless digital camera with at least 24 megapixels, paired with a fixed-focus lens (such as 35mm or 50mm), fixed on a sturdy tripod, set the camera to manual mode (M mode), and fix the aperture value (f / 8), shutter speed, and ISO (sensitivity). 100), turn off automatic white balance, and shoot in RAW format to retain maximum image information; along the building facade, set up fixed collection points in a grid pattern (e.g., every 10 meters × 10 meters), and affix physical markers (adhesive labels with cross-shaped markings) to each point on the wall as a reference for image alignment and subsequent reshoots; choose clear, cloudless weather, and shoot during periods of stable solar altitude angle and uniform lighting from 9:00 AM to 11:00 AM or 2:00 PM to 4:00 PM to minimize the effects of shadows and glare; the two shots before and after cleaning must be completed under the same weather conditions and within the same time period, no more than 24 hours apart; keep the camera lens axis perpendicular to the wall and shoot at each marked point to ensure that the framing and angle of the images before and after cleaning are completely consistent, and record the point simultaneously with each shot. The system collects GPS coordinates, timestamps, and shooting parameters, and stores them in the project database. The acquired images then enter an automated processing pipeline: First, the system uses feature point matching algorithms (such as SIFT or ORB) to automatically and accurately align the cleaned image with the uncleaned image, eliminating errors caused by minor shooting offsets. Then, based on a standard color chart or grayscale chart (included in the corner of the image during shooting), the system performs color and brightness correction on the images, ensuring both images are in the same color space and lighting standard. To avoid interference from non-cleaned areas such as window frames and structural seams, the system can automatically or manually delineate the main exterior areas requiring stain analysis (e.g., stone walls, glass panels). First, the color images are converted to grayscale images, and then an adaptive threshold segmentation algorithm (e.g., Otsu's method) is used. The method involves processing the image before cleaning to obtain a binary mask image of the stains before cleaning. For the image after cleaning, a segmentation method based on gray-level difference and texture analysis is used. Specifically: the gray-level difference map between the cleaned and pre-cleaning images within the ROI is calculated; threshold segmentation is applied to this difference map (the threshold needs to be set according to the noise level, for example, a difference greater than 15 gray levels); simultaneously, the Local Binary Pattern (LBP) texture features of the cleaned image are extracted and compared with a standard clean texture library to identify texture abnormal areas; finally, the gray-level difference segmentation results and texture abnormality detection results are fused to obtain a binary mask image of the stains after cleaning; the total number of white (stain) pixels in the binary mask images before and after cleaning is counted respectively; based on the known size of the reference object at the time of shooting (e.g., the spacing between marker points is 1 meter), the actual area represented by a single pixel in the image is calculated, and the total area of ​​the stains before and after cleaning is obtained through calculation.Convert the corrected color images before and after cleaning (ROI area only) into grayscale images, calculate the grayscale histogram of each grayscale image, which shows the grayscale distribution of all pixels (0-255) in the image, and calculate the arithmetic mean of the grayscale values ​​of all pixels in the image based on the grayscale histogram, which is the surface average grayscale value. In this embodiment, the weights are determined using the analytic hierarchy process (AHP) combined with expert experience. The value is 0.25. The value is 0.75.

[0022] In one embodiment of the present invention, step S1, the gloss recovery rate, includes the following steps: Obtain the standard gloss value of the exterior building materials in their brand-new state. Then, measure the gloss value before and after cleaning at the same testing point. Calculate the gloss recovery rate by the percentage difference. Specifically: in, Gloss recovery rate This is the gloss value after cleaning. This is the gloss value before cleaning. This refers to the standard gloss value of building materials in their brand-new state.

[0023] Specifically, before conducting the assessment, a database containing standard gloss values ​​for common exterior building materials should be established or accessed. This database should cover mainstream building materials such as granite, marble, glass curtain walls, aluminum composite panels, stone paint, fluorocarbon paint, and metals (aluminum panels, stainless steel). For each building material, standard samples in their brand-new condition should be collected, or samples should be taken from brand-new parts of the building not exposed to the outdoor environment (such as indoor material preparation areas or warehouse inventory). Under standard experimental conditions (temperature 23±2°C, humidity 50±5%), measurements should be taken using a calibrated portable gloss meter according to the standards. At least five points should be randomly selected on each standard sample for measurement. After removing outliers, the arithmetic mean should be taken as the standard gloss value for that building material at a specified angle. The building material type, brand / model, measurement angle, standard gloss value, and relevant environmental parameters should be stored in the database. For example, the database might record: "Shandong white granite (polished), 60°." =85GU”; The selection principles for monitoring points are as follows: Points should cover areas with different orientations and heights of the building, as well as varying degrees of contamination (areas prone to splashing near the ground, normal areas in the middle, and less contaminated areas at the top); ensure safe and convenient measurement before and after construction; and ensure each point corresponds to a single, homogeneous building material, avoiding cross-material measurements; use environmentally friendly markers that will not damage the facade or affix waterproof coded labels, clearly marking each selected testing point (numbered A-01). (B-02) Use a mobile terminal APP to record the GPS location, facade, material type and number of each marker point, and take photos of the surrounding environment to form a "Gloss Detection Point Distribution Map"; complete the initial measurement of all detection points within 24 hours before the start of the cleaning operation; for each detection point, the system automatically matches the corresponding standard gloss value from the standard gloss database according to the recorded material type and measurement angle, and substitutes it into the formula to calculate the gloss recovery rate of each detection point; the system averages the gloss recovery rate values ​​of all detection points of the same material to obtain the overall gloss recovery rate of that building material; at the same time, according to the area ratio of different materials on the facade, the overall recovery rate of each material is weighted and averaged to obtain the comprehensive gloss recovery rate of the entire building facade.

[0024] In one embodiment of the present invention, step S1, determining the detailed stain residue rate, includes the following steps: Establish a standard image library of ideal clean-state details for various building facade details, including joints, edges, and decorative components of different materials; The detailed area images acquired after cleaning are registered and aligned with the corresponding standard images in the standard image library. The overall difference between the cleaned image and the standard image is calculated as a characterization value for stain residue in this type of detailed area. Specifically: ;in, These are the weighting coefficients. It is a structural similarity index. For the first The actual image after cleaning of the detailed areas. The corresponding image in the standard image library Standard clean images of detailed areas It is an L1 norm. The total number of pixels in a single image; Calculate the average degree of difference across all evaluated detail categories as the final detail stain retention rate. Specifically: ,in, This refers to the number of detailed area categories involved in this assessment.

[0025] Specifically, firstly, a detailed classification system for specific areas should be established, including at least the following: structural types: flat joints, inside corners, outside corners, window frame perimeters, decorative lines, gaps between louver slats, and the junction of drainage pipes and walls; material combinations: glass-aluminum alloy, stone-sealant, paint-metal, and splicing of different stones; and surface treatments: glossy, matte, and textured surfaces. Newly built or perfectly maintained buildings are selected as standard samples; distortion correction and color normalization are performed on the original images, and ROI regions containing complete details are cropped. Then, structured labels (classification system information) are added to each image, and scale-invariant feature transform or directional fast rotation simplified feature descriptors are extracted and stored in the database for subsequent fast matching. After the cleaning operation is completed, systematic sampling and photography are carried out on various detailed areas of the building facade. The system automatically extracts feature points (such as ORB features) from the images acquired after cleaning. Then, in the standard image library, according to the classification labels recorded on site, the corresponding standard images are retrieved, the Hamming distance between the two feature descriptors is calculated, and feature point matching is performed. The Random Sampling Consensus Algorithm (RANSAC) is used to remove erroneous matching point pairs (outside points), and a homography matrix is ​​estimated. Using the calculated homography matrix, the perspective transformation is performed on the cleaned image to make it perfectly aligned with the standard image in terms of geometric structure. After alignment, "histogram matching" or "color constancy algorithm based on Retinex theory" is used to adjust the lighting and overall tone of the cleaned image to be as consistent as possible with the standard image to eliminate color differences caused by different lighting conditions during shooting, and to ensure that subsequent comparisons are mainly based on the stain residue itself. Among them, based on experience, set The value is 0.6.

[0026] In one embodiment of the present invention, step S1, which generates a quantitative value of comprehensive cleaning effect by weighted addition, includes the following steps: Based on the scene attributes and key requirements of the facade, the corresponding weight combination is dynamically selected from the pre-trained weight configuration pattern library; The weight configuration mode library is a classification model trained based on historical successful project data. Its output is a recommended weight coefficient combination of three factors: cleanliness index, gloss recovery rate, and detail stain residue rate. The cleanliness index, gloss recovery rate, and detail stain residue rate were normalized. The normalized values ​​were then weighted and added to the recommended weighting coefficients to obtain a quantitative value for the overall cleaning effect. Specifically: ,in, , and These are the normalized values ​​of cleanliness index, gloss recovery rate, and detail stain residue rate, respectively. , and Recommended weight coefficient combinations.

[0027] Specifically, collect a large amount of historical cleaning project data that has been completed and passed high-quality acceptance. The data package for each project should include: Project Feature Vector: a structured description of the project's "scene attributes" and "key requirements"; where, scene attributes: use classification labels or one-hot encoding, for example, building type = [commercial office, residential community, historical preservation, public facilities], main material = [glass curtain wall, stone, paint, metal], building height level = [low-rise, mid-rise, high-rise, super high-rise]; key requirements: extract the quality focus clearly stated in the project contract or acceptance minutes and encode it as numerical features, for example, visual transparency requirement = [0.1, 1.0], material protection requirement = [0.1, 1.0], long-term cleaning requirement = [0.1, 1.0]; A review panel of 3-5 senior industry experts (e.g., client representatives, supervising engineers, senior project managers) is invited to form a panel. Based on the project's actual performance, focus, and industry experience, they assign what they consider to be the most reasonable weight combination for three indicators: cleanliness index, gloss recovery rate, and detail stain residue rate. This weight set serves as the "supervised learning label" for the sample. The "project feature vector" is used as the input feature, and the "ideal weight label" is used as the output target. A machine learning model combining classification and regression is used for training; in this embodiment, a random forest regressor or a multilayer perceptron is used. The model learns the mapping relationship between project features and the reasonable weights recognized by the experts. The training objective is to minimize the mean squared error between the model's predicted weights and the expert-labeled weights. The trained model, along with the corresponding feature encoder and data normalizer, is packaged and deployed as an online weight configuration pattern library. Inputting new project features will output the predicted recommended weights. In this embodiment... , and The corresponding values ​​are 0.55, 0.3 and 0.15, respectively.

[0028] In one embodiment of the present invention, step S2, the nonlinear mapping function includes the following steps: Preset effect benchmark threshold It also defines two effect fluctuation ranges, including an incentive range and a penalty range; Based on the aforementioned effect fluctuation range, a nonlinear mapping function is constructed to calculate the effect fluctuation coefficient. Specifically: in, This represents the theoretical maximum value of the overall cleaning effect quantification. For the excitation intensity coefficient, The penalty intensity coefficient, and This is the curvature adjustment coefficient.

[0029] Specifically, the performance benchmark threshold is a comprehensive quantitative value representing the minimum qualified standard that the cleaning service should achieve as agreed by both parties. It is the critical point for dividing "incentives" and "penalties". The value is usually determined through negotiation based on industry standards, past cooperation experience or the specific difficulty of the project. For example, for the routine maintenance and cleaning of a regular office building, the performance benchmark threshold can be set at 75.0. in, The excitation intensity coefficient is defined as the factor at which the cleaning effect reaches its theoretical optimum ( = When the maximum positive fluctuation ratio relative to the base price is 0, it represents the highest premium that Party A is willing to pay to incentivize Party B to pursue the ultimate quality; the value ranges from 0 to 0.3; it reflects the project's willingness to pay for "superior quality". The penalty intensity coefficient is defined as the point at which the cleaning effect reaches its theoretical worst ( When =0), the maximum negative fluctuation ratio relative to the base price represents the maximum amount of compensation or discount that Party A can demand when Party B fails to fulfill its basic contractual obligations; the value ranges from 0 to 0.5; it reflects the project's tolerance for the "quality bottom line" and Party A's degree of risk aversion. For projects with high risks (such as the possibility of damaging historical building materials) or serious consequences of quality accidents, the value is higher. in, The curvature adjustment coefficient controls the excitation range. Internally, the price fluctuation coefficient increases with quality improvement. The shape of the growth curve determines the "acceleration" of the reward; if >1 (for example, =1.5): Convex (diminishing marginal returns) excitation, initially (just exceeding) (At the time) the reward growth rate was fast, but as near The reward percentage for each unit improvement gradually decreases to prevent excessive rewards in high-quality areas and encourage the contractor to allocate resources more evenly; if 0 < <1 (for example, =0.8): Initial reward growth is slow, but the closer to the perfect score, the faster the reward growth. This is suitable for strategic projects that require strong motivation for the contractor to overcome technical difficulties and pursue maximum results; if =1, linear incentive, a fixed percentage reward is given for each unit improvement in performance; simple and fair, but lacks additional incentives or constraints for "striving for excellence"; The curvature adjustment coefficient controls the penalty range. Price fluctuation coefficient As quality declines The shape of the decreasing curve determines the "severity variation" of the punishment; if >1 (for example, =1.5): Concave (marginally increasing) penalty; when the effect is just below the benchmark, the penalty is lighter, allowing some tolerance for minor errors; however, when the effect is seriously below the benchmark ( much smaller When 0 < 0, the penalty rate will increase sharply; if 0 < 0, the penalty rate will increase sharply. <1 (for example, =0.7): Concave (diminishing marginal returns) penalty, a heavier penalty is applied as soon as the effect decreases, but the subsequent increase in penalty intensity slows down; if =1, linear incentive, for every unit decrease in effect, a fixed percentage penalty is imposed.

[0030] In one embodiment of the present invention, step S3 includes the following steps: Obtain the static operating parameters of the building facade, quantify each parameter according to the preset classification rules, and form the corresponding static operating condition influence factors. Based on predefined coupling evaluation rules, the interaction between the static operating condition influencing factors is nonlinearly evaluated, and the coupling correction coefficient characterizing the complexity is calculated. The coupling evaluation rule is established by analyzing the correlation between multi-factor combinations and corresponding construction efficiency in historical working condition data; Access the real-time monitoring data stream of the construction process, identify risky operation periods based on predefined risk assessment rules, and calculate dynamic risk adjustment factors based on the proportion of risky operation time to total working hours; The static operating condition influence factor, coupling correction coefficient, and dynamic risk adjustment factor are comprehensively calculated to generate the final operating condition adjustment coefficient.

[0031] Specifically, the main material types of the facade (e.g., glass, granite, aluminum panels, stone paint, etc.) are confirmed through architectural drawings, site surveys, or material samples; key attributes such as surface hardness (Mohs hardness or pencil hardness), chemical resistance (acid and alkali resistance level), and water absorption rate (%) are further collected; the main stain types (dust, rust, oil, moss, paint residue) are determined through preliminary testing and stain sample analysis; their adhesion strength (quantified by an adhesion tester), chemical composition (acidity and alkalinity, organic matter content), and penetration depth (visual inspection or microscopic observation) are assessed; building height, facade inclination angle (vertical, inward, outward), working space (whether there are obstacles such as balconies and awnings), and surrounding environment (whether it is adjacent to main traffic arteries or sensitive facilities) are obtained; the system has a built-in hierarchical quantification table, which converts the above parameters into standardized influence factors according to custom preset rules, including material factors, stain factors, and accessibility factors, assigns weights to different factors, and calculates the initial weighted sum; A large amount of historical project data was collected. Each data sample included: static factor combinations, overall construction time efficiency (standard time / actual time), and actual cost overrun rate. Machine learning methods (decision trees, gradient boosting machines, or neural networks) were used to analyze the data. The goal of model learning was to predict the additional difficulty gain (the part exceeding the weighted sum) caused by a set of static factor inputs. The trained model is essentially a complex, non-linear "coupling function." For ease of deployment and interpretation, its prediction results can be simplified into a multi-dimensional lookup table, i.e., a "coupling rule matrix," as shown in Table 1. Material grade Stain grade Accessibility level Coupling correction coefficient Level 2 (Base difficulty ~1.0) Level 1 (Light stain ~ 0.9) Level 2 (Medium altitude ~ 1.2 meters) 0.00 Level 4 (High Difficulty ~ 1.8) Level 2 (Moderate stain ~ 1.3) Level 3 (High Difficulty ~ 1.6) 0.15 Level 5 (Extremely high difficulty ~ 2.3) Level 3 (Heavy stain ~ 1.7) Level 4 (Extremely high difficulty ~ 2.0) 0.40 Table 1 For the current project, the system queries the above "coupling rule matrix" based on its quantified static factor value (or level) and obtains the corresponding coupling correction coefficient through an interpolation algorithm (nearest neighbor or linear interpolation); Real-time data from the construction process is incorporated to dynamically adjust the pricing to reflect actual risks. During cleaning operations, the following real-time data streams are accessed: environmental monitoring sensors, equipment status monitoring, personnel positioning, and safety equipment. Simultaneously, the system pre-sets a series of risk assessment rules. The system automatically calculates the total duration of "risky work periods" throughout the entire construction cycle, obtains the total planned or actual construction time, calculates the proportion of risky work, and applies a risk sensitivity coefficient, which is set to 0.5 in this embodiment.

[0032] In one embodiment of the present invention, the operating condition adjustment coefficient includes the following steps: in, This is the operating condition adjustment factor. This refers to the number of types of static operating condition parameters. For the first The weighting coefficients of static operating condition factors, For the first Quantitative influencing factors of quasi-static operating condition factors, For coupling correction coefficients, It serves as a dynamic risk adjustment factor.

[0033] Specifically, among them, ,in, For risk sensitivity coefficient, Total man-hours for risky operations Total working hours.

[0034] In one embodiment of the present invention, step S4 includes the following steps: A semantic segmentation model is used to automatically identify and segment areas of different materials on the overall image of the building facade. Based on the segmentation results and the spatial scale calibration information of the image, the cleaning area of ​​each material component is automatically calculated. ,in, For each material type, a baseline unit price library is constructed, where the unit price library is for each material. A preset basic cleaning unit price is established, and the aforementioned working condition adjustment coefficient is used as a difficulty multiplier to initially adjust the basic unit price, resulting in a dynamic sub-item benchmark unit price. ,in, Basic cleaning unit price; Introducing the aforementioned effect fluctuation coefficient, calculate the adjusted price for each material component. Summarize all itemized prices to obtain the final total price. .

[0035] Specifically, the system imports the overall high-definition image of the building facade (usually a panoramic image captured from multiple angles and stitched together) acquired in step S1, performs distortion correction, color equalization, and perspective correction on the image to obtain a facade unfolded image with normalized perspective and uniform scale; the system calls a pre-trained semantic segmentation deep learning model, preferably using the U-Net or DeepLabv3+ architecture, and trains it on a large, labeled dataset of building facade images. The dataset labels include at least: glass, stone (granite / marble), metal (aluminum / stainless steel), paint / stone paint, concrete, windows / window frames, etc.; the pre-processed facade image is input into the model, and the model classifies each pixel in the image and outputs a segmentation mask image. In this image, different colored blocks represent different materials (e.g., blue = glass, gray = stone, silver = metal, beige = paint); morphological operations (such as opening and closing operations) are performed on the segmentation results output by the model to remove noise and small regions, and edge smoothing is performed to obtain the final quality segmentation image; During image acquisition, reference objects with known dimensions on the building facade (such as a standard window with a width of 1.5 meters, or using known floor heights) need to be marked. The system calculates the "pixel-actual size" scaling factor (e.g., 1 pixel = 0.01 square meters) by recognizing the pixel width of these reference objects in the image. The system then iterates through the segmentation mask image and counts the total number of pixels occupied by each material type. ; Calculate the actual cleaning area of ​​each component material based on the scaling factor. , As a scaling factor; the system generates a "Detailed Table of Exterior Facade Material Area", which clearly lists: material type, pixel area, actual area (㎡) and percentage; The benchmark unit price database records the cleaning unit price of different materials under "standard working conditions" (working condition adjustment coefficient value is 1.0).

[0036] In one embodiment of the present invention, step S5 includes the following steps: Based on the quantitative value of the overall cleaning effect and the residual rate of detailed stains, the overall cleaning achievement rate and the detailed cleaning rate are calculated, and a visual verification report containing the percentage indicators, key process data and image comparisons is automatically generated. Calculate the hash value of the verification report and key process data, and upload the hash value, key indicator triples, timestamp, and GPS location information together to form a notarized transaction to the blockchain network to obtain a notarized certificate containing a unique transaction hash TxID; wherein, the key indicator triples ,in, To achieve a comprehensive cleanliness achievement rate, For detailed cleaning rate; Generate an embedded transaction hash TxID and key indicators. Intelligent pricing sheet; A public verification interface based on the transaction hash TxID is provided, supporting the verification of data integrity and consistency through a blockchain explorer.

[0037] Specifically, the system obtains a quantitative value of the overall cleaning effect from step S1, converts it into a percentage, and sets a maximum standard value agreed upon in the industry or by the contract. The overall cleanliness achievement rate was obtained through calculation. The system obtains the detailed stain residue rate from step S1 and calculates the detailed cleaning rate. .

[0038] The above embodiments are only used to illustrate the technical methods of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical methods of the present invention without departing from the spirit and scope of the technical methods of the present invention.

Claims

1. A method for quantitatively evaluating and pricing the cleaning effect of building facades, characterized in that, Includes the following steps: S1: Collect images and physical parameters of the building facade before and after cleaning, calculate the cleanliness index, gloss recovery rate and detail stain residue rate through preset algorithms, and generate a comprehensive cleaning effect quantification value by dynamically weighting and adding them together. S2: Based on the basic quotation for cleaning construction, the quantitative value of the comprehensive cleaning effect is mapped into an effect fluctuation coefficient through a non-linear function; S3: Adjustment coefficient for construction conditions based on facade material, stain type, and construction risk; S4: Identify the area of ​​different materials on the building facade, combine the effect fluctuation coefficient and working condition adjustment coefficient, and perform linkage correction and summary on the benchmark unit price based on the material to generate the final price; S5: Binds the data and calculation results of the entire process to spatiotemporal information, and generates tamper-proof and traceable electronic archives and verification reports through blockchain technology.

2. The method for quantitative evaluation and pricing of building facade cleaning effect according to claim 1, characterized in that, In step S1, the cleanliness index includes the following steps: High-definition imaging equipment was used to collect images of the exterior facade of the same area before and after cleaning under the same lighting and shooting angle. Image processing technology was used to segment and count the total area of ​​stains before and after cleaning. The average gray value of the surface before cleaning and the average gray value of the surface after cleaning are extracted using a gray-level histogram algorithm. The cleanliness index is obtained by determining preset weight coefficients through the analytic hierarchy process (AHP) combined with expert experience, and then weighting and summing the results. Specifically: ,in, and For preset weighting coefficients, This represents the total area of ​​the stain after cleaning. This represents the total area of ​​the stain before cleaning. This represents the average grayscale value of the surface after cleaning. This represents the average grayscale value of the surface before cleaning.

3. The method for quantitative evaluation and pricing of building facade cleaning effect according to claim 1, characterized in that, In step S1, the gloss recovery rate includes the following steps: Obtain the standard gloss value of the exterior building materials in their brand-new state. Then, measure the gloss value before and after cleaning at the same testing point. Calculate the gloss recovery rate by the percentage difference. Specifically: in, Gloss recovery rate This is the gloss value after cleaning. This is the gloss value before cleaning. This refers to the standard gloss value of building materials in their brand-new condition.

4. The method for quantitative evaluation and pricing of building facade cleaning effect according to claim 1, characterized in that, In step S1, the detailed stain residue rate includes the following steps: Establish a standard image library of ideal clean-state details for various building facade details, including joints, edges, and decorative components of different materials; The detailed area images acquired after cleaning are registered and aligned with the corresponding standard images in the standard image library. The overall difference between the cleaned image and the standard image is calculated as a characterization value for stain residue in this type of detailed area. Specifically: ;in, These are the weighting coefficients. It is a structural similarity index. For the first The actual image after cleaning of the detailed areas. The corresponding image in the standard image library Standard clean images of detailed areas It is an L1 norm. The total number of pixels in a single image; Calculate the average degree of difference across all evaluated detail categories as the final detail stain retention rate. Specifically: ,in, This refers to the number of detailed area categories involved in this assessment.

5. The method for quantitative evaluation and pricing of building facade cleaning effect according to claim 1, characterized in that, In step S1, a comprehensive cleaning effect quantification value is generated by weighted addition and fusion, including the following steps: Based on the scene attributes and key requirements of the facade, the corresponding weight combination is dynamically selected from the pre-trained weight configuration pattern library; The weight configuration mode library is a classification model trained based on historical successful project data. Its output is a recommended weight coefficient combination of three factors: cleanliness index, gloss recovery rate, and detail stain residue rate. The cleanliness index, gloss recovery rate, and detail stain residue rate were normalized. The normalized values ​​were then weighted and added to the recommended weighting coefficients to obtain a quantitative value for the overall cleaning effect. Specifically: ,in, , and These are the normalized values ​​of cleanliness index, gloss recovery rate, and detail stain residue rate, respectively. , and Recommended weight coefficient combinations.

6. The method for quantitative evaluation and pricing of building facade cleaning effect according to claim 1, characterized in that, In step S2, the nonlinear mapping function includes the following steps: Preset effect benchmark threshold It also defines two effect fluctuation ranges, including an incentive range and a penalty range; Based on the aforementioned effect fluctuation range, a nonlinear mapping function is constructed to calculate the effect fluctuation coefficient. Specifically: in, This represents the theoretical maximum value of the overall cleaning effect quantification. For the excitation intensity coefficient, The penalty intensity coefficient, and This is the curvature adjustment coefficient.

7. The method for quantitative evaluation and pricing of building facade cleaning effect according to claim 1, characterized in that, Step S3 includes the following steps: Obtain the static operating parameters of the building facade, quantify each parameter according to the preset classification rules, and form the corresponding static operating condition influence factors. Based on predefined coupling evaluation rules, the interaction between the static operating condition influencing factors is nonlinearly evaluated, and the coupling correction coefficient characterizing the complexity is calculated. The coupling evaluation rule is established by analyzing the correlation between multi-factor combinations and corresponding construction efficiency in historical working condition data; Access the real-time monitoring data stream of the construction process, identify risky operation periods based on predefined risk assessment rules, and calculate dynamic risk adjustment factors based on the proportion of risky operation time to total working hours; The static operating condition influence factor, coupling correction coefficient, and dynamic risk adjustment factor are comprehensively calculated to generate the final operating condition adjustment coefficient.

8. The method for quantitative evaluation and pricing of building facade cleaning effect according to claim 7, characterized in that, The operating condition adjustment coefficient includes the following steps: in, This is the operating condition adjustment factor. The number of types of static operating condition parameters, where is the th... The weighting coefficients of static operating condition factors, For the first Quantitative influencing factors of quasi-static operating condition factors, For coupling correction coefficients, It serves as a dynamic risk adjustment factor.

9. The method for quantitative evaluation and pricing of building facade cleaning effect according to claim 1, characterized in that, Step S4 includes the following steps: A semantic segmentation model is used to automatically identify and segment areas of different materials on the overall image of the building facade. Based on the segmentation results and the spatial scale calibration information of the image, the cleaning area of ​​each material component is automatically calculated. ,in, Material type; Construct a benchmark unit price library, whereby the unit price library is for each material. A preset basic cleaning unit price is established, and the aforementioned working condition adjustment coefficient is used as a difficulty multiplier to initially adjust the basic unit price, resulting in a dynamic sub-item benchmark unit price. ,in, Basic cleaning unit price; Introducing the aforementioned effect fluctuation coefficient, calculate the adjusted price for each material component. Summarize all itemized prices to obtain the final total price. .

10. The method for quantitative evaluation and pricing of building facade cleaning effect according to claim 1, characterized in that, Step S5 includes the following steps: Based on the quantitative value of the overall cleaning effect and the residual rate of detailed stains, the overall cleaning achievement rate and the detailed cleaning rate are calculated, and a visual verification report containing the percentage indicators, key process data and image comparisons is automatically generated. Calculate the hash value of the verification report and key process data, and upload the hash value, key indicator triples, timestamp, and GPS location information together to form a notarized transaction to the blockchain network to obtain a notarized certificate containing a unique transaction hash TxID; wherein, the key indicator triples ,in, To achieve a comprehensive cleanliness achievement rate, For detailed cleaning rate; Generate an embedded transaction hash TxID and key indicators. Intelligent pricing sheet; A public verification interface based on the transaction hash TxID is provided, supporting the verification of data integrity and consistency through a blockchain explorer.

Images