A construction waste identification and recycling method, device, equipment and storage medium

Abstract

The present invention relates to the technical field of garbage identification, and provides a method, device, equipment and storage medium for identifying and recycling construction waste, including obtaining image data of construction waste and segmenting to obtain image segments; wherein the image data includes construction waste images; processing the image segments to obtain preprocessed images and extracting features to obtain image feature data; matching the image feature data with a preset construction waste feature library to obtain matching feature data and analyzing them to obtain classification results of the matching feature data, using the classification results to identify the construction waste images to obtain recognition results, and inputting the recognition results into a preset recycling decision system to obtain a recycling plan. Through automated processing and intelligent recognition, high-accuracy and rapid classification and recycling of construction waste is achieved, and the problems of being unable to process a large number of construction waste images in real time and having low recognition accuracy for different types of garbage in actual applications are improved.

Description

Technical Field

[0001] The present application relates to the technical field of garbage identification, and in particular to a method, device, equipment and storage medium for identifying and recycling construction waste. Background Art

[0002] With the acceleration of urbanization and the rapid development of the construction industry, the amount of construction waste generated has also increased year by year. Construction waste not only occupies a large amount of land resources, but also causes serious pollution to the environment. Therefore, how to efficiently identify and recycle construction waste has become one of the important topics in the field of environmental protection. Identifying and recycling construction waste in an intelligent and automated way can not only effectively improve resource utilization, but also reduce negative impacts on the environment.

[0003] Among the related technical means, the traditional construction waste identification and recycling method mainly relies on manual sorting and processing. This method is not only inefficient, but also has problems such as high misjudgment rate and high labor intensity. In recent years, with the development of computer vision and machine learning technology, some studies have proposed construction waste recycling methods based on image recognition, which can intelligently identify and recycle construction waste through steps such as image acquisition, preprocessing, feature extraction, classification and recognition. This method has achieved certain results in improving recognition accuracy and processing efficiency.

[0004] Regarding the above technical solution, although the intelligent identification and recycling of construction waste can be achieved through methods based on image recognition and machine learning, in actual application, there are problems such as the inability to process a large number of construction waste images in real time and the low accuracy in identifying different types of waste. Summary of the invention

[0005] In order to improve the problems of being unable to process a large amount of construction waste images in real time and having low accuracy in identifying different types of waste in practical applications, the present application provides a construction waste identification and recycling method, device, equipment and storage medium.

[0006] The present invention provides a construction waste identification and recycling method, comprising: acquiring image data of construction waste, segmenting the image data to obtain a plurality of image segments; wherein the image data includes construction waste images; performing image preprocessing on all the image segments to obtain a preprocessed image of each of the image segments; performing image feature extraction on each of the preprocessed images to obtain image feature data of each of the preprocessed images; using a preset feature matching algorithm to match each image feature data with a preset construction waste feature library to obtain a plurality of matching feature data; performing cluster analysis on all the matching feature data to obtain a classification result of each of the matching feature data; using all the classification results to identify the construction waste image to obtain an identification result, inputting the identification result into a preset recycling decision system to make a recycling decision, obtaining a recycling plan, and recycling the construction waste based on the recycling plan.

[0007] As a preferred solution, the steps of acquiring image data of construction waste, segmenting the image data, and obtaining a plurality of image segments include: acquiring image data of construction waste in real time based on a preset image acquisition device; wherein the image acquisition device includes a high-resolution camera, a lidar, and a depth sensor; preprocessing the image data, inputting the preprocessed image data into a preset image segmentation algorithm, and segmenting the preprocessed image data by image segmentation technology in the image segmentation algorithm to obtain a plurality of image segments.

[0008] As a preferred solution, the step of performing image preprocessing on all the image segments to obtain a preprocessed image of each of the image segments includes: adjusting the brightness of all the image segments according to a preset image brightness adjustment model to obtain a brightness adjusted image; smoothing the brightness adjusted image by Gaussian filtering to obtain a smoothed image and a noise reduced image; performing edge detection on the smoothed image and the noise reduced image to obtain an edge detected image, extracting the image contour in the edge detected image to obtain a contour extracted image; performing color enhancement on the contour extracted image by histogram equalization to obtain a color enhanced image; performing scale transformation on the color enhanced image to obtain a scale change, and performing feature standardization on the color enhanced image based on the scale change to obtain a feature standardized image; performing refinement on the feature standardized image by morphological operation to obtain a refined image and a structural image; performing color space conversion on the refined image and the structural image to obtain a color space image; and performing texture feature enhancement on all the original image segments by using the color space image and the refined image to obtain a preprocessed image of each of the image segments.

[0009] As a preferred embodiment, the step of performing image feature extraction on each of the preprocessed images to obtain image feature data of each of the preprocessed images includes: performing texture feature extraction on all of the preprocessed images according to the feature standardized image to obtain texture feature data and texture description data; performing classification learning on the texture feature data through a deep neural network to obtain classification feature data; performing geometric feature extraction on the classification feature data using geometric transformation to obtain geometric feature data and shape description data; performing vector quantization on the shape description data to obtain vectorized feature data, and performing object positioning on the original preprocessed image based on the vectorized feature data to obtain object positioning data; performing object recognition on the original preprocessed image using the object positioning data and the geometric feature data to obtain object recognition data, and inputting the object recognition data into a preset feature comparison model to obtain image feature data of each of the preprocessed images.

[0010] As a preferred solution, the step of using a preset feature matching algorithm to match each image feature data with a preset construction waste feature library to obtain multiple matching feature data includes: using a feature clustering algorithm to perform manifold learning on the image feature data to obtain local feature data, global feature data and feature distribution data; classifying the local feature data through a support vector machine to obtain classification results and probability evaluation data, and probabilistically fusing the probability evaluation data with the global feature data to obtain probabilistic fusion data; performing pattern recognition on the probability fusion data and the feature distribution data to obtain pattern recognition data, and performing feature matching on the pattern recognition data through a feature matching algorithm to obtain matching results; inputting the matching results into a preset matching evaluation system, and in the matching In the matching evaluation system, a rule inference algorithm is used to compare the similarity between the pattern recognition data and the feature data in the preset construction waste feature library to obtain a preliminary matching result; a preset similarity threshold is used to screen out feature data with a similarity above the preset threshold in the preliminary matching result, and invalid matches below the similarity threshold are removed to obtain a matching feature data set; the matching feature data set is statistically analyzed and processed through multidimensional feature analysis to obtain cluster feature data and abnormal feature data; similarity calculation is performed on the cluster feature data based on a preset feature similarity calculation algorithm to obtain a similarity calculation result; a feature comprehensive evaluation algorithm is used to comprehensively evaluate the similarity calculation result to obtain multiple valid matching feature data, and the multiple valid matching feature data are used as multiple matching feature data.

[0011] As a preferred solution, the step of performing cluster analysis on all the matching feature data to obtain the classification result of each of the matching feature data includes: using a data standardization algorithm to standardize all the matching feature data, normalizing the matching feature data of different scales, and obtaining the matching feature data after standardization; using a dimensionality reduction algorithm to perform dimensionality reduction on the matching feature data after standardization, reducing the dimension of the feature data, retaining the main feature information, and obtaining the matching feature data after dimensionality reduction; using a K-means clustering algorithm to perform preliminary cluster analysis on the feature data after dimensionality reduction, assigning the feature data after dimensionality reduction to multiple cluster centers, and obtaining a number of cluster feature data; classifying each of the cluster feature data based on a preset classification model to obtain a classification label for each cluster feature data, and using the classification label of each cluster feature data as the classification result of each of the matching feature data.

[0012] As a preferred solution, the step of utilizing all the classification results to identify the construction waste image to obtain a recognition result, inputting the recognition result into a preset recycling decision system to make a recycling decision and obtain a recycling plan, and recycling the construction waste based on the recycling plan includes: utilizing a classification result fusion algorithm to fuse all the classification results to obtain a fused classification result and a comprehensive classification result; reconstructing the fused classification result through an image reconstruction algorithm to obtain a recognition image; comparing and analyzing the recognition image with the construction waste image to obtain a recognition result; based on the recognition result, performing a recycling decision analysis on the construction waste image in combination with a preset recycling strategy database to obtain a recycling decision result; inputting the recycling decision result into a preset recycling decision model, performing recycling evaluation and recycling decision based on the preset recycling strategy in the recycling decision model, and generating a recycling plan; performing actual recycling operations on the construction waste based on the recycling plan, and classifying and recycling the construction waste using automated recycling equipment and preset recycling paths.

[0013] The present application also provides a construction waste identification and recycling device, comprising: an acquisition module, used to acquire image data of construction waste, and segment the image data to obtain multiple image segments; wherein the image data includes construction waste images; a processing module, used to perform image preprocessing on all the image segments to obtain a preprocessed image of each of the image segments; an extraction module, used to perform image feature extraction on each of the preprocessed images to obtain image feature data of each of the preprocessed images; a matching module, used to match each image feature data with a preset construction waste feature library using a preset feature matching algorithm to obtain multiple matching feature data; a classification module, used to perform cluster analysis on all the matching feature data to obtain a classification result for each of the matching feature data; a recycling module, used to identify the construction waste image using all the classification results to obtain an identification result, input the identification result into a preset recycling decision system for recycling decision making, obtain a recycling plan, and recycle the construction waste based on the recycling plan.

[0014] The present application also provides an electronic device, including a memory and a processor, wherein the memory stores a computer program that can be run on the processor, and when the processor executes the computer program, it implements any of the above-mentioned construction waste identification and recycling methods.

[0015] The present application also provides a computer-readable storage medium having a computer program stored thereon. When the computer program is executed by a processor, the processor executes the construction waste identification and recycling method as described in any one of the above.

[0016] Compared with the prior art, the present application has the following beneficial effects: rapid classification and accurate identification. Through the automated processing and intelligent identification of construction waste image data, the classification accuracy and sorting efficiency are improved; automated image acquisition and segmentation reduce the amount of manual work and dependence on complex environments, and preprocessing and feature extraction convert the original image data into a form more suitable for computer processing; and the feature matching algorithm combined with the construction waste feature library greatly improves the recognition reliability, and further cluster analysis and decision-making system input realize highly accurate and automated construction waste classification and recycling, improving the problem of being unable to process a large number of construction waste images in real time and the low accuracy of identifying different types of garbage in actual application. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.

[0018] The structures, proportions, sizes, etc. illustrated in the drawings of this specification are only used to match the contents disclosed in the specification so as to facilitate understanding and reading by persons familiar with this technology. They are not used to limit the conditions under which the present invention can be implemented, and therefore have no substantive technical significance. Any structural modification, change in proportion or adjustment of size, without affecting the effects and purposes that can be achieved by the present invention, should still fall within the scope of the technical contents disclosed by the present invention.

[0019] Figure 1 It is a schematic diagram of the flow of the construction waste identification and recycling method provided by an embodiment of the present invention;

[0020] Figure 2 is a schematic block diagram of the structure of a construction waste identification and recycling device provided by an embodiment of the present invention;

[0021] Figure 3 It is a schematic block diagram of the structure of an electronic device provided by an embodiment of the present invention.

[0022] Description of reference numerals:

[0023] 10. Construction waste identification and recycling device; 11. Acquisition module; 12. Processing module; 13. Extraction module; 14. Matching module; 15. Classification module; 16. Recycling module; 20. Electronic equipment; 21. Memory; 22. Processor. DETAILED DESCRIPTION

[0024] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0025] The flowcharts shown in the accompanying drawings are only examples and do not necessarily include all the contents and operations / steps, nor must they be executed in the order described. For example, some operations / steps may also be decomposed, combined or partially merged, so the actual execution order may change according to actual conditions.

[0026] It should also be understood that the terms used in this application specification are only for the purpose of describing specific embodiments and are not intended to limit the application. As used in this application specification and the appended claims, unless the context clearly indicates otherwise, the singular forms "a", "an" and "the" are intended to include plural forms.

[0027] It should be further understood that the term “and / or” used in the specification and appended claims refers to any combination and all possible combinations of one or more of the associated listed items, and includes these combinations.

[0028] The technical solution of the present invention is further described below with reference to the accompanying drawings and through specific implementation methods.

[0029] Embodiment 1:

[0030] like Figure 1 As shown, a construction waste identification and recycling method provided in an embodiment of the present application includes steps S100 to S600.

[0031] Step S100: Acquire image data of construction waste, and segment the image data to obtain a plurality of image segments; wherein the image data includes images of construction waste.

[0032] In this step, a high-resolution camera device is used to take photos or capture real-time video of the piled-up construction waste, so as to obtain image data of the construction waste; specifically, image segmentation algorithms such as threshold segmentation, region growing or edge detection techniques can be used to segment the image data into multiple separate image segments for subsequent image processing and recognition.

[0033] For example, an edge detection algorithm is used to first identify the contours of different objects in a construction waste image, and then the image is cut using these contours to form separate image segments containing each independent construction waste object.

[0034] Step S200: performing image preprocessing on all image segments to obtain a preprocessed image of each image segment.

[0035] In this step, the image segments obtained after segmentation are preprocessed to enhance image quality and reduce interference; specifically, it may include operations such as adjusting brightness and contrast, noise reduction, standardization, normalization processing, and color space conversion.

[0036] For example, the image segments are grayed, and then a median filter is applied to remove image noise, followed by a histogram equalization to improve the image contrast, thereby obtaining a clearer, easier to analyze pre-processed image.

[0037] Step S300: extract image features from each preprocessed image to obtain image feature data of each preprocessed image.

[0038] In this step, image processing technology is used to extract feature data from the preprocessed image; specifically, various image feature extraction algorithms, such as SIFT, SURF, or HOG, can be used to extract information such as image shape, texture, color, and edge features.

[0039] For example, the SIFT algorithm is used to extract key points and corresponding descriptors of the construction waste image, focusing on the local features of the image to form specific image feature data of the preprocessed image.

[0040] Step S400: using a preset feature matching algorithm, match each image feature data with a preset construction waste feature library to obtain a plurality of matching feature data.

[0041] In this step, the extracted image feature data is compared and matched with the data in the feature library; specifically, a preset machine learning algorithm such as support vector machine (SVM), random forest or deep learning model is used to identify and match the corresponding construction waste type.

[0042] For example, a deep convolutional neural network (CNN) model is used for feature matching. After being trained with a large amount of construction waste image data, the model can improve the accuracy of feature data matching and the ability to identify different types of construction waste.

[0043] Step S500: Perform cluster analysis on all matching feature data to obtain a classification result for each matching feature data.

[0044] In this step, the matched feature data are further clustered; specifically, clustering algorithms such as K-Means, hierarchical clustering analysis or DBSCAN are used to group the matched feature data according to feature similarity to achieve automatic classification.

[0045] For example, the K-Means clustering algorithm is used to classify construction waste into different categories, such as concrete blocks, bricks, metal sheets, plastics, etc., based on the texture, color, and shape features in the matching feature data.

[0046] Step S600: Use all classification results to identify the construction waste image to obtain a recognition result, input the recognition result into a preset recycling decision system to make a recycling decision, obtain a recycling plan, and recycle the construction waste based on the recycling plan.

[0047] In this step, all classification results are used comprehensively to make the final identification of the construction waste images; specifically, the classification results are combined with the actual recycling needs, analyzed and processed by the recycling decision system, and a corresponding recycling plan is formulated.

[0048] For example, the recycling decision system evaluates the utilization value and treatment methods of each category of construction waste based on the identification results, and provides reasonable sorting and recycling strategies for reuse, resource processing or harmless disposal.

[0049] In this embodiment, the image data of construction waste is obtained and segmented to generate multiple image segments; then, each image segment is subjected to image preprocessing to generate a preprocessed image; then, image feature extraction is performed on each preprocessed image to obtain image feature data; each image feature data is matched with a preset construction waste feature library through a preset feature matching algorithm to generate matching feature data. These matching feature data are further subjected to cluster analysis to obtain classification results; finally, the construction waste images are identified using these classification results, and the identification results are output, which are input into a preset recycling decision system to form a recycling plan.

[0050] By utilizing image preprocessing and feature extraction methods, the accuracy of identifying the type of construction waste through image data is improved; furthermore, the feature matching algorithm plus the application of the construction waste feature library enables the identification process to accurately match the characteristics of various construction waste, and cluster analysis further optimizes the classification effect. The final automated identification and recycling decision-making system not only improves the speed of construction waste processing, reduces the labor intensity and error rate of manual sorting, but also improves the problems of being unable to process a large number of construction waste images in real time and having low accuracy in identifying different types of waste in actual application.

[0051] Embodiment 2:

[0052] In step S100, image data of construction waste is collected in real time based on a preset image acquisition device; wherein the image acquisition device includes a high-resolution camera, a laser radar, and a depth sensor.

[0053] By installing multiple image acquisition devices at the construction waste treatment site, comprehensive coverage of the on-site work area can be achieved; specifically, high-resolution cameras are used to image various types of construction waste, lidar is used to capture the spatial information of the waste to assist in segmentation, and depth sensors can perceive the volume and depth information of the construction waste.

[0054] For example, a high-resolution camera is installed above the conveyor belt to continuously photograph the construction waste passing through the conveyor belt, while the lidar is responsible for scanning the 3D contour data of the garbage. The data from both are fused to generate detailed image data.

[0055] The image data is preprocessed, and the preprocessed image data is input into a preset image segmentation algorithm. In the image segmentation algorithm, the preprocessed image data is segmented by image segmentation technology to obtain multiple image segments.

[0056] By automatically adjusting the image contrast and clarity and using Gaussian filtering to reduce image noise, a preset segmentation algorithm, such as a threshold-based segmentation algorithm, is further used to perform preliminary image segmentation processing; specifically, the segmentation algorithm is used to identify the boundaries of construction waste, thereby segmenting the continuous image into independent image segments containing individual construction waste.

[0057] For example, the software can automatically analyze the brightness differences in the image to separate construction waste from the background and other waste, and then use the region growing method to refine the segmentation results to obtain clearly defined garbage image fragments.

[0058] In step S200, brightness adjustment is performed on all image segments according to a preset image brightness adjustment model to obtain brightness-adjusted images.

[0059] By introducing an algorithm model to dynamically analyze the current brightness conditions of the image fragments and make real-time adjustments, the image is ensured to be in the optimal brightness state; specifically, the brightness level of each image fragment is adjusted to be within the preset standardized brightness range to improve image quality and the accuracy of subsequent processing.

[0060] For example, using a brightness adjustment method based on histogram analysis, for each image segment, its histogram is calculated and a stretching or compression technique is applied to adjust the brightness, thereby making the visual details of the image clearer.

[0061] The brightness-adjusted image is smoothed by Gaussian filtering to obtain a smoothed image and an image with reduced noise.

[0062] The Gaussian filter algorithm is used to eliminate random noise and small errors in the image and retain important structural features. Specifically, the Gaussian filter kernel is applied to the brightness-adjusted image to smooth the high-frequency noise of the image without affecting the clarity of the image edges and textures.

[0063] For example, appropriate Gaussian kernel parameters are set and the image is subjected to multiple Gaussian filtering processes until the noise is effectively suppressed while maintaining the authenticity and integrity of the image structure.

[0064] Edge detection is performed on the smoothed image and the image with reduced noise to obtain an edge detection image, and the contour of the image in the edge detection image is extracted to obtain a contour extraction image.

[0065] The structural edges in the image are highlighted by edge detection algorithms such as Canny or Sobel operators to facilitate contour extraction; specifically, each pixel point in the image after edge detection is traversed, the pixel points whose edge strength exceeds a specific threshold are marked, and these points are connected to form a continuous contour line.

[0066] For example, the Canny edge detection algorithm is implemented. The image is first processed using Gaussian blur to remove noise, and then the gradient and direction of the image are calculated. Finally, a clear garbage outline is determined and extracted through double threshold detection.

[0067] The contour extraction image is color enhanced by using histogram equalization to obtain a color enhanced image; the color enhanced image is scale transformed to obtain a scale change; and the color enhanced image is feature normalized based on the scale change to obtain a feature normalized image.

[0068] The contrast of the image is improved by the histogram equalization method, making the brightness distribution of the entire image more uniform; specifically, each pixel in the contour extraction image is reallocated according to its brightness value to form a color enhanced image, and then the size and resolution of these images are unified, and then the feature unification processing of the color enhanced image is implemented.

[0069] For example, histogram equalization is performed on the contour image to make the transition from dark areas to bright areas more natural, and then the image is uniformly scaled to a specified size to ensure consistency and comparability of subsequent processing.

[0070] The feature normalized image is refined through morphological operations to obtain a refined image and a structural image.

[0071] By applying morphological operations such as dilation, erosion or opening / closing operations, the structural characteristics of the image are further improved; specifically, the outline and internal details of the construction waste are highlighted through refinement processing, and irrelevant small blocks and spots are removed to obtain an image describing the morphological structure of the garbage.

[0072] For example, an erosion operation is first performed on the feature-normalized image to remove pixels at the object boundary, and then a dilation operation is performed to accurately refine the contours of the garbage, and finally a structural image with rich details that is beneficial for analysis is obtained.

[0073] Performing color space conversion on the refinement image and the structure image to obtain a color space image;

[0074] The color space image and the refined image are used to enhance the texture features of all the original image segments, and a preprocessed image of each image segment is obtained.

[0075] The color space conversion makes the image features more suitable for subsequent analysis and processing; specifically, the color representation of the image is changed, such as converting from RGB to HSV or Lab color space, to emphasize the characteristics of texture and color, and to enhance the texture characteristics of the original image fragment.

[0076] For example, an image is converted from RGB color space to HSV color space to better distinguish color changes and texture details in the image, and then this enhanced texture feature is applied to the original image segment to provide richer information for subsequent feature extraction.

[0077] In step S300, texture features are extracted from all pre-processed images according to the feature normalized images to obtain texture feature data and texture description data.

[0078] The texture information of the image is extracted through advanced image analysis techniques, such as gray-level co-occurrence matrix or local binary pattern; specifically, the texture characteristics of the preprocessed image are analyzed through these techniques and converted into numerical data and descriptive information that are easy to process later.

[0079] For example, the local binary pattern algorithm is used to encode the contrast relationship between each pixel of the preprocessed image and its surrounding pixels, generating feature data representing the texture of construction waste and metadata describing these features.

[0080] The texture feature data is classified and learned through a deep neural network to obtain classification feature data.

[0081] By training deep learning models, such as convolutional neural networks (CNNs), we learn the relationship between the texture features of images and different types of garbage. Specifically, based on a large number of labeled samples, we adjust the network parameters so that the network can effectively classify the input texture feature data and output the classification feature data of the corresponding category.

[0082] For example, a multi-layer deep convolutional network structure is designed and trained using a labeled construction waste image dataset so that the network can learn how to identify different categories of construction waste from texture features.

[0083] The geometric transformation is used to extract the geometric features of the classified feature data to obtain the geometric feature data and shape description data.

[0084] By calculating the geometric properties of different classification feature data, such as area, perimeter, compactness and other geometric indicators, the feature data describing the shape of construction waste are extracted; specifically, through these geometric transformations, the spatial position and size of garbage objects in the image are analyzed and measured to provide more complete shape description information.

[0085] For example, geometric transformation algorithms are used to calculate the contour characteristics of garbage objects in the image, such as polygon approximation or roundness measurement, and these geometric attributes are organized into shape description data.

[0086] The shape description data is vectorized to obtain vectorized feature data, and the original preprocessed image is object located based on the vectorized feature data to obtain object location data.

[0087] The complexity of shape features is simplified by converting shape description data into a vector form suitable for machine learning algorithms; specifically, key points that make a decisive contribution to the description of the object shape are extracted, and these key point data are vectorized to facilitate rapid identification of key parts that constitute the outline of construction waste in the image.

[0088] For example, by calculating the key points on the outline of construction waste and converting their coordinates into vectors, these vectorized data are used to locate objects and quickly identify the exact locations of separated construction waste objects in the image.

[0089] The object positioning data and the geometric feature data are used to perform object recognition on the original preprocessed image to obtain object recognition data, and the object recognition data is input into a preset feature comparison model to obtain image feature data of each preprocessed image.

[0090] Through object recognition technology, such as template matching-based recognition methods, the precise location and type of specific garbage objects in the image can be found; specifically, the object positioning data and geometric feature data are combined and compared with the preset feature comparison model to extract the positioning data with a high degree of match with the known garbage features, and finally obtain accurate image feature information.

[0091] For example, by using the shape template matching method, by comparing the shape features of various types of construction waste in the object positioning data and the feature comparison model, the object positioning points with a high degree of matching are found, thereby identifying the specific type of waste in the image and outputting the corresponding image feature data.

[0092] In step S400, a feature clustering algorithm is used to perform manifold learning on the image feature data to obtain local feature data, global feature data, and feature distribution data.

[0093] By utilizing manifold learning techniques, such as t-SNE or UMAP, the distribution of image feature data in multidimensional space is analyzed to discover the intrinsic structure of the data and the global and local similarities of image features; specifically, the multidimensional image feature data is mapped to a low-dimensional space in order to more intuitively understand the relationship and distribution pattern of junk features.

[0094] For example, the t-SNE algorithm is used to reduce the dimensionality of global feature data, so that similar features are clustered in a low-dimensional space, while different features are dispersed, so that a clear feature clustering distribution can be observed in the low-dimensional graph.

[0095] The local feature data is classified by support vector machine to obtain classification results and probability evaluation data, and the probability evaluation data is probabilistically fused with the global feature data to obtain probabilistic fusion data.

[0096] Through the support vector machine (SVM) algorithm, a decision interface is established to classify the local feature data according to their positions in the feature space; specifically, the optimal decision boundary is found according to the distribution of the local feature data, and classification is performed according to the decision interface, and the classification results and their corresponding probability evaluation data are generated. Then, the global feature data are comprehensively considered and data fusion is performed to obtain a more accurate and robust final classification result.

[0097] For example, an SVM classifier is trained to distinguish different types of construction waste feature vectors, and a probability score is provided for each predicted classification based on the distance from the decision boundary. The statistical probability model of the global feature data is then combined to obtain a more reliable prediction result.

[0098] The probability fusion data and the feature distribution data are subjected to pattern recognition to obtain pattern recognition data, and feature matching is performed on the pattern recognition data through a feature matching algorithm to obtain a matching result.

[0099] Through pattern recognition technology, the similarity between the recognition probability fusion data and the feature distribution data is compared to discover the classification rules of construction waste images; specifically, feature matching algorithms, such as Euclidean distance or cosine similarity, are used to compare the distance or direction between pattern recognition data and known feature distributions, and results with high matching scores are extracted to determine the feature classification of the image.

[0100] For example, the obtained probability fusion data is positioned in the space represented by the feature distribution data, and the distance measurement is used to evaluate its similarity with the cluster centers of each known pattern, and it is classified according to the pattern with the highest matching degree to obtain an exact matching result.

[0101] The matching results are input into a preset matching evaluation system, and a rule inference algorithm is used in the matching evaluation system to compare the similarity between the pattern recognition data and the feature data in the preset construction waste feature library to obtain a preliminary matching result.

[0102] By comparing the pattern recognition data with a large feature library, the most matching type of construction waste is identified; specifically, using rule reasoning algorithms, such as decision trees or fuzzy logic algorithms, the data in the matching evaluation system are subjected to detailed similarity analysis with the samples in the feature library to extract the closest matches and generate preliminary matching results.

[0103] For example, the extracted garbage image feature vector is input into a pre-trained decision tree model. The model compares the features of each category of construction waste in the feature library layer by layer according to each element of the feature vector, and gives the type of construction waste with the highest matching degree.

[0104] The preset similarity threshold is used to filter out feature data with similarity above the preset threshold in the preliminary matching results, and invalid matches below the similarity threshold are removed to obtain a matching feature data set. The matching feature data set is statistically analyzed and processed through multidimensional feature analysis to obtain cluster feature data and abnormal feature data.

[0105] By setting a similarity threshold, data with a low match degree compared with known construction waste characteristics are removed; specifically, the preliminary matching results are screened, and the data exceeding the similarity threshold is retained and regarded as valid. The matching feature data set obtained is processed using multidimensional statistical analysis to distinguish normal clustering features from abnormal features.

[0106] For example, if the similarity score between the matching feature data and a feature in the feature library exceeds the set threshold (such as 85%), it is considered a valid match and is included in the clustering feature data set; conversely, if the score is lower than the threshold, it is determined to be abnormal feature data and is excluded.

[0107] Based on a preset feature similarity calculation algorithm, similarity calculation is performed on the cluster feature data to obtain a similarity calculation result, and a feature comprehensive evaluation algorithm is used to comprehensively evaluate the similarity calculation result to obtain multiple valid matching feature data, and the multiple valid matching feature data are used as multiple matching feature data.

[0108] The similarity between different feature data is calculated through an efficient algorithm, and the similarity results are evaluated as a whole; specifically, a feature similarity calculation algorithm, such as cosine similarity calculation, is used to analyze each group of cluster feature data, and a feature comprehensive evaluation algorithm, such as weighted scoring method, is used to score each group of results, thereby extracting effective matching feature data with high scores.

[0109] For example, the similarity between the cluster feature data is calculated, and different weights are assigned according to the importance of the features. The weighted scoring method is used to evaluate the influence of each feature, and finally those features with high scores are screened out as valid data sets for determining matching.

[0110] In step S500, all matching feature data are standardized using a data standardization algorithm, and matching feature data of different scales are normalized to obtain standardized matching feature data.

[0111] Through standardization processing, all the obtained matching feature data are converted to the same scale to facilitate comparison and analysis between data; specifically, each feature data is converted into a dimensionless form using methods such as Z-score standardization or minimum-maximum standardization to ensure consistency in data processing.

[0112] For example, all matching feature data are Z-score standardized, the deviation of each feature value relative to its mean is calculated and divided by the standard deviation, and all data are mapped to the same standard deviation and mean range to ensure that the data are comparable in subsequent processing.

[0113] The dimension reduction algorithm is used to perform dimension reduction processing on the standardized matching feature data, reduce the dimension of the feature data, retain the main feature information, and obtain the matching feature data after dimension reduction.

[0114] Dimensionality reduction technology is used to reduce the complexity of data and improve algorithm efficiency. Specifically, dimensionality reduction methods such as principal component analysis (PCA) or linear discriminant analysis (LDA) are used to remove redundant and unimportant features and retain the main features that have an important impact on classification decisions.

[0115] For example, PCA is used to transform the obtained standardized matching feature data into a new feature space, retaining the feature components with the largest variance as the main features, while removing some unimportant components. The resulting dimensionality-reduced data set has a lower dimension and still contains sufficient information for effective classification.

[0116] The K-means clustering algorithm is used to perform preliminary clustering analysis on the feature data after dimensionality reduction, and the feature data after dimensionality reduction are assigned to multiple cluster centers to obtain several cluster feature data.

[0117] The K-means cluster analysis is used to assign the reduced-dimensional data points to the nearest cluster center, thereby classifying the feature data. Specifically, the number of clusters is set in advance, the distance between each feature data point and each cluster center is calculated, and mapped to the nearest cluster center, and gradually adjusted until each cluster center is stabilized.

[0118] For example, after determining the K value, the dimensionality-reduced matching feature data set is iteratively processed, and the data points are assigned to different cluster centers according to the Euclidean distance from each cluster center. The iterations are repeated until the clusters no longer change, thereby obtaining a series of cluster feature data representing different types of construction waste.

[0119] Each cluster feature data is classified based on a preset classification model to obtain a classification label for each cluster feature data, and the classification label for each cluster feature data is used as a classification result for each matching feature data.

[0120] Through a pre-trained classification model, such as a random forest or neural network model, a specific category label is assigned to each cluster feature data; specifically, the classification model is used to perform classification prediction on each data point in the cluster feature data set, and the predicted result label is combined with the original matching feature data as the final classification result.

[0121] For example, each cluster feature data obtained by the K-means algorithm is input into the previously trained random forest classifier, and the model is used to perform detailed analysis and label assignment on each cluster result. The data point of each cluster obtains an exact classification label based on the classification probability output by the model.

[0122] In step S600, all classification results are fused using a classification result fusion algorithm to obtain a fused classification result and a comprehensive classification result.

[0123] By aggregating the output results of multiple classification algorithms, more accurate and comprehensive garbage classification decisions can be obtained; specifically, ensemble learning or fusion methods, such as voting or stacking, are used to integrate and weigh different classification results to obtain the final fusion classification result.

[0124] For example, a multi-algorithm integration method is used to integrate the classification results of different models, and the final category label of each garbage category is determined according to the majority voting principle, thereby improving the accuracy and stability of the overall classification.

[0125] The fusion classification result is reconstructed through the image reconstruction algorithm to obtain the recognition image.

[0126] Image processing technology is used to combine the identified classification labels with the original image to reconstruct the recognition image containing the classification information. Specifically, image reconstruction algorithms are applied, such as reconstruction technology based on pixel-level segmentation, to add classification label information to the corresponding image area to generate the final recognition image, showing clear classification boundaries and logos.

[0127] For example, a bounding box of different colors is added to each image segment in the preprocessed image according to its classification label, and the reconstructed image clearly shows the type and location of each piece of construction waste.

[0128] The recognition image is compared and analyzed with the construction waste image to obtain the recognition result.

[0129] The accuracy and reliability of recognition are verified by analyzing the consistency and differences in features and structures between the recognized image and the original construction waste image; specifically, the shapes and features of each marked area in the recognized image are compared with the actual construction waste objects in the image to confirm the correctness of classification and recognition.

[0130] For example, compare the construction waste image with type identification obtained through image reconstruction with the actual construction waste distribution in the scene, check whether each classification mark is consistent with the characteristics of the real object, and record the classification success rate and misidentification as feedback on the recognition result.

[0131] Based on the recognition results, the recycling decision analysis of the construction waste images is carried out in combination with the preset recycling strategy database to obtain the recycling decision results.

[0132] Further analysis and decision-making are carried out on the identification results according to specific recycling strategies and relevant regulations; in particular, the treatment methods and resource values ​​of different construction wastes provided by the recycling strategy database are used to formulate appropriate recycling or treatment plans for each type of construction waste.

[0133] For example, by utilizing the recycling policies and guidelines for different construction waste materials such as concrete, metal, and wood in the recycling strategy database and combining them with image recognition results, specific recycling processes and treatment methods can be developed for each type of waste, as well as subsequent resource utilization pathways, to form a practical recycling strategy plan.

[0134] The recycling decision results are input into a preset recycling decision model, in which recycling evaluation and recycling decision are performed based on the preset recycling strategy to generate a recycling plan.

[0135] By introducing professional recycling knowledge and optimization algorithms into the decision-making model, we ensure that the generated recycling plan is economical, effective and environmentally friendly. Specifically, by combining the cost-benefit analysis, environmental impact assessment and resource recovery value calculation functions in the recycling decision model, we comprehensively consider the identification results based on the recycling strategy and output the optimized recycling plan.

[0136] For example, based on the different types of construction waste identified, a recycling decision model is used to evaluate the processing costs, potential resource recovery value and environmental benefits of various types of waste. These factors are combined to design a set of specific operation paths and recycling methods to achieve the best economic and environmental results.

[0137] Carry out actual recycling operations on construction waste based on the recycling plan, and use automated recycling equipment and preset recycling routes to classify and recycle construction waste.

[0138] According to the recycling plan obtained, mechanization and automation technology are used to implement the specific recycling operations of construction waste; specifically, automatic sorting systems such as robotic arms and conveyor belts are used in combination with computer control programs to sort, transport and centrally process various types of construction waste according to predetermined paths and methods.

[0139] For example, according to the guidance in the recycling plan, a robotic arm is set up on the automatic sorting line to perform precise sorting based on the recognition image, and then different categories of construction waste are transported to designated processing areas or recycling containers, ultimately achieving effective and automated resource utilization of construction waste.

[0140] In this embodiment, by combining advanced image processing, machine learning and data analysis technologies, it is possible to accurately and efficiently identify and recycle construction waste from the acquired construction waste image data. First, various types of image acquisition devices are used to acquire construction waste images in real time, and then the image data is refined through multi-level preprocessing and feature extraction. Next, deep learning and clustering algorithms are used to effectively analyze and classify the feature data. Finally, an economical and environmentally friendly recycling plan is generated by combining the recycling strategy database and the recycling decision model, and the actual construction waste classification and recycling is carried out in conjunction with automated equipment. The whole process not only improves the recycling efficiency and reduces labor intensity, but also promotes the resource utilization of construction waste, which is conducive to environmental protection and sustainable development.

[0141] Embodiment 3:

[0142] like Figure 2 As shown, the present application also provides a construction waste identification and recycling device 10 , which includes an acquisition module 11 , a processing module 12 , an extraction module 13 , a matching module 14 , a classification module 15 and a recycling module 16 .

[0143] The acquisition module 11 is mainly used to acquire image data of construction waste and segment the image data to obtain multiple image segments; wherein the image data includes images of construction waste.

[0144] The processing module 12 is mainly used to perform image preprocessing on all image segments to obtain a preprocessed image of each image segment.

[0145] The extraction module 13 is mainly used to extract image features from each pre-processed image to obtain image feature data of each pre-processed image.

[0146] The matching module 14 is mainly used to match each image feature data with a preset construction waste feature library using a preset feature matching algorithm to obtain a plurality of matching feature data.

[0147] The classification module 15 is mainly used to perform cluster analysis on all matching feature data to obtain a classification result for each matching feature data.

[0148] The recycling module 16 is mainly used to identify the construction waste image using all the classification results, obtain the identification result, input the identification result into a preset recycling decision system to make a recycling decision, obtain a recycling plan, and recycle the construction waste based on the recycling plan.

[0149] In this embodiment, the acquisition module 11 is responsible for collecting and segmenting the image data of construction waste; the processing module 12 pre-processes the segmented image segments to optimize the image quality and prepare for subsequent feature extraction; the extraction module 13 extracts key image feature data from the pre-processed image; the matching module 14 matches the extracted features with the pre-established construction waste feature library to find the most similar feature data; the classification module 15 performs cluster analysis on the garbage type according to the matching results, so as to achieve accurate classification; the recycling module 16 makes recycling decisions according to the classification results and guides the actual recycling process. The device has a high degree of automation and is easy to operate. It can improve the efficiency and accuracy of construction waste sorting and recycling, and is conducive to resource recycling and environmental protection.

[0150] It should be noted that technicians in the relevant technical field can clearly understand that for the convenience and simplicity of description, the specific working process of the above-described device and each module can refer to the corresponding process in the construction waste identification and recycling method of the aforementioned Example 1, and will not be repeated here.

[0151] Embodiment 4:

[0152] like Figure 3 As shown, the present application also provides an electronic device 20, including a memory 21 and a processor 22, the memory 21 stores a computer program that can be run on the processor 22, and the processor 22 implements the construction waste identification and recycling method of Example 1 when executing the computer program.

[0153] In this embodiment, the construction waste identification and recycling method described in Embodiment 1 is implemented by an electronic device 20. The electronic device 20 includes a memory 21 and a processor 22, wherein the memory 21 stores a computer program that can be run on the processor 22. When the processor 22 executes the computer program, it can sequentially complete multiple steps such as image acquisition, image preprocessing, feature extraction, feature matching, cluster analysis and recycling decision of construction waste. The electronic device 20 can be applied to various intelligent terminals or industrial automation equipment to realize intelligent identification and automatic recycling of construction waste, further improve the intelligence level of the recycling process, reduce labor costs, and optimize resource allocation efficiency.

[0154] Embodiment 5:

[0155] The present application also provides a computer-readable storage medium on which a computer program is stored. When the computer program is executed by a processor, the processor executes the construction waste identification and recycling method as in Example 1.

[0156] In this embodiment, a computer-readable storage medium is provided, and a computer program of the construction waste identification and recycling method described in Example 1 is embedded therein. The storage medium can be in the form of a hard disk drive, a solid-state hard disk, a memory card, a USB flash drive, or a CD. When the computer program is loaded and run on the processor, it instructs the processor to execute a complete process of construction waste image recognition and classification, including but not limited to a series of steps such as image acquisition, image processing, feature extraction, matching comparison, and recycling analysis. This computer-readable storage medium is easy to transport and deploy, and can be easily integrated into different hardware platforms and systems, bringing flexibility and scalability to modern intelligent recycling systems, and supporting environmental protection projects and intelligent upgrades of various sizes.

[0157] As described above, the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features thereof may be replaced by equivalents. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A construction waste identification and recycling method, characterized in that: include: Acquire image data of construction waste, and segment the image data to obtain a plurality of image segments; wherein the image data includes images of construction waste; Performing image preprocessing on all the image segments to obtain a preprocessed image of each of the image segments; Performing image feature extraction on each of the preprocessed images to obtain image feature data of each of the preprocessed images; The image feature data is subjected to manifold learning by using a feature clustering algorithm to obtain local feature data, global feature data, and feature distribution data; the local feature data is classified by using a support vector machine to obtain classification results and probability evaluation data, and the probability evaluation data is probabilistically fused with the global feature data to obtain probability fusion data; the probability fusion data and the feature distribution data are subjected to pattern recognition to obtain pattern recognition data, and the pattern recognition data is subjected to feature matching by using a feature matching algorithm to obtain matching results; the matching results are input into a preset matching evaluation system, and the pattern recognition data is matched with a preset construction waste feature distribution data by using a rule inference algorithm in the matching evaluation system. The similarities between the feature data in the feature database are compared to obtain a preliminary matching result; the feature data with similarities above the preset threshold in the preliminary matching result are screened out using a preset similarity threshold, and invalid matches below the similarity threshold are removed to obtain a matching feature data set; the matching feature data set is statistically analyzed and processed by multidimensional feature analysis to obtain cluster feature data and abnormal feature data; similarity calculation is performed on the cluster feature data based on a preset feature similarity calculation algorithm to obtain a similarity calculation result; the similarity calculation result is comprehensively evaluated using a feature comprehensive evaluation algorithm to obtain multiple valid matching feature data, and the multiple valid matching feature data are used as multiple matching feature data; Performing cluster analysis on all the matching feature data to obtain a classification result for each of the matching feature data; The construction waste image is identified using all the classification results to obtain an identification result, the identification result is input into a preset recycling decision system to make a recycling decision, a recycling plan is obtained, and the construction waste is recycled based on the recycling plan.

2. The construction waste identification and recycling method according to claim 1, characterized in that: The step of acquiring image data of construction waste and segmenting the image data to obtain a plurality of image segments comprises: Based on a preset image acquisition device, image data of construction waste is collected in real time; wherein the image acquisition device includes a high-resolution camera, a laser radar and a depth sensor; The image data is preprocessed, and the preprocessed image data is input into a preset image segmentation algorithm, in which the preprocessed image data is segmented by an image segmentation technique to obtain a plurality of image segments.

3. The construction waste identification and recycling method according to claim 1, characterized in that: The step of performing image preprocessing on all the image segments to obtain a preprocessed image of each of the image segments comprises: Adjusting the brightness of all the image segments according to a preset image brightness adjustment model to obtain an image after brightness adjustment; Smoothing the brightness-adjusted image by Gaussian filtering to obtain a smoothed image and an image with reduced noise; Performing edge detection on the smoothed image and the noise-reduced image to obtain an edge-detected image, and extracting the contour of the image in the edge-detected image to obtain a contour-extracted image; Performing color enhancement on the contour extraction image by using histogram equalization to obtain a color enhanced image; performing scale transformation on the color enhanced image to obtain a scale change, and performing feature normalization on the color enhanced image based on the scale change to obtain a feature normalized image; Performing a thinning process on the characteristic standardized image through morphological operations to obtain a thinned image and a structural image; Performing color space conversion on the refined image and the structural image to obtain a color space image; The color space image and the refined image are used to perform texture feature enhancement on all the original image segments to obtain a preprocessed image of each image segment.

4. The construction waste identification and recycling method according to claim 3 is characterized in that: The step of extracting image features from each of the preprocessed images to obtain image feature data of each of the preprocessed images comprises: Extracting texture features from all the preprocessed images according to the feature standardized images to obtain texture feature data and texture description data; Classify and learn the texture feature data through a deep neural network to obtain classification feature data; Using geometric transformation to extract geometric features from the classification feature data, to obtain geometric feature data and shape description data; Performing vector quantization on the shape description data to obtain vectorized feature data, and performing object positioning on the original preprocessed image based on the vectorized feature data to obtain object positioning data; The object positioning data and the geometric feature data are used to perform object recognition on the original pre-processed image to obtain object recognition data, and the object recognition data is input into a preset feature comparison model to obtain image feature data of each pre-processed image.

5. The construction waste identification and recycling method according to claim 1, characterized in that: The step of performing cluster analysis on all the matching feature data to obtain a classification result for each matching feature data includes: All matching feature data are standardized using a data standardization algorithm, matching feature data of different scales are normalized to obtain standardized matching feature data; Performing dimensionality reduction processing on the standardized matching feature data using a dimensionality reduction algorithm to reduce the dimension of the feature data, retain the main feature information, and obtain the matching feature data after dimensionality reduction; Performing a preliminary clustering analysis on the feature data after dimensionality reduction by using a K-means clustering algorithm, assigning the feature data after dimensionality reduction to a plurality of cluster centers, and obtaining a plurality of cluster feature data; Each of the cluster feature data is classified based on a preset classification model to obtain a classification label for each cluster feature data, and the classification label for each cluster feature data is used as a classification result for each of the matching feature data.

6. The construction waste identification and recycling method according to claim 1, characterized in that: The steps of using all the classification results to identify the construction waste image to obtain an identification result, inputting the identification result into a preset recycling decision system to make a recycling decision, obtaining a recycling plan, and recycling the construction waste based on the recycling plan include: Using a classification result fusion algorithm to fuse all the classification results, to obtain a fused classification result and a comprehensive classification result; Reconstructing the fusion classification result by an image reconstruction algorithm to obtain a recognition image; Compare and analyze the recognition image with the construction waste image to obtain a recognition result; Based on the recognition result, a recycling decision analysis is performed on the construction waste image in combination with a preset recycling strategy database to obtain a recycling decision result; Inputting the recycling decision result into a preset recycling decision model, performing recycling evaluation and recycling decision based on a preset recycling strategy in the recycling decision model, and generating a recycling plan; Based on the recycling scheme, actual recycling operations are carried out on the construction waste, and the construction waste is classified and recycled using automated recycling equipment and preset recycling paths.

7. A construction waste identification and recycling device, characterized in that: include: An acquisition module, used to acquire image data of construction waste, and segment the image data to obtain a plurality of image segments; wherein the image data includes images of construction waste; A processing module, used for performing image preprocessing on all the image segments to obtain a preprocessed image of each of the image segments; An extraction module, used for performing image feature extraction on each of the preprocessed images to obtain image feature data of each of the preprocessed images; The matching module is used to perform manifold learning on the image feature data using a feature clustering algorithm to obtain local feature data, global feature data and feature distribution data; classify the local feature data using a support vector machine to obtain classification results and probability evaluation data, and perform probability fusion on the probability evaluation data and the global feature data to obtain probability fusion data; perform pattern recognition on the probability fusion data and the feature distribution data to obtain pattern recognition data, and perform feature matching on the pattern recognition data using a feature matching algorithm to obtain matching results; input the matching results into a preset matching evaluation system, and use a rule inference algorithm in the matching evaluation system to compare the pattern recognition data with a preset construction data. The similarities between the feature data in the garbage feature library are compared to obtain a preliminary matching result; the feature data with similarities above the preset threshold in the preliminary matching result are screened out using a preset similarity threshold, and invalid matches below the similarity threshold are removed to obtain a matching feature data set; the matching feature data set is statistically analyzed and processed by multidimensional feature analysis to obtain clustering feature data and abnormal feature data; similarity calculation is performed on the clustering feature data based on a preset feature similarity calculation algorithm to obtain a similarity calculation result; the similarity calculation result is comprehensively evaluated using a feature comprehensive evaluation algorithm to obtain multiple valid matching feature data, and the multiple valid matching feature data are used as multiple matching feature data; A classification module, used for performing cluster analysis on all the matching feature data to obtain a classification result for each of the matching feature data; The recycling module is used to use all the classification results to identify the construction waste image, obtain an identification result, input the identification result into a preset recycling decision system to make a recycling decision, obtain a recycling plan, and recycle the construction waste based on the recycling plan.

8. An electronic device, characterized in that: It comprises a memory and a processor, wherein the memory stores a computer program that can be run on the processor, and when the processor executes the computer program, the method for identifying and recycling construction waste as described in any one of claims 1 to 6 is implemented.

9. A computer-readable storage medium, characterized in that: A computer program is stored thereon, and when the computer program is executed by a processor, the processor is enabled to execute the construction waste identification and recycling method as described in any one of claims 1 to 6.

Images