A method and system for measuring the volume of recycled aggregate from construction waste

Abstract

The present invention relates to the field of measurement technology, and in particular to a volume measurement method and system for recycled aggregate from construction waste, including removing background point clouds in a scene to obtain three-dimensional point cloud data of aggregates; performing plane fitting on the point clouds of aggregates and correcting spatial positions; designing a logistic regression model of an adaptive grid for calculating the projection area of ​​the point clouds, and obtaining the volume and weight of the aggregates based on the projection area. The present invention solves the problem that the measurement of renewable aggregates by traditional belt scales is limited by the characteristics of the aggregates themselves, resulting in the difficulty of belt scale maintenance and inaccurate measurement results.

Description

Technical Field

[0001] The invention relates to the field of measurement technology, and in particular to a volume measurement method and system for recycled aggregate from construction waste. Background Art

[0002] Volume is an important parameter for analyzing the morphology of objects in space. Once the exact volume of an object is calculated and the density per unit volume of the object is known, the weight of the object can be quickly calculated.

[0003] The volume of regular bodies can be accurately measured using existing calculation formulas. However, since irregular objects have different shapes, for example, the weight of recycled aggregates from construction waste usually needs to be estimated after processing, but the volume of recycled aggregates is irregular in shape, so measuring the volume of irregular recycled aggregates has always been a difficult problem; recycled aggregates are simply referred to as aggregates.

[0004] In the field of computer vision applications, the three-dimensional representation of point cloud data has become one of the current research hotspots and is widely used in fields such as robotics and autonomous driving. For many practical application scenarios, relying on manual calculation of the volume of objects is inefficient and prone to errors. By applying the image processing method of point cloud data to the volume calculation of objects, the manpower input is reduced and the processing efficiency is improved.

[0005] Existing methods for measuring object volume based on 3D point clouds include: convex hull method, slicing method, surface reconstruction method and triangulation method; the convex hull method is only applicable to convex objects, and the measurement error for non-convex objects is large; the surface reconstruction method is affected by the point cloud density, accuracy and the number of generated grids, and the measurement effect for non-convex objects is not ideal, and the surface is prone to overfitting; the slicing method is only applicable to closed point clouds, and the triangulation method needs to fit the point cloud into a surface model. When faced with complex point clouds or non-convex point clouds, the point cloud triangulation is prone to overfitting.

[0006] In addition, existing aggregates are still measured using traditional belt scales, but due to the characteristics of the aggregates themselves, there is a problem of difficulty in maintaining the belt scales. Summary of the invention

[0007] In view of the shortcomings of the existing methods, the present invention solves the problems of using traditional belt scales to measure renewable aggregates, which are limited by the characteristics of the aggregates themselves, resulting in difficult maintenance of the belt scales and inaccurate measurement results, and the existing three-dimensional point cloud method of measuring the volume of objects has large measurement errors.

[0008] The technical solution adopted by the present invention is: a method for measuring the volume of recycled aggregate from construction waste comprises the following steps:

[0009] Step 1: Remove the background point cloud in the scene to obtain the three-dimensional point cloud data of the aggregate;

[0010] Specifically include:

[0011] Step 11, obtaining an RGB-D image of the aggregate;

[0012] Step 12: Segment the pixels occupied by the aggregate to obtain an aggregate mask that is consistent with the width and height of the RGB-D image; create an image D with all pixel values ​​​​0 that is consistent with the width and height of the depth image D ’ , record the index of the pixel position where the pixel value is 1 in the mask as M, and copy the depth value of the position corresponding to M in the depth image D to the image D ’ At the corresponding M position, image D ’ The remaining pixel values ​​are 0;

[0013] Step 13: Initialize the depth image D x, y Axis coordinate array, expand the coordinate array to a two-dimensional grid, and convert the two-dimensional grid to a three-dimensional point cloud.

[0014] As a preferred embodiment of the present invention, the formula for converting a two-dimensional grid into a three-dimensional point cloud is:

[0015] ;

[0016] in, arange Indicates initializing the width and height value sequence from 0 to 1 according to the width and height of the image; width, height is the width and height of the image; X o , Y o Represents each column in the image x Axis coordinates and each row y Axis coordinates; meshgrid Represents a one-dimensional array X o , Y o Combine into a two-dimensional coordinate matrix; depth Indicates the depth value; scale Indicates the scaling factor of the depth value; C x and C y is the principal point coordinate of the camera; f x and f y is the focal length of the camera; stack is the concatenation function according to the channel dimension; X map and Y map is a two-dimensional grid.

[0017] Step 2: plane fitting of the point cloud of the aggregate and correction of the spatial position;

[0018] As a preferred embodiment of the present invention, step 2 specifically includes:

[0019] Step 21: In the aggregate point cloud cloud A part of data points are randomly selected as the internal point set, and the RANSAC The method performs plane fitting according to the set number of iterations to obtain the first fitting plane;

[0020] As a preferred embodiment of the present invention, step 21 specifically includes: constructing an initial fitting plane based on the inner point set, calculating the distance from the remaining data points outside the inner point set to the initial fitting plane, and if the distance is less than a set threshold, updating the remaining data points to the inner point set; repeating the execution, and selecting the plane with the largest number of inner points under the set number of iterations as the first fitting plane.

[0021] Step 22: Calculate two vectors on the first fitting plane v 1. v 2. Calculate the cross product of the two vectors to get the plane normal vector n ;

[0022] Step 23: Set the normal vector to the plane n Vertical and XOY Vectors parallel to the plane v ,by v is the rotation axis of the plane, calculate the plane normal vector n and z Unit vector in the positive direction of the axis m The cosine of the angle between ;

[0023] Step 24, calculate the Rodriguez rotation matrix;

[0024] Step 25: After the point cloud is rotated by the Rodriguez rotation matrix, the point cloud is translated based on the height mean of all points in the point cloud of the first fitting plane, and the first fitting plane is aligned with XOY The planes overlap to obtain a corrected second fitting plane.

[0025] As a preferred embodiment of the present invention, the Rodriguez rotation matrix is ​​as follows:

[0026] ;

[0027] In the formula, I represents the identity matrix, V x Represented as a rotation axis V The antisymmetric matrix of a vector, θ The rotation angle is the angle between the normal vector and the unit vector, a and b is a vectorv of x and y Axis component, H is the Rodriguez rotation matrix.

[0028] Step 3: Design a logistic regression model of adaptive grid to calculate the projection area of ​​point cloud, and obtain the volume and weight of aggregate based on the projection area;

[0029] As a preferred embodiment of the present invention, designing a logistic regression model of an adaptive grid for calculating the point cloud projection area includes:

[0030] Step 31: Select in the second fitting plane m’ points, according to m’ Point x Axis and y The coordinate values ​​of the upper left and lower right points of the axis determine the length and width of the minimum enclosing rectangle;

[0031] Step 32: m’ Take one of the points as the starting point, calculate the point closest to that point, and then traverse the next point one by one until all points are calculated. m’ Find the neighboring points of a point m’ The average distance of the neighboring points of the point; then select m’ points, repeat L times, and get L groups of training data;

[0032] Step 33: training a logistic regression model based on the length and width of the minimum circumscribed rectangle, the average distance between adjacent points, and the manually measured real projection area, outputting the predicted side length of the square grid, thereby obtaining a regression grid;

[0033] Step 34: Project the point cloud to XOY Regression grid on the plane, extract x and y The axis coordinate value is traversed through the number of square grid logical values ​​1 in the regression grid, and the total plane projection area of ​​the point cloud is accumulated.

[0034] As a preferred embodiment of the present invention, obtaining the volume and weight of aggregate based on the projected area includes:

[0035] ;

[0036] in, M is the weight of aggregate, S is the total plane projection area of ​​the point cloud, is the volume of aggregate, is the unit density of aggregate, m’ is the number of points selected in the second fitting plane; For point cloudi The height value of a point.

[0037] As a preferred embodiment of the present invention, a volume measurement system for recycled aggregate from construction waste includes: a memory for storing instructions executable by a processor; and a processor for executing the instructions to implement a volume measurement method for recycled aggregate from construction waste.

[0038] As a preferred embodiment of the present invention, a computer readable medium stores a computer program code, and when the computer program code is executed by a processor, the method for measuring the volume of recycled aggregate from construction waste is implemented.

[0039] Beneficial effects of the present invention:

[0040] 1. Remove aggregate background noise, reduce volume calculation errors, reduce the amount of point cloud data, and improve processing efficiency;

[0041] 2. When calculating the area by ordinary gridding, the grid side length is usually a preset fixed value. Affected by the density, shape and projection area differences of the point cloud, the fixed value has poor universality for all point cloud data, which affects the measurement accuracy of the projection area. The logistic regression model is used to optimize the grid side length, improve robustness and improve calculation accuracy.

[0042] 3. Measure the bottom area of ​​a single point, and then get the volume of the point cloud data by adding and multiplying the bottom area of ​​a single point with the height value of each point. Get the weight of the aggregate based on the known density of the aggregate per unit volume. The method for calculating the weight of the aggregate is simple and convenient for production application. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] Figure 1 The present invention is a flow chart of the method for measuring the volume of recycled aggregate from construction waste. DETAILED DESCRIPTION

[0044] The present invention is further described below in conjunction with the accompanying drawings and embodiments. This figure is a simplified schematic diagram, which only illustrates the basic structure of the present invention in a schematic manner, and therefore it only shows the components related to the present invention.

[0045] like Figure 1 As shown, a method for measuring the volume of recycled aggregate from construction waste comprises the following steps:

[0046] Step 1: Remove the background point cloud in the scene to obtain the three-dimensional point cloud data of the aggregate;

[0047] When using a depth camera to collect RGB-D data of an aggregate stockpile scene, not only the aggregate is collected, but also the surrounding environment, which contains background noise. In order to achieve accurate measurement, the pixels belonging to the aggregate should be focused on from the image perspective. The addition of a cluttered background will affect the calculation of the aggregate volume, thereby affecting the calculation of the final weight of the aggregate.

[0048] Step 11: Obtain aggregate image data, use a depth perception camera to collect the RGB image and depth image D of the aggregate, and perform RGB-D image alignment; wherein, call Intel's official SDK package for image alignment, and use Intel's D435i depth camera as the depth perception camera.

[0049] Step 12: Filter out the valid depth value area where the aggregates in the scene are located, use the open source segmentation model SAM (Segment Anything Mode) to segment the aggregates in the current scene, segment the pixels occupied by the aggregates in the current RGB image, and obtain an aggregate mask that is consistent with the width and height of RGB-D; create an image D with all pixel values ​​​​0 that is consistent with the width and height of the depth image D ’ , record the index of the position of the pixel with a pixel value of 1 in the mask as M, and copy the depth value corresponding to the position in M ​​in the depth image D to the image D ’ At the corresponding M position, image D ’ The remaining pixel values ​​remain 0.

[0050] In this embodiment, SAM segments the pixels occupied by the aggregate in the currently collected image to obtain the aggregate mask, which is consistent with the width and height of RGB-D. The pixel value is composed of a value of either 0 or 1, 1 represents the aggregate part, and 0 represents the background; initialize an image D with a pixel value of all 0 that is consistent with the width and height of the depth image D ’ And 1 index M used to represent the index of the pixel position whose value is all false. M is a logical matrix. Map the index of the pixel position where the pixel value in mask is 1 to M, and change the logical value in M ​​to true. Copy the depth value of the position corresponding to the logical value of true in M ​​in the depth image D to image D. ’ At the corresponding M position, image D ’ The remaining pixel values ​​remain 0.

[0051] Step 13: First, according to the width of the depth image D width ,high height Initialize each column of the depth image D x Axis coordinate array X o , X o is [0,1,2… width]; and each row of y Axis coordinate array Y o , Y o is [0,1,2… height ]; then the coordinate array X o , Y o Extension to 2D grids X map and Y map ; Then convert it to generate dimensions N ×3 point cloud cloud , that is, the three-dimensional point cloud data of the aggregate in the scene is obtained.

[0052] in, N Indicates the number of points, each point is represented by ( point_x , point_y , point_z ) respectively represent the three-dimensional space x Axis coordinates, y Axis coordinates, z The coordinate point on the axis coordinate, the conversion process is expressed by the following formula:

[0053] ;

[0054] in, arange Indicates initializing the width and height value sequence from 0 to 1 according to the width and height of the image; X o , Y o Represents each column in the image x Axis coordinates and each row y Axis coordinates; meshgrid Represents a one-dimensional array X o , Y o Combine into a two-dimensional coordinate matrix; depth Indicates the depth value; scale Indicates the scaling factor of the depth value; C x and C y is the principal point coordinate of the camera, that is, the coordinate of the center of the optical axis; f x and f y is the focal length of the camera, in pixels; stack is the concatenation function according to the channel dimension; scale The scaling factor value is 1000.

[0055] In this embodiment, it is assumed that a one-dimensional array with a dimension of 1×3 X is a one-dimensional array of [1, 2, 3] and 1×2 Y is [4, 5], combining these two one-dimensional arrays into a coordinate matrix, meshgrid ( X , Y ) will return two 2D arrays of dimensions 2×3 X [[1, 2, 3], [1, 2, 3]] and Y [[4, 4, 4], [5, 5, 5]]; Finally, calculate the point cloud data, each point z The coordinate point on the coordinate is obtained by the depth value of the corresponding depth image D pixel after the scaling factor; x , y The coordinate point on the coordinate is calculated by the camera's principal point coordinate and focal length, and the final result is point_x , point_y , point_z All are expressed as being consistent with the width and height of the depth image D width × height A two-dimensional matrix, point_x , point_y , point_z conduct stack , get the point cloud data of each point.

[0056] in, stack Including: First, point_x , point_y , point_z Splice according to the channel dimension to get 1 width × height ×3 three-dimensional matrix, and then change the dimension to ( width × height )×3 two-dimensional matrix, indicating that there are width × height points, each point is represented by ( point_x , point_y , point_z ) respectively represent the three-dimensional space x coordinate, y coordinate, z The coordinate point on the coordinates.

[0057] Step 2: plane fitting of the point cloud of the aggregate and correction of the spatial position;

[0058] Step 21: In the aggregate point cloud cloud A part of data points are randomly selected as the internal point set, and the RANSAC The method performs plane fitting according to the set number of iterations to obtain the first fitting plane;

[0059] Specifically, it includes: constructing an initial fitting plane according to the inner point set, calculating the distance between the remaining data points outside the inner point set and the initial fitting plane, and if the distance is less than a set threshold, updating the remaining data points to the inner point set; repeating the execution, and selecting the plane with the largest number of inner points under the set number of iterations as the first fitting plane;

[0060] In this embodiment, the number of iterations is set to 1000 times and the threshold is 0.05.

[0061] Step 22: First, calculate two vectors on the first fitting plane v 1. v 2. In order to make the two vectors neither collinear nor parallel, select three non-collinear points on the first fitting plane and set them as: P 1( x 1, y 1, z 1) P 2( x 2, y 2, z 2) P 3( x 3, y 3, z 3); then calculate the cross product of the two vectors, that is, the plane normal vector n , the calculation formula is:

[0062] ;

[0063] in, i , j and k Symbols representing unit vectors in three-dimensional space; i express x The unit vector in the direction (1,0,0), j express y The unit vector in the direction (0,1,0), k express z The unit vector in the direction (0,0,1).

[0064] if n The third component of is close to zero, that is, the third component is z axis component, the first fitting plane is considered to be XOY The planes are parallel and the rotation matrix is ​​the identity matrix; if n If the third component of is not close to zero, the required Rodriguez rotation matrix is ​​calculated based on the normal vector.

[0065] Step 23. Next, set the normal vector to the plane n Vertical andXOY Vectors parallel to the plane v ( a ,- b ,0), a and b is a vector v of x and y Axis component; v is the rotation axis of the plane. If the plane rotates counterclockwise by default, the rotation angle is the normal vector n With unit vector m The angle of (0, 0, 1) θ ; Calculate the plane normal vector n and z Unit vector in the positive direction of the axis m The angle of (0, 0, 1) θ The cosine value of is:

[0066] ;

[0067] in, m is a unit vector, n is the plane normal vector.

[0068] Step 24. Next, calculate the Rodriguez rotation matrix, the formula is:

[0069] ;

[0070] In the formula, I represents the identity matrix; V x Represented as a rotation axis V Antisymmetric matrix of vector; H is the Rodriguez rotation matrix, which represents any vector in three-dimensional space around the rotation axis V Vector in counterclockwise direction θ The rotation matrix for angular rotations.

[0071] Step 25: Finally, the point cloud cloud Rodriguez rotation matrix H The mean height of all points in the point cloud after rotation and based on the first fitting plane e (0, 0, - e ) to translate the point cloud and align the first fitting plane with XOY The planes overlap to obtain a corrected second fitting plane.

[0072] Step 3: Design a logistic regression model of adaptive grid to calculate the projection area of ​​point cloud, and obtain the volume and weight of aggregate based on the projection area;

[0073] The logistic regression model of adaptive grid is referred to as logistic regression model (Logistic Regression);

[0074] Step 31: Select in the second fitting plane m’ points, according to m’ Point x Axis and y The coordinate values ​​of the upper left and lower right points of the axis determine the length and width of the minimum enclosing rectangle;

[0075] Step 32: m’ Take one of the points as the starting point, calculate the point closest to the point, and then traverse the next point one by one until all points are calculated. m’ Find the neighboring points of a point m’ The average distance of the neighboring points of the point; then select m’ points, repeat L times, and get L groups of training data;

[0076] Step 33: train a logistic regression model based on the length and width of the minimum circumscribed rectangle, the average distance between adjacent points, and the manually measured real projection area, output the predicted side length of the square grid, and thus obtain a regression grid; wherein the regression grid is composed of a plurality of square grids.

[0077] In this example, m’= 120, calculate 120 points x Axis and y The coordinate values ​​of the upper left and lower right points of the axis are used to determine the minimum enclosing rectangle, thereby obtaining the length and width of the rectangle. At the same time, the distance between each of the 120 points and its nearest neighbor is calculated and the sum is accumulated to obtain the average distance between the 120 points. Repeat the operation to obtain 100 sets of training data. The length and width of the minimum enclosing rectangle and the average distance are used as extracted features, and the real projection area measured manually is used as the real label. The root mean square error is used to train the logistic regression model, and the data set is divided into training set and test set according to 8:2.

[0078] Step 34: Point Cloud cloud Project to XOY Regression grid on the plane, extract x and y axis coordinate value, set the z-axis coordinate value to 0, create a two-dimensional array, mark the square grid containing the point cloud data as a logical value of 1, and calculate the total plane projection area of ​​the point cloud by traversing each square grid in the regression grid and counting the number of square grids with a logical value of 1;

[0079] In this embodiment, it is assumed that n’ points, the coordinates of each point are ( x i,y i , 0), where i From 1 to n’ , create a two-dimensional array, each row represents the coordinates of a point, such as: [[ x 1, y 1], [ x 2, y 2], … [ x n , y n ]], in a two-dimensional array, each row represents a point in the point cloud. XOY Projection on the plane; then mark the square grid containing the point cloud data as a logical value of 1, and calculate the total plane projection area of ​​the point cloud data by traversing each square grid in the grid output by the logistic regression model and counting the number of cells with a logical value of 1.

[0080] Based on point cloud cloud The total plane projection area is used to obtain the bottom area of ​​a single point, and then the volume of the point cloud data is obtained by adding and multiplying the bottom area of ​​a single point and the height value of each point; finally, the weight of the aggregate is obtained according to the known aggregate density per unit volume. The formula is:

[0081] ;

[0082] in, M is the weight of aggregate, S is the total plane projection area of ​​the point cloud, is the volume of aggregate, is the unit density of aggregate, m’ is the number of points selected in the second fitting plane; For point cloud i The height value of each point ( x , y , z ) z The coordinate value of .

[0083] The present invention removes aggregate background noise, reduces volume calculation errors, reduces the amount of point cloud data, and improves processing efficiency. When calculating the area by ordinary gridding, the grid side length is usually a preset fixed value. Affected by the density, shape and projection area differences of the point cloud, the fixed value has poor universality for all point cloud data, which affects the calculation accuracy of the projection area. A logistic regression model is used to optimize the grid side length, improve robustness, and improve calculation accuracy.

[0084] Based on the above ideal embodiments of the present invention, the relevant staff can make various changes and modifications without departing from the technical concept of the present invention through the above description. The technical scope of the present invention is not limited to the contents of the specification, and its technical scope must be determined according to the scope of the claims.

Claims

1. A method for measuring the volume of recycled aggregate from construction waste, characterized in that: The following steps are involved: Step 1: Remove the background point cloud in the scene to obtain the three-dimensional point cloud data of the aggregate; Specifically include: Step 11, obtaining an RGB-D image of the aggregate; Step 12: Segment the pixels occupied by the aggregate to obtain an aggregate mask that is consistent with the width and height of the RGB-D image; create an image D with all pixel values ​​​​0 that is consistent with the width and height of the depth image D ’ , record the index of the pixel position where the pixel value is 1 in the mask as M, and copy the depth value of the position corresponding to M in the depth image D to the image D ’ At the corresponding M position, image D ’ The remaining pixel values ​​are 0; Step 13: Initialize the depth image D x, y Axis coordinate array, expand the coordinate array to a two-dimensional grid, and convert the two-dimensional grid into a three-dimensional point cloud; Step 2: plane fitting of the point cloud of the aggregate and correction of the spatial position; Step 2 specifically includes: Step 21: In the aggregate point cloud cloud A part of data points are randomly selected as the internal point set, and the RANSAC The method performs plane fitting according to the set number of iterations to obtain the first fitting plane; Step 22: Calculate two vectors on the first fitting plane v 1. v 2. Calculate the cross product of the two vectors to get the plane normal vector n ; Step 23: Set the normal vector to the plane n Vertical and XOY Vectors parallel to the plane v ,by v is the rotation axis of the plane, calculate the plane normal vector n and z Unit vector in the positive direction of the axis m The cosine of the angle between ; Step 24, calculate the Rodriguez rotation matrix; Rodriguez rotation matrix, the formula is: In the formula, I represents the identity matrix, V x Represented as a rotation axis V The antisymmetric matrix of a vector, θ The rotation angle is the angle between the normal vector and the unit vector, a and b is a vector v of x and y Axis component, H Rotation matrix for Rodriguez; Step 25: After the point cloud is rotated by the Rodriguez rotation matrix, the point cloud is translated based on the height mean of all points in the point cloud of the first fitting plane, and the first fitting plane is aligned with XOY The planes overlap to obtain a corrected second fitting plane; Step 3: Design a logistic regression model of adaptive grid to calculate the projection area of ​​point cloud, and obtain the volume and weight of aggregate based on the projection area; The logistic regression model of the adaptive grid is designed to calculate the projection area of ​​the point cloud including: Step 31: Select in the second fitting plane m’ points, according to m’ Point x Axis and y The coordinate values ​​of the upper left and lower right points of the axis determine the length and width of the minimum enclosing rectangle; Step 32: m’ Take one of the points as the starting point, calculate the point closest to that point, and then traverse the next point one by one until all points are calculated. m’ Find the neighboring points of a point m’ The average distance of the neighboring points of the point; then select m’ points, repeat L times, and get L groups of training data; Step 33: training a logistic regression model based on the length and width of the minimum circumscribed rectangle, the average distance between adjacent points, and the manually measured real projection area, outputting the predicted side length of the square grid, thereby obtaining a regression grid; Step 34: Project the point cloud to XOY Regression grid on the plane, extract x and y Axis coordinate value, traverse the number of square grids with logical value 1 in the regression grid, and accumulate them to get the total plane projection area of ​​the point cloud; Obtaining the volume and weight of aggregate based on projected area includes: ； in, M is the weight of aggregate, S is the total plane projection area of ​​the point cloud, is the volume of aggregate, is the unit density of aggregate, m’ is the number of points selected in the second fitting plane; For point cloud i The height value of a point.

2. The method for measuring the volume of recycled aggregate from construction waste according to claim 1, characterized in that: The formula for converting a 2D mesh into a 3D point cloud is: ； in, arange Indicates initializing the width and height value sequence from 0 to 1 according to the width and height of the image; width、height is the width and height of the image; X o , Y o Represents each column in the image x Axis coordinates and each row y Axis coordinates; meshgrid Indicates that a one-dimensional array X o , Y o Combine into a two-dimensional coordinate matrix; depth Indicates the depth value; scale Indicates the scaling factor of the depth value; C x and C y is the principal point coordinate of the camera; f x and f y is the focal length of the camera; stack is the concatenation function according to the channel dimension; X map and Y map is a two-dimensional grid.

3. The method for measuring the volume of recycled aggregate from construction waste according to claim 1, characterized in that: Step 21 specifically includes: constructing an initial fitting plane based on the inner point set, calculating the distance from the remaining data points outside the inner point set to the initial fitting plane, and if the distance is less than a set threshold, updating the remaining data points to the inner point set; repeating the process, and selecting the plane with the largest number of inner points under the set number of iterations as the first fitting plane.

4. Construction waste recycled aggregate volume measurement system, characterized in that: include: a memory for storing instructions executable by a processor; A processor, used to execute instructions to implement the method for measuring the volume of recycled aggregate from construction waste as described in any one of claims 1 to 3.

5. A computer readable medium storing a computer program code, characterized in that: When the computer program code is executed by a processor, the method for measuring the volume of recycled aggregate from construction waste as claimed in any one of claims 1 to 3 is implemented.

Images