A method for laying tiles based on structured light projection guidance

Abstract

The application provides a paving brick method based on structured light projection guidance, a control execution mechanism is used to place a current brick in a position close to a reference brick and a side column brick; a laser projection mechanism one and a laser projection mechanism two are respectively used to form a projection pattern on the current brick, the reference brick and / or the side column brick and collect images; coarse positioning processing and fine positioning processing are performed on brick edge features in the projection pattern to identify the contour of the current brick, the reference brick and / or the side column brick; the target position of the current brick is calculated through the contour of the reference brick and / or the side column brick and the width of the paving brick joint, the angle deviation value and the distance deviation value of the current brick and the target position of the current brick are calculated, and the execution mechanism is controlled by using the angle deviation value and the distance deviation value to pave the current brick. The application has strong anti-interference and robustness, can guide the execution mechanism to complete high-precision brick paving under a harsh construction environment, and can ensure high-precision joint width control and edge and corner alignment, thereby meeting the precision requirement of the paving brick joint.

Description

Technical Field

[0001] This invention relates to the field of floor tile laying technology, and more specifically, to a method for laying floor tiles based on structured light projection guidance. Background Technology

[0002] Tile laying is a crucial step in home renovation, and its quality directly affects the aesthetics, flatness, and durability of the floor decoration.

[0003] Currently, tile laying still relies primarily on manual labor, and the construction environment has the following typical characteristics: high dust concentration at the construction site, affecting the respiratory health and visibility of workers; complex lighting conditions, often with shadows or uneven lighting, which is not conducive to accurate positioning; and limited working space, especially in areas such as kitchens, bathrooms, balconies, and corridors, making it inconvenient for personnel to move around. Under these conditions, the traditional manual laying method generally adopts the following steps: first, the ground base is leveled, then cement mortar or tile adhesive is applied, and then the tiles are manually moved and placed in the predetermined position. Tools such as rubber mallets and spirit levels are used to tap, press, and fine-tune the tiles to control flatness, joint width, and alignment accuracy.

[0004] However, this experience-based construction method has many inherent drawbacks:

[0005] 1. The tiling process relies on the hand-eye coordination and operational feel of skilled workers. The tiling speed of workers is limited, which not only restricts the efficiency of tiling, but also results in high labor intensity and high labor costs.

[0006] 2. The paving quality is unstable, making it difficult to ensure the consistency of the width of each brick joint and the precise alignment of the brick edges and corners. Defects such as misaligned joints and height differences are prone to occur, and the paving quality cannot be guaranteed.

[0007] 3. The construction site has prominent safety hazards such as dust and confined spaces, which pose a threat to the personal safety of workers. Summary of the Invention

[0008] The purpose of this invention is to overcome the shortcomings and deficiencies of the prior art and provide a method for laying floor tiles based on structured light projection guidance. This method has strong anti-interference capabilities and can guide the actuator to complete high-precision floor tile laying in harsh construction environments, while ensuring high-precision joint width control and corner alignment, thereby meeting the requirements for tile joint accuracy.

[0009] To achieve the above objectives, the present invention is implemented through the following technical solution: a method for laying floor tiles based on structured light projection guidance, characterized in that: an execution mechanism for grasping the current tile is provided, and a laser projection mechanism one, a camera one, a laser projection mechanism two, and a camera two are installed on the execution mechanism;

[0010] The method for laying floor tiles is as follows: The control actuator places the current tile close to the reference tile and / or adjacent tiles; Laser projection mechanism one and laser projection mechanism two are activated, and they respectively form projection patterns on the current tile, reference tile, and / or adjacent tiles, and images are acquired by camera one and camera two respectively; In the acquired images, coarse positioning processing and fine positioning processing are performed on the tile edge features in the projection pattern to identify the outlines of the current tile, reference tile, and / or adjacent tiles; The target position for laying the current tile is calculated based on the outlines of the reference tile and / or adjacent tiles and the width of the tile joints, and the angular deviation value and distance deviation value between the current tile and the target position are calculated. The angular deviation value and distance deviation value are used to control the actuator to lay the current tile.

[0011] Specifically, in the acquired image, the coarse localization processing of the brick edge features in the projection pattern refers to: obtaining the mask of the projection pattern and an image containing only the projection pattern itself; identifying the brick edge features in the image containing only the projection pattern itself in the mask; obtaining the brick corner feature points of the reference brick, the adjacent bricks, and the current brick; fitting these features to obtain the horizontal and vertical edges corresponding to the brick corners; and outputting the set of feature points of the brick corners and corresponding brick edges of the reference brick, the adjacent bricks, and the current brick as the result of coarse localization.

[0012] To elaborate further, it includes the following steps:

[0013] The first step is to perform spatial preprocessing on the original image Img0 to obtain image Img1;

[0014] The second step is to perform frequency domain preprocessing on image Img1 to obtain image Img2;

[0015] The third step is to extract the effective region of the projection pattern from the image Img1 using the brightness binarization threshold segmentation method, and then fit the minimum circumscribed convex polygon of the effective region to obtain the mask region of the projection pattern.

[0016] The fourth step is to binarize the image Img2, extract the highlight parts, and segment out the image Img3, which contains only the projection pattern itself.

[0017] The fifth step is to retrieve the edge feature points of the brick in the projection pattern: the abrupt change features in the projection pattern are used as the edge feature pattern of the brick; in the image Img3, all pixels in the masked area are traversed, and it is determined whether the feature of the pixel area matches the edge feature pattern. If they match, the pixel is added to the feature point set PointList.

[0018] The sixth step is to use the iterative model parameter estimation algorithm RANSAC to fit the brick corners of all feature points in PointList, and obtain the reference brick corner feature points, the side column brick corner feature points, and the current brick corner feature points.

[0019] Step 7: Traverse the feature points of the reference brick corner, the feature points of the adjacent brick corner, and the feature points of the current brick corner, fit the horizontal and vertical edges corresponding to the brick corner, and output the set of feature points of the brick corner and corresponding brick edge of the reference brick, adjacent brick, and current brick, preliminRsts, as the result of coarse localization.

[0020] In the first step, Gaussian filtering is used to smooth and filter out noise points in the original image Img0, and histogram equalization and gamma transformation are used to enhance contrast to obtain image Img1.

[0021] In the second step, the spectral distribution of image Img1 is calculated by fast Fourier transform, and a bandpass filter for frequency and phase is constructed. The spectrum of image Img1 is filtered out by the bandpass filter to extract the main frequency feature of image Img1. Then, the main frequency feature is restored by inverse Fourier transform to obtain image Img2.

[0022] Specifically, the fine localization process for the coarse localization results includes the following steps:

[0023] Step 1: In the feature point set of the coarse localization result, preliminRsts, each brick corner has several feature point coordinates: preliminPtsi = (p1, p2, p3…pi). Iterate through each feature coordinate point pi in preliminPtsi, and extract an image Img_win with a width and height of (w, h) within a window range centered on pi from the original image Img0. This brick edge includes both horizontal and vertical edges.

[0024] Step 2: Enhance the contrast of the image Img_win: Use histogram equalization to stretch the grayscale range of the image Img_win to 0~255, and use gamma transform to improve the contrast, to obtain the enhanced image Img_win_eh;

[0025] Step 3: Perform integral image calculation on the image Img_win_eh to obtain the integral image Integral;

[0026] Step 4: Calculate the gradient at each point in the image Img_win_eh using the integral image Integral to obtain the gradient map Img_win_grad;

[0027] Step 5: Set the edge gradient threshold to thr, traverse all pixels of the gradient map Img_win_grad, extract the extreme points that satisfy the edge gradient threshold thr as fine-localized edge points, and obtain the set of fine-localized edge points edgePts.

[0028] Step 6: Use the iterative model parameter estimation algorithm RANSAC to fit the set of finely located edge points edgePts to obtain the horizontal line line_h_acc and the vertical line line line_v_acc of the finely located edge points.

[0029] Step 7: Based on the geometric relationship, find the intersection point of the horizontal line line_h_acc and the vertical line line_v_acc of the fine positioning edge as the fine positioning brick corner point BrickPti, so as to realize the identification of the outline of the reference brick, the adjacent brick and the current brick.

[0030] In step four, the width and height of the gradient calculation window are (pw, ph). Therefore, the gradient value of point pt(x, y) in the gradient map Img_win_grad is calculated as follows:

[0031] Img_win_grad(x,y) = Sum1 – Sum2 – Sum3 + Sum4;

[0032] Sum1 = Integral(x + pw, y + ph);

[0033] Sum2 = Integral(x, y-ph);

[0034] Sum3 = Integral(x, y + ph);

[0035] Sum4 = Integral(x-pw,y-ph).

[0036] In step five, the finely located edge points in the set edgePts must satisfy the following condition: Img_win_grad(x,y) > thr

[0037] ∪ Img_win_grad(x,y) > Img_win_grad(x,y - 1)

[0038] ∪ Img_win_grad(x,y) > Img_win_grad(x,y + 1).

[0039] Specifically, the step of calculating the target position of the current brick by referencing the outline of the reference brick and / or the adjacent bricks and the width of the grout lines, calculating the angular and distance deviations between the current brick and the target position, and using the angular and distance deviations to control the actuator to lay the current brick refers to: setting the width of the grout lines, calculating the target brick corner point based on the horizontal and vertical lines of the precise positioning edge of the reference brick and the corner point of the precise positioning brick, and / or the horizontal and vertical lines of the precise positioning edge of the adjacent brick and the corner point of the precise positioning brick, thus obtaining the target position of the current brick; calculating the angular and distance deviations between the target brick corner point and the corner point of the precise positioning brick, and using the angular and distance deviations to control the actuator to move the current brick to the target position to achieve laying.

[0040] Furthermore, based on step seven, the following points are obtained: reference brick precise positioning brick corner D, reference brick precise positioning brick corner E, side brick precise positioning brick corner F, side brick precise positioning brick corner G, side brick precise positioning brick corner H, current brick precise positioning brick corner A1, current brick precise positioning brick corner B1, current brick precise positioning brick corner C1, the horizontal straight line of the precise positioning edge of the reference brick, and the vertical straight line of the precise positioning edge of the side brick.

[0041] Includes the following steps:

[0042] Step 1: Set the width Gap of the grout line. Using the width Gap, the reference brick's precise positioning brick corner feature point D, the reference brick's precise positioning brick corner point E, the adjacent brick's precise positioning brick corner point F, the adjacent brick's precise positioning brick corner point G, the adjacent brick's precise positioning brick corner point H, the horizontal straight line of the reference brick's precise positioning edge, and the vertical straight line of the adjacent brick's precise positioning edge, calculate the target brick corner point A2, the target brick corner point B2, and the target brick corner point C2 of the current brick laying, thus obtaining the target position of the current brick laying.

[0043] Step 2: The current brick corner A1, B1, and C1 points need to be aligned with the target brick corners A2, B2, and C2 points respectively. Calculate the angle deviation value offset_r and the distance deviation values ​​offset_x and offset_y.

[0044] center1 = (A1 + B1 + C1) / 3;

[0045] center2 = (A2 + B2 + C2) / 3;

[0046] Decentralized point sets:

[0047] ;

[0048] ;

[0049] ;

[0050] ;

[0051] ;

[0052] ;

[0053] Calculate matrix H:

[0054] ;

[0055] Perform SVD decomposition on H: ;

[0056] make ;

[0057] ;

[0058] ;

[0059] ;

[0060] ;

[0061] Step 3: Use the angle deviation value offset_r and the distance deviation values ​​offset_x and offset_y to control the actuator to move the current brick to the target position for laying the current brick to achieve the laying.

[0062] Both the laser projection mechanism one and the laser projection mechanism two of this invention are vision systems composed of laser emitters and DOE diffraction gratings. Compared with traditional visual lighting methods, which are easily affected by the color and pattern of the brick surface and the background adhesive color when extracting edges, this invention uses structured light projection patterns of the laser projection mechanism one and the laser projection mechanism two to form significant graphic features on the brick surfaces of the current brick, the reference brick, and the adjacent bricks. The graphic features are then coarsely and finely positioned to accurately identify the edges of the current brick, the reference brick, and the adjacent bricks. Finally, a guiding execution mechanism completes the high-precision tiling. The tile laying method of this invention has strong anti-interference and robustness, and can adapt to various tile colors, complex patterns, different background adhesive colors and complex lighting conditions for tile laying. The structured light projection pattern is not sensitive to the color, pattern and adhesive color of the tile surface, ensuring accurate identification of the tile edge under different tile colors, patterns and adhesive colors. It can also guide the execution mechanism to complete high-precision tile laying in harsh construction environments, while ensuring high-precision joint width control and corner alignment, thereby meeting the tile joint accuracy requirements.

[0063] Meanwhile, the structured light module used in this invention is less expensive than traditional light sources, thus reducing the cost of this laying method. Furthermore, this method for laying floor tiles can be extended to decorative tiles, irregularly shaped tiles, wall tiles, and other construction scenarios requiring high-precision placement.

[0064] Compared with the prior art, the present invention has the following advantages and beneficial effects: the tile laying method based on structured light projection has strong anti-interference ability and can guide the execution mechanism to complete high-precision tile laying in harsh construction environments, while ensuring high-precision joint width control and corner alignment, thereby meeting the tile joint accuracy requirements. Attached Figure Description

[0065] Figure 1 This is a flowchart of the method for laying floor tiles according to the present invention;

[0066] Figure 2 This is a schematic diagram showing the layout of laser projection mechanism one, camera one, laser projection mechanism two, and camera two in the method for laying floor tiles according to the present invention.

[0067] Figure 3 This is a schematic diagram of the reference brick, the adjacent brick, and the current brick in the method of laying floor tiles according to the present invention;

[0068] Among them, 1 is laser projection mechanism one, 2 is camera one, 3 is laser projection mechanism two, 4 is camera two, 5 is the current brick, 6 is the reference brick, 7 is the side brick one, and 8 is the side brick two. Detailed Implementation

[0069] The present invention will now be described in further detail with reference to the accompanying drawings and specific embodiments.

[0070] Example 1

[0071] In this embodiment, floor tiles are laid between a reference tile and two adjacent tiles.

[0072] like Figures 1 to 3 As shown, the method for laying floor tiles based on structured light projection guidance of the present invention is as follows: setting up an execution mechanism for grasping the current tile 5, and a laser projection mechanism 1, a camera 2, a laser projection mechanism 3 and a camera 4 installed on the execution mechanism;

[0073] The method for laying floor tiles is as follows: The control actuator places the current tile 5 close to the reference tile 6, the first adjacent tile 7, and the second adjacent tile 8; the laser projection mechanism 1 and the laser projection mechanism 2 are activated, and the laser projection mechanism 1 and the laser projection mechanism 2 form projection patterns on the current tile 5, the reference tile 6, and the first adjacent tile 7 and the second adjacent tile 8, respectively, and the images are acquired by the camera 12 and the camera 24, respectively; in the acquired images, the brick edge features in the projection pattern are subjected to coarse positioning processing and fine positioning processing to identify the outlines of the current tile 5, the reference tile 6, and the first adjacent tile 7 and the second adjacent tile 8; the target position of the current tile is calculated based on the outlines of the reference tile 6, the first adjacent tile 7, and the second adjacent tile 8 and the width of the tile joints; the angle deviation value and the distance deviation value between the current tile and the target position of the current tile are calculated; the angle deviation value and the distance deviation value are used to control the actuator to lay the current tile.

[0074] The specific steps of this invention include:

[0075] (1) The control actuator places the current brick 5 close to the reference brick 6, the first adjacent brick 7, and the second adjacent brick 8; the laser projection mechanism 1 and the laser projection mechanism 2 are activated, and the laser projection mechanism 1 and the laser projection mechanism 2 form projection patterns on the current brick 5, the reference brick 6, the first adjacent brick 7, and the second adjacent brick 8, respectively, and the images are acquired by the camera 12 and the camera 24, respectively. Among them, the laser projection mechanism 1 and the laser projection mechanism 2 are both vision systems composed of laser emitters and DOE diffraction gratings. The image acquired by the camera 12 includes the current brick 5 and the reference brick 6, and the image acquired by the camera 24 includes the current brick 5, the reference brick 6, the first adjacent brick 7, and the second adjacent brick 8.

[0076] (2) In the images acquired by camera 12 and camera 24, the brick edge features in the projection pattern are coarsely located: the mask of the projection pattern and the image containing only the projection pattern itself are obtained. The brick edge features are identified in the image containing only the projection pattern itself in the mask. The brick corner feature points of reference brick 6, side brick 17, side brick 28 and current brick 5 are obtained and fitted to obtain the horizontal and vertical edges corresponding to the brick corners. The set of feature points of brick corners and corresponding brick edges of reference brick 6, side brick 17, side brick 28 and current brick 5 is output as the result of coarse localization.

[0077] Specifically, it includes:

[0078] The first step is to perform spatial preprocessing on the original images Img0 from camera 12 and camera 24: Gaussian filtering is used to smooth and filter out noise points in the original image Img0, and histogram equalization and gamma transformation are used to enhance the contrast to obtain image Img1.

[0079] The second step is to perform frequency domain preprocessing on image Img1: the spectral distribution of image Img1 is calculated by fast Fourier transform, and a bandpass filter for frequency and phase is constructed. The spectrum of image Img1 is filtered out by the bandpass filter to extract the main frequency features of image Img1. Then, the main frequency features are restored by inverse Fourier transform to obtain image Img2.

[0080] The third step is to extract the effective region of the projection pattern from the image Img1 using the brightness binarization threshold segmentation method, and then fit the minimum circumscribed convex polygon of the effective region to obtain the mask region of the projection pattern.

[0081] The fourth step is to binarize the image Img2, extract the highlight parts, and segment out the image Img3, which contains only the projection pattern itself.

[0082] The fifth step is to retrieve the edge feature points of the brick in the projection pattern: the abrupt change features in the projection pattern are used as the edge feature pattern of the brick; in the image Img3, all pixels in the masked area are traversed, and it is determined whether the feature of the pixel area matches the edge feature pattern. If they match, the pixel is added to the feature point set PointList.

[0083] The sixth step is to use the iterative model parameter estimation algorithm RANSAC to fit the brick corners of all feature points in PointList, and obtain the reference brick corner D feature point, the side column brick corner feature points, and the current brick corner feature point.

[0084] Step 7: Traverse the feature points of the reference brick corner, the feature points of the adjacent brick corners E, F, G, and H, the feature point of the current brick corner A1, and the feature point of the current brick corner B1, fit the horizontal and vertical sides corresponding to the brick corner, and output the set of feature points of the above brick corner and corresponding brick side, preliminRsts, as the result of coarse localization.

[0085] (3) Perform fine positioning processing on the coarse positioning results, including:

[0086] Step 1: In the feature point set of the coarse localization result, preliminRsts, each brick corner has several feature point coordinates: preliminPtsi = (p1, p2, p3…pi). Iterate through each feature coordinate point pi in preliminPtsi, and extract an image Img_win with a width and height of (w, h) within a window range centered on pi from the original image Img0. This brick edge includes both horizontal and vertical edges.

[0087] Step 2: Enhance the contrast of the image Img_win: Use histogram equalization to stretch the grayscale range of the image Img_win to 0~255, and use gamma transform to improve the contrast, to obtain the enhanced image Img_win_eh;

[0088] Step 3: Perform integral image calculation on the image Img_win_eh to obtain the integral image Integral;

[0089] Step 4: Calculate the gradient at each point in the image Img_win_eh using the integral image Integral to obtain the gradient map Img_win_grad;

[0090] If the width and height of the gradient calculation window are (pw, ph), then the gradient value of point pt(x, y) in the gradient map Img_win_grad is calculated as follows:

[0091] Img_win_grad(x,y) = Sum1 – Sum2 – Sum3 + Sum4;

[0092] Sum1 = Integral(x + pw, y + ph);

[0093] Sum2 = Integral(x, y-ph);

[0094] Sum3 = Integral(x, y + ph);

[0095] Sum4 = Integral(x-pw,y-ph);

[0096] Step 5: Set the edge gradient threshold to thr, traverse all pixels in the gradient map Img_win_grad, and extract the extreme points that satisfy the edge gradient threshold thr as fine-localized edge points, obtaining the fine-localized edge point set edgePts. The fine-localized edge points in edgePts must satisfy the following condition: Img_win_grad(x,y) > thr

[0097] ∪ Img_win_grad(x,y) > Img_win_grad(x,y - 1)

[0098] ∪ Img_win_grad(x,y) > Img_win_grad(x,y + 1);

[0099] Step 6: Use the iterative model parameter estimation algorithm RANSAC to fit the set of finely located edge points edgePts to obtain the horizontal line line_h_acc and the vertical line line line_v_acc of the finely located edge points.

[0100] Step 7: Based on geometric relationships, determine the intersection point of the horizontal line of the fine positioning edge (line_h_acc) and the vertical line of the fine positioning edge (line_v_acc) as the fine positioning brick corner point (BrickPti), thus recognizing the contours of the reference brick, adjacent bricks, and the current brick. Specifically, obtain the fine positioning brick corner points D and E of the reference brick, F, G, and H of the adjacent bricks, A1 and B1 of the current brick, the horizontal line of the fine positioning edge of the reference brick, and the vertical line of the fine positioning edge of the adjacent brick. Calculate the fine positioning brick corner point C1 of the current brick using points A1 and B1.

[0101] (4) Set the width Gap of the paving joint. Calculate the target brick corner point of the current brick based on the horizontal and vertical straight lines of the precise positioning edge of the reference brick and the corner point of the precise positioning brick, as well as the horizontal and vertical straight lines of the precise positioning edge of adjacent brick 1 and adjacent brick 2, thus obtaining the target position of the current brick. Calculate the angle deviation and distance deviation between the target brick corner point of the current brick and the corner point of the precise positioning brick of the current brick. Use the angle deviation and distance deviation to control the actuator to drive the current brick 5 to the target position of the current brick to achieve paving.

[0102] include:

[0103] Step 1: Set the width Gap of the grout line. Using the width Gap, the reference brick's precise positioning brick corner feature point D, the reference brick's precise positioning brick corner point E, the adjacent brick's precise positioning brick corner point F, the adjacent brick's precise positioning brick corner point G, the adjacent brick's precise positioning brick corner point H, the horizontal straight line of the reference brick's precise positioning edge, and the vertical straight line of the adjacent brick's precise positioning edge, calculate the target brick corner point A2, the target brick corner point B2, and the target brick corner point C2 of the current brick laying, thus obtaining the target position of the current brick laying.

[0104] Step 2: The current brick corner A1, B1, and C1 points need to be aligned with the target brick corners A2, B2, and C2 points respectively. Calculate the angle deviation value offset_r and the distance deviation values ​​offset_x and offset_y.

[0105] center1 = (A1 + B1 + C1) / 3;

[0106] center2 = (A2 + B2 + C2) / 3;

[0107] Decentralized point sets:

[0108] ;

[0109] ;

[0110] ;

[0111] ;

[0112] ;

[0113] ;

[0114] Calculate matrix H:

[0115] ;

[0116] Perform SVD decomposition on H: ;

[0117] make ;

[0118] ;

[0119] ;

[0120] ;

[0121] ;

[0122] Step 3: Use the angle deviation value offset_r and the distance deviation values ​​offset_x and offset_y to control the actuator to move the current brick 5 to the target position of the current brick to achieve the laying.

[0123] Example 2

[0124] The only difference between this embodiment and embodiment one is that this embodiment can lay floor tiles on a reference tile using an AI-guided positioning method.

[0125] This embodiment obtains the two precisely positioned corner points of the reference brick and the horizontal or vertical straight line of the precisely positioned edge of the reference brick through width gap and precise positioning processing. The target corner point of the current brick is then calculated, thus obtaining the target position for the current brick. The angle deviation value offset_r and distance deviation values ​​offset_x and offset_y are calculated using the method in step 2 of embodiment one. These angle deviation values ​​offset_r and distance deviation values ​​offset_x and offset_y are then used to control the actuator to move the current brick to the target position for laying.

[0126] This embodiment can also use an AI-guided method to lay floor tiles between a reference tile and a side tile.

[0127] The other methods in this embodiment are the same as in Embodiment 1.

[0128] The above embodiments are preferred embodiments of the present invention, but the embodiments of the present invention are not limited to the above embodiments. Any changes, modifications, substitutions, combinations, or simplifications made without departing from the spirit and principle of the present invention shall be considered equivalent substitutions and shall be included within the protection scope of the present invention.

Claims

1. A method for laying floor tiles based on structured light projection guidance, characterized in that: The system includes an actuator for grasping the current brick, and a laser projection mechanism 1, a camera 1, a laser projection mechanism 2, and a camera 2 mounted on the actuator. The method for laying floor tiles is as follows: The control actuator places the current tile close to the reference tile and / or adjacent tiles; Laser projection mechanism one and laser projection mechanism two are activated, and they respectively form projection patterns on the current tile, reference tile, and / or adjacent tiles, and images are acquired by camera one and camera two respectively; In the acquired images, coarse positioning processing and fine positioning processing are performed on the tile edge features in the projection pattern to identify the outlines of the current tile, reference tile, and / or adjacent tiles; The target position for laying the current tile is calculated based on the outlines of the reference tile and / or adjacent tiles and the width of the tile joints, and the angular deviation value and distance deviation value between the current tile and the target position are calculated. The angular deviation value and distance deviation value are used to control the actuator to lay the current tile.

2. The method for laying floor tiles based on structured light projection guidance according to claim 1, characterized in that: In the acquired images, coarse localization of brick edge features in the projection pattern refers to: obtaining the masked area of ​​the projection pattern and an image containing only the projection pattern itself; identifying brick edge features in the corresponding image containing only the projection pattern itself in the masked area; obtaining the brick corner feature points of the reference brick, adjacent bricks, and the current brick; fitting these features to obtain the horizontal and vertical edges corresponding to the brick corners; and outputting the set of feature points of the brick corners and corresponding brick edges of the reference brick, adjacent bricks, and the current brick as the result of coarse localization.

3. The method for laying floor tiles based on structured light projection guidance according to claim 2, characterized in that: Includes the following steps: The first step is to perform spatial preprocessing on the original image Img0 to obtain image Img1; The second step is to perform frequency domain preprocessing on image Img1 to obtain image Img2; The third step is to extract the effective region of the projection pattern from the image Img1 using the brightness binarization threshold segmentation method, and then fit the minimum circumscribed convex polygon of the effective region to obtain the mask region of the projection pattern. The fourth step is to binarize the image Img2, extract the highlight parts, and segment out the image Img3, which contains only the projection pattern itself. The fifth step is to retrieve the edge feature points of the brick in the projection pattern: the abrupt change features in the projection pattern are used as the edge feature pattern of the brick; in the image Img3, all pixels in the masked area are traversed, and it is determined whether the feature of the pixel area matches the edge feature pattern. If they match, the pixel is added to the feature point set PointList. The sixth step is to use the iterative model parameter estimation algorithm RANSAC to fit the brick corners of all feature points in PointList, and obtain the reference brick corner feature points, the side column brick corner feature points, and the current brick corner feature points. Step 7: Traverse the feature points of the reference brick corner, the feature points of the adjacent brick corner, and the feature points of the current brick corner, fit the horizontal and vertical edges corresponding to the brick corner, and output the set of feature points of the brick corner and corresponding brick edge of the reference brick, adjacent brick, and current brick, preliminRsts, as the result of coarse localization.

4. The method for laying floor tiles based on structured light projection guidance according to claim 2, characterized in that: In the first step, Gaussian filtering is used to smooth and filter out noise points in the original image Img0, and histogram equalization and gamma transformation are used to enhance contrast to obtain image Img1.

5. The method for laying floor tiles based on structured light projection guidance according to claim 2, characterized in that: In the second step, the spectral distribution of image Img1 is calculated by fast Fourier transform, and a bandpass filter for frequency and phase is constructed. The spectrum of image Img1 is filtered out by the bandpass filter to extract the main frequency feature of image Img1. Then, the main frequency feature is restored by inverse Fourier transform to obtain image Img2.

6. The method for laying floor tiles based on structured light projection guidance according to claim 3, characterized in that: The fine localization process for the coarse localization results includes the following steps: Step 1: In the feature point set of the coarse localization result, preliminRsts, each brick corner has several feature point coordinates: preliminPtsi = (p1, p2, p3…pi). Iterate through each feature coordinate point pi in preliminPtsi, and extract an image Img_win with a width and height of (w, h) within a window range centered on pi from the original image Img0. This brick edge includes both horizontal and vertical edges. Step 2: Enhance the contrast of the image Img_win: Use histogram equalization to stretch the grayscale range of the image Img_win to 0~255, and use gamma transform to improve the contrast, to obtain the enhanced image Img_win_eh; Step 3: Perform integral image calculation on the image Img_win_eh to obtain the integral image Integral; Step 4: Calculate the gradient at each point in the image Img_win_eh using the integral image Integral to obtain the gradient map Img_win_grad; Step 5: Set the edge gradient threshold to thr, traverse all pixels of the gradient map Img_win_grad, extract the extreme points that satisfy the edge gradient threshold thr as fine-localized edge points, and obtain the set of fine-localized edge points edgePts. Step 6: Use the iterative model parameter estimation algorithm RANSAC to fit the set of finely located edge points edgePts to obtain the horizontal line line_h_acc and the vertical line line line_v_acc of the finely located edge points. Step 7: Based on the geometric relationship, find the intersection point of the horizontal line line_h_acc and the vertical line line_v_acc of the fine positioning edge as the fine positioning brick corner point BrickPti, so as to realize the identification of the outline of the reference brick, the adjacent brick and the current brick.

7. The method for laying floor tiles based on structured light projection guidance according to claim 6, characterized in that: In step four, the width and height of the gradient calculation window are (pw, ph). Therefore, the gradient value of point pt(x, y) in the gradient map Img_win_grad is calculated as follows: Img_win_grad(x,y) = Sum1 – Sum2 – Sum3 + Sum4; Sum1 = Integral(x + pw, y + ph); Sum2 = Integral(x, y-ph); Sum3 = Integral(x, y + ph); Sum4 = Integral(x-pw,y-ph).

8. The method for laying floor tiles based on structured light projection guidance according to claim 6, characterized in that: In step five, the finely located edge points in the set edgePts must satisfy the following condition: Img_win_grad(x,y) >thr ∪ Img_win_grad(x,y) > Img_win_grad(x,y - 1) ∪ Img_win_grad(x,y) > Img_win_grad(x,y + 1).

9. The method for laying floor tiles based on structured light projection guidance according to claim 6, characterized in that: The process of calculating the target position of the current brick by referencing the outline of the reference brick and / or the adjacent bricks and the width of the grout lines, calculating the angular and distance deviations between the current brick and the target position, and using these angular and distance deviations to control the actuator to lay the current brick refers to: setting the width of the grout lines, calculating the target brick corner point based on the horizontal and vertical lines of the precise positioning edge of the reference brick and the corner point of the precise positioning brick, and / or the horizontal and vertical lines of the precise positioning edge of the adjacent brick and the corner point of the precise positioning brick, thus obtaining the target position of the current brick; calculating the angular and distance deviations between the target brick corner point and the precise positioning brick corner point, and using these angular and distance deviations to control the actuator to move the current brick to the target position to achieve laying.

10. The method for laying floor tiles based on structured light projection guidance according to claim 9, characterized in that: Based on step seven, we obtain the following points: reference brick precise positioning brick corner D, reference brick precise positioning brick corner E, side brick precise positioning brick corner F, side brick precise positioning brick corner G, side brick precise positioning brick corner H, current brick precise positioning brick corner A1, current brick precise positioning brick corner B1, current brick precise positioning brick corner C1, reference brick precise positioning edge horizontal line, and side brick precise positioning edge vertical line. Includes the following steps: Step 1: Set the width Gap of the grout line. Using the width Gap, the reference brick's precise positioning brick corner feature point D, the reference brick's precise positioning brick corner point E, the adjacent brick's precise positioning brick corner point F, the adjacent brick's precise positioning brick corner point G, the adjacent brick's precise positioning brick corner point H, the horizontal straight line of the reference brick's precise positioning edge, and the vertical straight line of the adjacent brick's precise positioning edge, calculate the target brick corner point A2, the target brick corner point B2, and the target brick corner point C2 of the current brick laying, thus obtaining the target position of the current brick laying. Step 2: The current brick corner A1, B1, and C1 points need to be aligned with the target brick corners A2, B2, and C2 points respectively. Calculate the angle deviation value offset_r and the distance deviation values ​​offset_x and offset_y. center1 = (A1 + B1 + C1) / 3; center2 = (A2 + B2 + C2) / 3; Decentralized point sets: ； ； ； ； ； ； Calculate matrix H: ； Perform SVD decomposition on H: ; make ; ； ； ； ； Step 3: Use the angle deviation value offset_r and the distance deviation values ​​offset_x and offset_y to control the actuator to move the current brick to the target position for laying the current brick to achieve the laying.

Images