Method for monitoring the consistency of slurry for autoclaved aerated concrete blocks

Abstract

The present application relates to the technical field of image processing, and more particularly to a slurry mixing consistency monitoring method for autoclaved aerated concrete blocks, which obtains continuous multiple frame overhead images of the slurry and obtains a grayscale image; performs edge detection on each grayscale image respectively, and screens at least one suspected slurry mixing texture edge in each grayscale image; divides all suspected slurry mixing texture edges into at least two sets of texture edge groups; splices all suspected slurry mixing texture edges in each set of texture edge groups, obtains a possibility index of a target texture edge corresponding to each set of texture edge groups being a real mixing texture, and obtains at least one mixing texture edge; according to the texture characteristics and recovery speed of all mixing texture edges, a slurry mixing consistency coefficient is obtained, which is used to determine the slurry mixing consistency grade, reduces the texture interference caused by reflection and machine vibration, and improves the slurry mixing consistency monitoring accuracy.

Description

Technical Field

[0001] This invention relates to the field of image processing technology, and in particular to a method for monitoring the consistency of grout mixing in autoclaved aerated concrete blocks. Background Technology

[0002] Autoclaved aerated concrete (AAC) blocks are lightweight, porous, and environmentally friendly building materials with excellent thermal insulation, sound insulation, and seismic resistance, making them widely used in modern construction. During their production, the consistency of the mortar is a key factor affecting the performance, quality, and production efficiency of the AAC blocks. If the mortar is too thin or too thick, it will affect the uniformity of the block's pore structure, thus impacting its weight, strength, thermal insulation, and water absorption. Therefore, monitoring the consistency of the mortar is a crucial step in the production of AAC blocks.

[0003] Existing methods for monitoring slurry consistency mainly utilize edge detection and other techniques to identify textured areas in images, analyzing the consistency based on the agitation texture features on the slurry surface during mixing. However, because equipment inevitably vibrates during mixing, and slurry has a certain fluidity, the captured images may contain not only agitation textures but also water ripples caused by vibration. These water ripples do not accurately reflect the relationship between agitation texture and slurry consistency, potentially affecting the judgment of slurry consistency and impacting monitoring accuracy.

[0004] Therefore, how to reduce the interference of water ripples caused by vibration on the monitoring of slurry mixing consistency and improve the monitoring accuracy has become an urgent problem to be solved. Summary of the Invention

[0005] In view of this, embodiments of the present invention provide a method for monitoring the consistency of grout mixing in autoclaved aerated concrete blocks, in order to solve the problem of how to reduce the interference of water ripples generated by vibration on the monitoring of grout mixing consistency and improve the monitoring accuracy.

[0006] This invention provides a method for monitoring the consistency of grout used in autoclaved aerated concrete (AAC) blocks, the method comprising the following steps:

[0007] Acquire multiple consecutive overhead images of the slurry, and convert each overhead image to grayscale to obtain the corresponding grayscale image;

[0008] Edge detection is performed on each grayscale image to obtain the edges in each grayscale image. Based on the grayscale difference between pixels on each edge in each grayscale image, at least one suspected slurry mixing texture edge in each grayscale image is selected.

[0009] Perform the same texture edge evaluation on every two suspected slurry mixing texture edges between two adjacent grayscale images, and divide all suspected slurry mixing texture edges into at least two sets of texture edges;

[0010] For any set of texture edges, all suspected slurry stirring texture edges in the set of texture edges are stitched together to obtain the target texture edge. A template image of the slurry stirring trajectory is obtained. The target texture edge is matched with the template texture edge in the template image to obtain the probability index that the target texture edge is a real stirring texture.

[0011] The probability index of the target texture edge corresponding to each set of texture edges being the real stirring texture is obtained, and at least one stirring texture edge is obtained. Based on the texture characteristics and recovery speed of all stirring texture edges, the slurry stirring consistency coefficient is obtained to determine the slurry stirring consistency level.

[0012] Preferably, the step of filtering out at least one suspected slurry mixing texture edge in each grayscale image based on the grayscale difference between pixels on each edge in each grayscale image includes:

[0013] For any edge in any grayscale image, any edge pixel on the edge is taken as the analysis pixel. Two non-edge pixels with symmetrical positions are obtained in the four neighborhoods of the analysis pixel and recorded as the pixels on both sides of the analysis pixel. The absolute value of the grayscale value difference between the pixels on both sides of the analysis pixel is calculated. The reciprocal of the sum of the absolute value of the grayscale value difference and a preset value is taken as the grayscale difference degree on both sides of the edge of the analysis pixel.

[0014] The grayscale difference between the two sides of each edge pixel on any edge is obtained. The cumulative value of the grayscale difference between the two sides of all edge pixels on any edge is normalized to obtain the probability index of any edge belonging to the slurry stirring texture edge. If the probability index of any edge belonging to the slurry stirring texture edge is greater than the preset probability index threshold, then any edge is regarded as a suspected slurry stirring texture edge.

[0015] Traverse each edge in each grayscale image to obtain at least one suspected slurry mixing texture edge in each grayscale image.

[0016] Preferably, the step of performing the same texture edge evaluation on every two suspected slurry stirring texture edges between two adjacent grayscale images, and dividing all suspected slurry stirring texture edges into at least two sets of texture edges, includes:

[0017] For the xth suspected slurry stirring texture edge in the i-th grayscale image, in the (i+1)-th grayscale image, obtain suspected slurry stirring texture edges that overlap with the xth suspected slurry stirring texture edge or whose edge endpoints are within a preset number of pixels, and record them as edges to be analyzed. If at least one edge to be analyzed is obtained, obtain the same texture probability index between the xth suspected slurry stirring texture edge and each edge to be analyzed.

[0018] Obtain the probability index of the same texture between each suspected slurry stirring texture edge in the i-th grayscale image and the corresponding edge to be analyzed in the (i+1)-th grayscale image, and obtain at least one edge pair consisting of two suspected slurry stirring texture edges belonging to the same texture between the i-th grayscale image and the (i+1)-th grayscale image.

[0019] Obtain all edge pairs between any two adjacent grayscale images. Based on the transitivity of parallel lines, edge pairs belonging to the same texture are grouped into a texture edge set, resulting in at least one texture edge set. Each texture edge set corresponds to one texture.

[0020] Preferably, obtaining the probability index of the same texture between the xth suspected slurry stirring texture edge and each edge to be analyzed includes:

[0021] For any edge to be analyzed, calculate the distance between the edge endpoint of the edge to be analyzed and the edge endpoint of the xth suspected slurry stirring texture edge, obtain the two edge endpoints corresponding to the minimum distance, and record them as edge end points. Take the edge end point as the last edge pixel point, obtain a first preset number of edge pixels on the edge to be analyzed to form a first pixel point sequence, and obtain a first preset number of edge pixels on the xth suspected slurry stirring texture edge to form a second pixel point sequence;

[0022] Based on the slope of the first pixel sequence and the slope of the second pixel sequence, a first probability index is obtained that any edge to be analyzed and the xth suspected slurry mixing texture edge belong to the same texture in the edge extension direction. With the position of the second pixel sequence as a reference, based on the difference between the fitted position and the actual position of the first pixel sequence, a second probability index is obtained that any edge to be analyzed and the xth suspected slurry mixing texture edge belong to the same texture in relative position.

[0023] The product of the first probability index and the second probability index is normalized to obtain the same texture probability index between the xth suspected slurry stirring texture edge and any edge to be analyzed.

[0024] Preferably, the step of obtaining a first probability index that any edge to be analyzed and the xth suspected slurry mixing texture edge belong to the same texture in the edge extension direction based on the slope of the first pixel sequence and the slope of the second pixel sequence includes:

[0025] Calculate the slope of the straight line between two adjacent pixels in the first pixel sequence to obtain the average slope, denoted as the first slope. Calculate the slope of the straight line between two adjacent pixels in the second pixel sequence to obtain the average slope, denoted as the second slope. Calculate the absolute value of the difference between the first slope and the second slope. The reciprocal of the sum of the absolute value of the difference and a preset value is used as the first probability index that any edge to be analyzed and the xth suspected slurry mixing texture edge belong to the same texture in the edge extension direction.

[0026] Preferably, the step of using the position of the second pixel sequence as a reference and obtaining a second probability index that any edge to be analyzed and the xth suspected slurry mixing texture edge belong to the same texture in relative position based on the difference between the fitted position and the actual position of the first pixel sequence includes:

[0027] Based on the abscissa value of each pixel in the second pixel sequence, the ordinate values ​​of all pixels are fitted using the least squares method to obtain a fitting curve. Using the fitting curve, the fitted value of the ordinate of each pixel in the first pixel sequence is obtained. The fitted value of the ordinate of each pixel in the first pixel sequence is subtracted from the actual value to obtain the absolute value of the difference for each pixel. The average value of the absolute values ​​of the differences for all pixels in the first pixel sequence is obtained. The reciprocal of the sum of the average value and the preset value is used as the second probability index that any edge to be analyzed and the xth suspected slurry mixing texture edge belong to the same texture in relative position.

[0028] Preferably, the method for obtaining at least one pair of edge pairs consisting of two suspected slurry mixing texture edges belonging to the same texture between the i-th grayscale image and the (i+1)-th grayscale image includes:

[0029] If the probability index of the same texture between any suspected slurry stirring texture edge in the i-th grayscale image and any corresponding edge to be analyzed in the (i+1)-th grayscale image is greater than the preset threshold of the same texture probability index, then any edge to be analyzed is taken as the target edge. If it is detected that any suspected slurry stirring texture edge in the i-th grayscale image corresponds to at least two target edges, then the target edge corresponding to the largest probability index of the same texture is selected and any suspected slurry stirring texture edge in the i-th grayscale image forms a pair of edges.

[0030] Preferably, the step of matching the target texture edge with the template texture edge in the template image to obtain a probability index that the target texture edge is a real stirring texture includes:

[0031] For any template texture edge in the template image, the minimum length is obtained as the size of the sliding window based on the length between the target texture edge and the template texture edge, and the edge corresponding to the minimum length is taken as the target window. Based on the size of the sliding window and the preset sliding step size, the sliding window is slid on the edge corresponding to the maximum length to obtain at least two sliding windows. The normalized correlation coefficient between the target window and each sliding window is calculated, and the maximum normalized correlation coefficient is taken as the matching degree between the target texture edge and the template texture edge.

[0032] The matching degree between the target texture edge and each template texture edge in the template image is obtained, and the maximum matching degree is used as the probability index that the target texture edge is a real stirring texture.

[0033] Preferably, obtaining the probability index that the target texture edge corresponding to each set of texture edges is a real stirring texture, and obtaining at least one stirring texture edge, includes:

[0034] A preset probability index threshold is obtained. If the probability index of the target texture edge corresponding to any set of texture edges is a real stirring texture is greater than or equal to the probability index threshold, then the target texture edge corresponding to any set of texture edges is determined to be a stirring texture edge.

[0035] Preferably, obtaining the slurry mixing consistency coefficient based on the texture features and recovery speed of all stirring texture edges includes:

[0036] For any stirring texture edge, each suspected slurry stirring texture edge in the texture edge set corresponding to any stirring texture edge is recorded as a marked edge. According to the grayscale image to which each marked edge belongs, the edge pixels of each marked edge are marked in the corresponding grayscale image to obtain the marked pixels. According to the sampling time of the grayscale image to which each marked edge belongs, the duration between the maximum sampling time and the minimum sampling time is obtained and recorded as the stirring texture duration.

[0037] Obtain the duration of the stirring texture corresponding to each stirring texture edge, and obtain the average duration of the stirring texture. Normalize the average duration of the stirring texture to obtain a normalized value. Based on each suspected slurry stirring texture edge in the texture edge set corresponding to each stirring texture edge, obtain all the marked pixels in each grayscale image, obtain the proportion of marked pixels in each grayscale image, and obtain the average proportion, which is recorded as the texture feature value.

[0038] The product of the normalized value and the texture feature value is used as the slurry mixing consistency coefficient.

[0039] The beneficial effects of the embodiments of the present invention compared with the prior art are as follows:

[0040] This invention acquires multiple consecutive overhead images of the slurry, converts each frame to grayscale to obtain a corresponding grayscale image, performs edge detection on each grayscale image to obtain the edges in each grayscale image, and filters out at least one suspected slurry mixing texture edge in each grayscale image based on the grayscale difference between pixels on each edge in each grayscale image. For every two suspected slurry mixing texture edges between two adjacent grayscale images, the same texture edge evaluation is performed, dividing all suspected slurry mixing texture edges into at least two sets of texture edges. For any set of texture edges, all suspected slurry mixing texture edges in the set are stitched together to obtain a target texture edge. A template image of the slurry mixing trajectory is obtained, and the target texture edge is matched with the template texture edge in the template image to obtain a probability index that the target texture edge is a real mixing texture. The probability index that the target texture edge corresponding to each set of texture edges is a real mixing texture is obtained, resulting in at least one mixing texture edge. Based on the texture features and recovery speed of all mixing texture edges, a slurry mixing consistency coefficient is obtained to determine the slurry mixing consistency level. Based on the regularity of the changes in the mixing texture during the slurry mixing process, continuous monitoring images are used to identify the real mixing texture, thereby reducing the impact of slurry surface reflection and machine vibration on the slurry mixing consistency analysis results, improving the accuracy of slurry mixing consistency monitoring, and thus ensuring the performance and quality of concrete blocks. Attached Figure Description

[0041] To more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0042] Figure 1This is a flowchart of a method for monitoring the consistency of grout mixing in autoclaved aerated concrete blocks, provided in Embodiment 1 of the present invention. Detailed Implementation

[0043] Embodiments of this disclosure are described in detail below, with examples of these embodiments illustrated in the accompanying drawings. The embodiments described below with reference to the accompanying drawings are exemplary and intended to explain this disclosure, and should not be construed as limiting it.

[0044] It should be noted that the terms "first," "second," etc., used in this disclosure and the accompanying drawings are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of this disclosure described herein can be implemented in orders other than those illustrated or described herein. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with this disclosure. Rather, they are merely examples of apparatuses and methods consistent with some aspects of this disclosure.

[0045] To illustrate the technical solution of the present invention, specific embodiments are described below.

[0046] See Figure 1 This is a flowchart of a method for monitoring the consistency of grout mixing in autoclaved aerated concrete blocks, provided in Embodiment 1 of the present invention. Figure 1 As shown, the method may include:

[0047] Step S101: Obtain multiple consecutive overhead images of the slurry, and convert each overhead image to grayscale to obtain a corresponding grayscale image.

[0048] In the process of monitoring the consistency of concrete block grout, the consistency is usually obtained based on the texture features of the grout surface during mixing. However, when the grout is thin, many water ripples appear during mixing, accompanied by changes in reflection, making the edge features of the grout surface complex and affecting the identification of the mixing texture features, thus affecting the accuracy of the grout consistency monitoring results. Therefore, in this embodiment of the invention, a camera is installed on top of the mixer to capture images of the grout from above. To analyze the consistency of the grout based on the changes in the grout surface texture during mixing, the camera is set to capture an image of the grout surface every 1 second and analyze the grout consistency every 5 minutes to prevent over-mixing. The image acquisition frequency and the consistency analysis frequency are not limited here and can be set according to the accuracy requirements. Considering that the occlusion of the mixing blades in the captured images is unavoidable, in order to prevent the interference of the mixing blades in the images on the grout feature analysis, an existing mixing blade image training model is used to identify the mixing blade area in each frame of the captured image and remove it, retaining only the grout surface image, and recording it as a series of consecutive top-down images of the grout. The impeller image training model is an existing technology and will not be elaborated here.

[0049] Each frame of the overhead image is converted to grayscale to obtain a corresponding grayscale image. This grayscale image is used to subsequently distinguish between normal textures and water ripple patterns during slurry mixing based on their directional and regularity differences, thereby improving the accuracy of slurry consistency monitoring results. Grayscale processing is an existing technology and will not be elaborated upon here.

[0050] Step S102: Perform edge detection on each grayscale image to obtain the edges in each grayscale image. Based on the grayscale difference between pixels on each edge in each grayscale image, filter out at least one suspected slurry mixing texture edge in each grayscale image.

[0051] The mortar of concrete blocks has a certain degree of reflectivity. The lower the consistency (the thinner the mortar), the more obvious the reflective surface. The reflection of the mortar surface will increase the edge complexity in the captured image to a certain extent, thus affecting the recognition of the mixing texture. Therefore, it is necessary to eliminate the interference of the edge of the reflective area and identify the pixels suspected of mortar mixing texture.

[0052] First, the edge pixels in each image are obtained using the Canny edge detection algorithm, and continuous edges are obtained using the closing operation to obtain the initial edges in each grayscale image. Because bubbles and particles may exist during the stirring process, interfering with the recognition of the stirring texture, and these bubbles and particles are relatively small, the initial edges with a total number of pixels less than or equal to 5 are identified as the edges of bubbles or particles. Then, the initial edges corresponding to a total number of pixels less than or equal to 5 are removed to obtain the edges in each grayscale image.

[0053] Then, considering that the texture generated during the stirring process is caused by the indentation of the slurry surface when the stirring paddle passes over the slurry, the gray values ​​on both sides of the edge of the texture are relatively close. However, one side of the reflective edge is a reflective area, and the other side is a normal area, resulting in a larger gray value in the reflective area and a relatively large gray value difference between the two sides. Therefore, in this embodiment of the invention, based on the gray value difference between pixels on each edge in each grayscale image, at least one suspected slurry stirring texture edge in each grayscale image is selected. The specific method for obtaining this edge is as follows:

[0054] For any edge in any grayscale image, any edge pixel on the edge is taken as the analysis pixel. Two non-edge pixels with symmetrical positions are obtained in the four neighborhoods of the analysis pixel and recorded as the pixels on both sides of the analysis pixel. The absolute value of the grayscale value difference between the pixels on both sides of the analysis pixel is calculated. The reciprocal of the sum of the absolute value of the grayscale value difference and a preset value is taken as the grayscale difference degree on both sides of the edge of the analysis pixel.

[0055] The grayscale difference between the two sides of each edge pixel on any edge is obtained. The cumulative value of the grayscale difference between the two sides of all edge pixels on any edge is normalized to obtain the probability index of any edge belonging to the slurry stirring texture edge. If the probability index of any edge belonging to the slurry stirring texture edge is greater than the preset probability index threshold, then any edge is regarded as a suspected slurry stirring texture edge.

[0056] Traverse each edge in each grayscale image to obtain at least one suspected slurry mixing texture edge in each grayscale image.

[0057] Taking the i-th edge as an example, the formula for calculating the probability index that the i-th edge belongs to the slurry mixing texture edge is as follows:

[0058]

[0059] in, This represents the probability index that the i-th edge belongs to the edge of the slurry mixing texture. Represents the normalization function. This represents the number of edge pixels on the i-th edge. This represents the grayscale difference between the two sides of the x-th edge pixel on the i-th edge. This represents a preset value used to ensure that the fraction is meaningful. 1 represents a constant.

[0060] It should be noted that, The smaller the grayscale difference between the two sides of the edge, the more likely it is to be the edge of a stirring texture. The smaller the value, the greater the probability that the i-th edge belongs to the slurry mixing texture edge.

[0061] Because the surface color of normal slurry is highly consistent, while the grayscale value of reflective areas is significantly higher than that of normal areas, the grayscale difference between the two sides of the edge pixels corresponding to the edge of the reflective area is significant. The difference in grayscale between the two sides of the texture edge pixel will be significantly higher than that in the normal area. After normalization using the above formula, the grayscale difference corresponding to the reflective edge will be significantly higher. It will concentrate in the range of [0, 0.5], while the texture edges in the normal area correspond to... It will concentrate in the range of [0.5, 1], therefore the preset probability index threshold is set to 0.5. When the i-th edge is considered to be a suspected slurry mixing texture edge, it is determined that the i-th edge is a suspected slurry mixing texture edge.

[0062] At this point, at least one suspected slurry mixing texture edge can be obtained in each grayscale image.

[0063] Step S103: Perform the same texture edge evaluation on every two suspected slurry stirring texture edges between two adjacent grayscale images, and divide all suspected slurry stirring texture edges into at least two sets of texture edges.

[0064] When using texture features in grayscale images to identify slurry consistency, generally, the deeper the texture formed by stirring, the greater the slurry consistency. However, during stirring, factors such as vibration may cause significant fluctuations on the slurry surface, forming numerous ripples. These ripples do not conform to the above rule; rather, they are ripple-like waves formed after disturbance due to the high fluidity of the slurry. These ripples change faster than the texture formed by stirring, which may make the slurry appear thinner. To avoid the influence of these ripples on the analysis results of slurry consistency, it is necessary to further identify the true stirring texture within the edges of suspected stirring textures.

[0065] Since the formation principles of ripple-like water ripples and stirring textures are different, the corresponding edges in the image have certain differences in direction, shape and dynamic change process. Therefore, in this embodiment of the invention, the true stirring texture edge is identified based on the change characteristics of the suspected slurry stirring texture edge in the continuously monitored grayscale image. According to step S102, the suspected slurry stirring texture edges in each grayscale image are obtained. The slurry stirring texture changes in a spiral shape following the movement of the stirring paddle, while the ripples generated by the vibration form circular or elliptical ripples at the vibration position and continue to affect and spread to the surroundings. Therefore, for the xth suspected slurry stirring texture edge in the i-th grayscale image, considering that the stirring texture is continuously formed following the stirring paddle and requires a certain recovery time after formation, the same stirring texture is continuous in position in adjacent frame images. Since the recovery time is shorter when the slurry is thinner, based on the image acquisition order, in the (i+1)-th grayscale image, the suspected slurry stirring texture edges that overlap with the xth suspected slurry stirring texture edge or whose edge endpoints are within a preset number of pixels are obtained and recorded as edges to be analyzed. Preferably, in order to avoid missed detections, the preset number should be set to a small number to ensure that as long as there are overlapping pixels, they are considered to belong to the same texture. Therefore, the preset number is set to be less than or equal to 3. This is not limited here, and the implementer can adjust it according to the requirements.

[0066] If no edge to be analyzed is found, the xth suspected slurry mixing texture edge is considered to be a ripple generated by vibration, and the xth suspected slurry mixing texture edge is discarded. Conversely, if at least one edge to be analyzed is found, the probability index of the same texture between the xth suspected slurry mixing texture edge and each edge to be analyzed is obtained, which is used for subsequent classification of suspected slurry mixing texture edges belonging to the same texture. Specifically, the method for obtaining the probability index of the same texture between the xth suspected slurry mixing texture edge and each edge to be analyzed is as follows:

[0067] The shape of the stirring texture is related to the stirring direction of the stirring paddle. The extension direction of the same stirring texture in adjacent frame images has a certain similarity. Therefore, for any edge to be analyzed, the distance between the edge endpoint of any edge to be analyzed and the edge endpoint of the xth suspected slurry stirring texture edge is calculated. The two edge endpoints corresponding to the minimum distance are obtained and recorded as edge end points. The edge end points are taken as the last edge pixel points. A first preset number of edge pixels are obtained on any edge to be analyzed to form a first pixel point sequence, and a first preset number of edge pixels are obtained on the xth suspected slurry stirring texture edge to form a second pixel point sequence. Preferably, the first preset number is set to 5, which is used to take the last 5 edge pixels of the edge to be analyzed and the xth suspected slurry stirring texture edge as the analysis objects of the edge extension direction.

[0068] First, based on the slope of the first pixel sequence and the slope of the second pixel sequence, a first probability index is obtained that any edge to be analyzed and the xth suspected slurry mixing texture edge belong to the same texture in the edge extension direction. Specifically, this includes: calculating the slope of the straight line between two adjacent pixels in the first pixel sequence to obtain the average straight line slope, denoted as the first slope; calculating the slope of the straight line between two adjacent pixels in the second pixel sequence to obtain the average straight line slope, denoted as the second slope; calculating the absolute value of the difference between the first slope and the second slope; and taking the reciprocal of the sum of the absolute value of the difference and a preset value as the first probability index that any edge to be analyzed and the xth suspected slurry mixing texture edge belong to the same texture in the edge extension direction.

[0069] The formula for calculating the first probability index that any edge to be analyzed and the xth suspected slurry mixing texture edge belong to the same texture in the edge extension direction is:

[0070]

[0071] in, Indicates the first probability indicator. This represents the second slope corresponding to the second pixel sequence. This represents the first slope corresponding to the first pixel sequence, where | represents the absolute value sign. This represents a preset value used to ensure that the fraction is meaningful. .

[0072] It should be noted that, This indicates the direction of the edge extension of the xth suspected slurry mixing texture edge. This indicates the direction of extension of the edge end of any edge to be analyzed, corresponding to... The smaller the value, the more consistent the edge extension direction of the two. The more the edge to be analyzed and the xth suspected slurry mixing texture edge belong to the same texture in the edge extension direction, the greater the corresponding first probability index.

[0073] Then, considering that multiple parallel stirring textures may appear during stirring, and that there may be more and denser textures when the slurry is thicker, in order to avoid judging parallel textures in adjacent frame images as the same stirring texture, the position of the second pixel point sequence is used as a reference. Based on the difference between the fitted position and the actual position of the first pixel point sequence, a second probability index is obtained that any edge to be analyzed and the xth suspected slurry stirring texture edge belong to the same texture in relative position. Specifically, this includes: performing least squares fitting on the ordinate values ​​of all pixels according to the abscissa value of each pixel in the second pixel point sequence to obtain a fitting curve; using the fitting curve, obtaining the fitted value of the ordinate of each pixel in the first pixel point sequence; subtracting the fitted value and the actual value of the ordinate of each pixel in the first pixel point sequence to obtain the absolute value of the difference for each pixel; obtaining the average value of the absolute values ​​of the differences for all pixels in the first pixel point sequence; and using the reciprocal of the sum of the average value and the preset value as the second probability index that any edge to be analyzed and the xth suspected slurry stirring texture edge belong to the same texture in relative position.

[0074] The formula for calculating the second probability index, which states that any edge to be analyzed and the xth suspected slurry mixing texture edge belong to the same texture in relative position, is as follows:

[0075]

[0076] in, This indicates the second probability indicator. This represents the absolute value of the difference between the fitted value and the actual value of the ordinate of the x-th pixel in the first pixel sequence. This represents a preset value used to ensure that the fraction is meaningful. 1 represents a constant.

[0077] It should be noted that, This represents the cumulative distance between the fitted value and the actual value of the ordinate of each pixel in the first pixel sequence. The smaller the value, the greater the probability that any edge to be analyzed belongs to the same texture as the xth suspected slurry mixing texture edge, and the greater the corresponding second probability index.

[0078] Finally, the product of the first probability index and the second probability index is normalized to obtain the same texture probability index between the x-th suspected slurry mixing texture edge and any of the edges to be analyzed. The formula for calculating the same texture probability index is:

[0079]

[0080] in, This represents the probability index of the x-th suspected slurry mixing texture edge being the same texture as any of the edges to be analyzed. Indicates the first probability indicator. This indicates the second probability indicator. This represents the normalization function.

[0081] Similarly, the probability index of the same texture between each suspected slurry stirring texture edge in the i-th grayscale image and the corresponding edge to be analyzed in the (i+1)-th grayscale image is obtained. This yields at least one pair of edge pairs consisting of two suspected slurry stirring texture edges belonging to the same texture between the i-th and (i+1)-th grayscale images. Because the same texture needs to have high regularity in both direction and position across multiple frames, and multiple textures may be generated simultaneously during stirring, these textures may only satisfy the condition of being close in position or having parallel extension directions. To avoid misidentifying these textures as the same texture, the probability index of the same texture between the x-th suspected slurry stirring texture edge and any of the edges to be analyzed is required. and All values ​​fall within the higher range of the calculated results, and after normalization using the above formula, only those values ​​are considered to be within the higher range. When the xth suspected slurry stirring texture edge and any edge to be analyzed may belong to the same texture, if the probability index of the same texture between any suspected slurry stirring texture edge in the i-th grayscale image and any corresponding edge to be analyzed in the (i+1)-th grayscale image is greater than the preset probability index threshold of the same texture, then any edge to be analyzed is taken as the target edge. If it is detected that any suspected slurry stirring texture edge in the i-th grayscale image corresponds to at least two target edges, then the target edge corresponding to the largest probability index of the same texture is selected to form an edge pair with any suspected slurry stirring texture edge in the i-th grayscale image.

[0082] Similarly, the probability index of the same texture between each suspected slurry stirring texture edge in the (i+1)th grayscale image and the corresponding edge to be analyzed in the (i+2)th grayscale image is obtained. This yields at least one pair of edge pairs consisting of two suspected slurry stirring texture edges belonging to the same texture between the (i+1)th and (i+2)th grayscale images. This process is repeated to obtain all edge pairs between every two adjacent grayscale images. Then, based on the transitivity of parallel lines, edge pairs belonging to the same texture are grouped into a texture edge set, resulting in at least one texture edge set. Each texture edge set corresponds to one texture. For example, suppose suspected slurry stirring texture edge 1 in the i-th grayscale image and suspected slurry stirring texture edge 2 in the (i+1)th grayscale image are one edge pair, and suspected slurry stirring texture edge 2 in the (i+1)th grayscale image and suspected slurry stirring texture edge 3 in the (i+2)th grayscale image are another edge pair. Then, suspected slurry stirring texture edges 1, 2, and 3 are constructed as a texture edge set, belonging to the same texture.

[0083] At this point, all suspected slurry mixing texture edges in all grayscale images are divided into multiple texture edge sets.

[0084] Step S104: For any set of texture edges, stitch together all suspected slurry stirring texture edges in the set of texture edges to obtain the target texture edge, obtain a template image of the slurry stirring trajectory, match the target texture edge with the template texture edge in the template image, and obtain the probability index that the target texture edge is the real stirring texture.

[0085] The stirring texture should exhibit a spiral shape following the movement trajectory of the agitator. Therefore, the direction of change of the same texture should be consistent with the movement trajectory of the agitator. Thus, in this embodiment of the invention, the edges of suspected slurry stirring textures belonging to the same texture are stitched together to obtain the change path of the texture within the specified time period (the acquisition time period corresponding to all grayscale images). This path is then compared with the movement path of the agitator to analyze the likelihood that the texture belongs to a genuine stirring texture, thereby screening and identifying genuine stirring textures and eliminating interference from vibration textures.

[0086] Therefore, firstly, the preset stirring trajectory of the agitator is drawn as a template to obtain a template image. An edge detection algorithm is used to obtain all edges in the template image, which are denoted as template texture edges. Then, for any set of texture edges, all suspected slurry stirring texture edges in that set are stitched together to obtain the target texture edge. The more similar the target texture edge is to the template texture edge, the more likely it is to be a real stirring texture. Therefore, the target texture edge is matched with the template texture edges in the template image to obtain an index indicating the probability that the target texture edge is a real stirring texture. The specific method for obtaining this index is as follows:

[0087] For any template texture edge in the template image, the minimum length is obtained as the size of the sliding window based on the length between the target texture edge and the template texture edge. The edge corresponding to the minimum length is taken as the target window. Based on the size of the sliding window and a preset sliding step size, the sliding window is moved along the edge corresponding to the maximum length to obtain at least two sliding windows. To ensure calculation accuracy, the sliding step size is set to one pixel. The normalized correlation coefficient between the target window and each sliding window is calculated. The maximum normalized correlation coefficient is used as the degree of matching between the target texture edge and any template texture edge. The normalized correlation coefficient is an existing technology and will not be elaborated here.

[0088] The matching degree between the target texture edge and each template texture edge in the template image is obtained, and the maximum matching degree is used as the probability index that the target texture edge is a real stirring texture.

[0089] The formula for calculating the probability index that the target texture edge is a real stirring texture is as follows:

[0090]

[0091] in, This indicates the likelihood that the edge of the target texture is a real stirring texture. Represents the maximum value function. This indicates the degree of matching between the edge of the target texture and the edge of any template texture.

[0092] Similarly, we obtain an index of the probability that the target texture edge corresponding to each set of texture edges is a real stirring texture.

[0093] Step S105: Obtain the probability index of the target texture edge corresponding to each set of texture edges being the real stirring texture, and obtain at least one stirring texture edge. Based on the texture characteristics and recovery speed of all stirring texture edges, obtain the slurry stirring consistency coefficient, which is used to determine the slurry stirring consistency level.

[0094] because The value of is in the range of 0 to 1, corresponding to The value of is also in the range of 0 to 1. When the normalized correlation coefficient is greater than or equal to 0.7, the two contrasting images are considered to be significantly correlated. Therefore, in this embodiment of the invention, a preset probability index threshold of 0.7 is set. If the probability index of the target texture edge corresponding to any set of texture edges is a real stirring texture is greater than or equal to 0.7, then the target texture edge corresponding to the set of texture edges is determined to be a stirring texture edge. Conversely, if the probability index of the target texture edge corresponding to any set of texture edges is less than 0.7, then the target texture edge corresponding to the set of texture edges does not conform to the stirring texture rule.

[0095] At this point, multiple stirring texture edges can be obtained, which represent the actual stirring texture. Since the thicker the slurry, the more stirring textures appear on its surface, and the slower their recovery speed, the slurry stirring consistency coefficient is obtained based on the texture characteristics and recovery speed of all stirring texture edges. The specific method for obtaining this coefficient is as follows:

[0096] For any stirring texture edge, each suspected slurry stirring texture edge in the texture edge set corresponding to the stirring texture edge is recorded as a marked edge. According to the grayscale image to which each marked edge belongs, the edge pixels of each marked edge are marked in the corresponding grayscale image to obtain the marked pixels. According to the sampling time of the grayscale image to which each marked edge belongs, the duration between the maximum sampling time and the minimum sampling time is obtained and recorded as the stirring texture duration T.

[0097] Obtain the duration of the stirring texture corresponding to the edge of each stirring texture, and get the average duration of the stirring texture. The average duration of the stirring texture is normalized to obtain a normalized value. Based on each suspected slurry stirring texture edge in the texture edge set corresponding to each stirring texture edge, all marked pixels in each grayscale image are obtained to obtain the percentage of marked pixels in each grayscale image. The average of the proportions is obtained and denoted as the texture feature value. The product of the normalized value and the texture feature value is used as the slurry mixing consistency coefficient.

[0098] The formula for calculating the consistency coefficient of the slurry is as follows:

[0099]

[0100] in, This indicates the consistency coefficient of the slurry after mixing. Represents texture feature values, This represents the normalized value of the average duration of the stirring texture.

[0101] It should be noted that the more marked pixels there are, the larger the texture feature value, the more the texture is stirred, and the thicker the corresponding slurry. The thicker the slurry, the slower the texture recovery. Therefore... The higher the value, the slower the overall texture recovery speed and the greater the consistency of the slurry.

[0102] Furthermore, based on the consistency coefficient of the slurry... To determine the consistency level of the slurry, in one embodiment, a random forest algorithm is used to train a slurry consistency grading model, based on the slurry consistency coefficient and corresponding consistency level of existing slurry images. This model can classify the slurry consistency level into three grades: relatively thin, moderate, and relatively thick. The slurry consistency coefficient is then used in real-time. As input to the slurry mixing consistency classification model, the corresponding output is the slurry mixing consistency level, and the slurry ratio is adjusted based on the result until the slurry mixing consistency reaches a suitable level.

[0103] It should be noted that determining the consistency grade of slurry based on the consistency coefficient of slurry mixing is existing technology, and training a slurry mixing consistency classification model is also existing technology, which will not be elaborated on here.

[0104] The above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention, and should all be included within the protection scope of the present invention.

Claims

1. A method for monitoring the consistency of mortar mixing in autoclaved aerated concrete (AAC) blocks, characterized in that, The method includes: Acquire multiple consecutive overhead images of the slurry, and convert each overhead image to grayscale to obtain the corresponding grayscale image; Edge detection is performed on each grayscale image to obtain the edges in each grayscale image. Based on the grayscale difference between pixels on each edge in each grayscale image, at least one suspected slurry mixing texture edge in each grayscale image is selected. Perform the same texture edge evaluation on every two suspected slurry mixing texture edges between two adjacent grayscale images, and divide all suspected slurry mixing texture edges into at least two sets of texture edges; For any set of texture edges, all suspected slurry stirring texture edges in the set of texture edges are stitched together to obtain the target texture edge. A template image of the slurry stirring trajectory is obtained. The target texture edge is matched with the template texture edge in the template image to obtain the probability index that the target texture edge is a real stirring texture. Obtain the probability index of the target texture edge corresponding to each set of texture edges as the real stirring texture, and obtain at least one stirring texture edge. Based on the texture characteristics and recovery speed of all stirring texture edges, obtain the slurry stirring consistency coefficient, which is used to determine the slurry stirring consistency level. The process involves evaluating the same texture edge for every two suspected slurry mixing texture edges between two adjacent grayscale images, dividing all suspected slurry mixing texture edges into at least two sets of texture edges, including: For the xth suspected slurry stirring texture edge in the i-th grayscale image, in the (i+1)-th grayscale image, obtain suspected slurry stirring texture edges that overlap with the xth suspected slurry stirring texture edge or whose edge endpoints are within a preset number of pixels, and record them as edges to be analyzed. If at least one edge to be analyzed is obtained, obtain the same texture probability index between the xth suspected slurry stirring texture edge and each edge to be analyzed. Obtain the probability index of the same texture between each suspected slurry stirring texture edge in the i-th grayscale image and the corresponding edge to be analyzed in the (i+1)-th grayscale image, and obtain at least one edge pair consisting of two suspected slurry stirring texture edges belonging to the same texture between the i-th grayscale image and the (i+1)-th grayscale image. Obtain all edge pairs between any two adjacent grayscale images. Based on the transitivity of parallel lines, edge pairs belonging to the same texture are grouped into a texture edge set, resulting in at least one texture edge set. Each texture edge set corresponds to one texture.

2. The method for monitoring the consistency of grout mixing for autoclaved aerated concrete blocks according to claim 1, characterized in that, The step of filtering out at least one suspected slurry mixing texture edge in each grayscale image based on the grayscale difference between pixels on each edge in each grayscale image includes: For any edge in any grayscale image, any edge pixel on the edge is taken as the analysis pixel. Two non-edge pixels with symmetrical positions are obtained in the four neighborhoods of the analysis pixel and recorded as the pixels on both sides of the analysis pixel. The absolute value of the grayscale value difference between the pixels on both sides of the analysis pixel is calculated. The reciprocal of the sum of the absolute value of the grayscale value difference and a preset value is taken as the grayscale difference degree on both sides of the edge of the analysis pixel. The grayscale difference between the two sides of each edge pixel on any edge is obtained. The cumulative value of the grayscale difference between the two sides of all edge pixels on any edge is normalized to obtain the probability index of any edge belonging to the slurry stirring texture edge. If the probability index of any edge belonging to the slurry stirring texture edge is greater than the preset probability index threshold, then any edge is regarded as a suspected slurry stirring texture edge. Traverse each edge in each grayscale image to obtain at least one suspected slurry mixing texture edge in each grayscale image.

3. The method for monitoring the consistency of grout mixing for autoclaved aerated concrete blocks according to claim 1, characterized in that, The step of obtaining the probability index of the same texture between the xth suspected slurry stirring texture edge and each edge to be analyzed includes: For any edge to be analyzed, calculate the distance between the edge endpoint of the edge to be analyzed and the edge endpoint of the xth suspected slurry stirring texture edge, obtain the two edge endpoints corresponding to the minimum distance, and record them as edge end points. Take the edge end point as the last edge pixel point, obtain a first preset number of edge pixels on the edge to be analyzed to form a first pixel point sequence, and obtain a first preset number of edge pixels on the xth suspected slurry stirring texture edge to form a second pixel point sequence; Based on the slope of the first pixel sequence and the slope of the second pixel sequence, a first probability index is obtained that any edge to be analyzed and the xth suspected slurry mixing texture edge belong to the same texture in the edge extension direction. With the position of the second pixel sequence as a reference, based on the difference between the fitted position and the actual position of the first pixel sequence, a second probability index is obtained that any edge to be analyzed and the xth suspected slurry mixing texture edge belong to the same texture in relative position. The product of the first probability index and the second probability index is normalized to obtain the same texture probability index between the xth suspected slurry stirring texture edge and any edge to be analyzed.

4. The method for monitoring the consistency of grout mixing for autoclaved aerated concrete blocks according to claim 3, characterized in that, The step of obtaining a first probability index that any edge to be analyzed and the xth suspected slurry mixing texture edge belong to the same texture in the edge extension direction based on the slope of the first pixel sequence and the slope of the second pixel sequence includes: Calculate the slope of the straight line between two adjacent pixels in the first pixel sequence to obtain the average slope, denoted as the first slope. Calculate the slope of the straight line between two adjacent pixels in the second pixel sequence to obtain the average slope, denoted as the second slope. Calculate the absolute value of the difference between the first slope and the second slope. The reciprocal of the sum of the absolute value of the difference and a preset value is used as the first probability index that any edge to be analyzed and the xth suspected slurry mixing texture edge belong to the same texture in the edge extension direction.

5. The method for monitoring the consistency of grout mixing for autoclaved aerated concrete blocks according to claim 3, characterized in that, The step of using the position of the second pixel sequence as a reference, and obtaining a second probability index of whether any edge to be analyzed and the xth suspected slurry mixing texture edge belong to the same texture in relative position based on the difference between the fitted position and the actual position of the first pixel sequence, includes: Based on the abscissa value of each pixel in the second pixel sequence, the ordinate values ​​of all pixels are fitted using the least squares method to obtain a fitting curve. Using the fitting curve, the fitted value of the ordinate of each pixel in the first pixel sequence is obtained. The fitted value of the ordinate of each pixel in the first pixel sequence is subtracted from the actual value to obtain the absolute value of the difference for each pixel. The average value of the absolute values ​​of the differences for all pixels in the first pixel sequence is obtained. The reciprocal of the sum of the average value and the preset value is used as the second probability index that any edge to be analyzed and the xth suspected slurry mixing texture edge belong to the same texture in relative position.

6. The method for monitoring the consistency of grout mixing for autoclaved aerated concrete blocks according to claim 1, characterized in that, The method for obtaining at least one edge pair consisting of two suspected slurry mixing texture edges belonging to the same texture between the i-th grayscale image and the (i+1)-th grayscale image includes: If the probability index of the same texture between any suspected slurry stirring texture edge in the i-th grayscale image and any corresponding edge to be analyzed in the (i+1)-th grayscale image is greater than the preset threshold of the same texture probability index, then any edge to be analyzed is taken as the target edge. If it is detected that any suspected slurry stirring texture edge in the i-th grayscale image corresponds to at least two target edges, then the target edge corresponding to the largest probability index of the same texture is selected and any suspected slurry stirring texture edge in the i-th grayscale image forms a pair of edges.

7. The method for monitoring the consistency of grout mixing for autoclaved aerated concrete blocks according to claim 1, characterized in that, The step of matching the target texture edge with the template texture edge in the template image to obtain a probability index that the target texture edge is a real stirring texture includes: For any template texture edge in the template image, the minimum length is obtained as the size of the sliding window based on the length between the target texture edge and the template texture edge, and the edge corresponding to the minimum length is taken as the target window. Based on the size of the sliding window and the preset sliding step size, the sliding window is slid on the edge corresponding to the maximum length to obtain at least two sliding windows. The normalized correlation coefficient between the target window and each sliding window is calculated, and the maximum normalized correlation coefficient is taken as the matching degree between the target texture edge and the template texture edge. The matching degree between the target texture edge and each template texture edge in the template image is obtained, and the maximum matching degree is used as the probability index that the target texture edge is a real stirring texture.

8. The method for monitoring the consistency of grout mixing for autoclaved aerated concrete blocks according to claim 1, characterized in that, The step of obtaining the probability index that the target texture edge corresponding to each set of texture edges is a real stirring texture, and obtaining at least one stirring texture edge, includes: A preset probability index threshold is obtained. If the probability index of the target texture edge corresponding to any set of texture edges is a real stirring texture is greater than or equal to the probability index threshold, then the target texture edge corresponding to any set of texture edges is determined to be a stirring texture edge.

9. The method for monitoring the consistency of grout mixing for autoclaved aerated concrete blocks according to claim 8, characterized in that, The process of obtaining the slurry mixing consistency coefficient based on the texture characteristics and recovery speed of all stirring texture edges includes: For any stirring texture edge, each suspected slurry stirring texture edge in the texture edge set corresponding to any stirring texture edge is recorded as a marked edge. According to the grayscale image to which each marked edge belongs, the edge pixels of each marked edge are marked in the corresponding grayscale image to obtain the marked pixels. According to the sampling time of the grayscale image to which each marked edge belongs, the duration between the maximum sampling time and the minimum sampling time is obtained and recorded as the stirring texture duration. Obtain the duration of the stirring texture corresponding to each stirring texture edge, and obtain the average duration of the stirring texture. Normalize the average duration of the stirring texture to obtain a normalized value. Based on each suspected slurry stirring texture edge in the texture edge set corresponding to each stirring texture edge, obtain all the marked pixels in each grayscale image, obtain the proportion of marked pixels in each grayscale image, and obtain the average proportion, which is recorded as the texture feature value. The product of the normalized value and the texture feature value is used as the slurry mixing consistency coefficient.

Images