A multifunctional logistics sorting machine

Abstract

The present invention relates to the technical field of logistics item sorting, and specifically to a multifunctional logistics sorting machine, including a conveyor belt, an image acquisition device and an image processing chip, wherein the image acquisition device captures the conveyor belt image, and the image processing chip analyzes the conveyor belt image, thereby realizing the sorting of logistics items, wherein during the analysis process, the first probability that the pixel point is a conveyor belt pixel point is determined by the gray value change and pixel point position of adjacent frames, and the second probability that the gray value is the conveyor belt background gray value is determined according to the first probability and the distribution proximity coefficient, and then the third probability that there is a logistics item occlusion in a row of pixels is determined according to the gray value of each pixel point in a row of pixels and the second probability, and the position and area of ​​the logistics item are determined according to the third probability, thereby sorting the logistics items. The present invention can improve the reliability and accuracy of logistics sorting and enhance the logistics sorting effect.

Description

Technical Field

[0001] The present invention relates to the technical field of logistics article sorting, and in particular to a multifunctional logistics sorting machine. Background Art

[0002] The logistics and warehousing industries are increasingly leaning towards automated solutions to improve efficiency and reduce labor costs. As automated equipment, multifunctional logistics sorters play a key role in sorting, packaging and handling items. The identification of logistics items is an important function of multifunctional logistics sorters.

[0003] In the related art, the logistics items on the conveyor belt are analyzed through grayscale information. In this way, since the grayscale texture of the logistics items themselves is relatively complex, it is impossible to accurately locate and analyze the size of the logistics items when analyzing the logistics items based on the grayscale information, which affects the subsequent sorting effect of the logistics items. The corresponding logistics item identification and sorting reliability and accuracy are poor, and the logistics sorting effect is insufficient. Summary of the invention

[0004] In order to solve the technical problems in the related art that the logistics items cannot be accurately located and the size analyzed, which affects the subsequent sorting effect of the logistics items, the corresponding logistics items identification and sorting reliability and accuracy are poor, and the logistics sorting effect is insufficient, the present invention provides a multifunctional logistics sorting machine, and the technical solutions adopted are as follows:

[0005] The present invention proposes a multifunctional logistics sorting machine, comprising a conveyor belt, an image acquisition device and an image processing chip. The image acquisition device is placed above the conveyor belt, periodically captures images of the conveyor belt, and then sends the images of the conveyor belt to the image processing chip. The image processing chip receives the images of the conveyor belt, analyzes at least two consecutive frames of the images of the conveyor belt, and controls the conveyor belt to sort logistics items according to the analysis results.

[0006] The step of analyzing at least two consecutive frames of the conveyor belt images and controlling the conveyor belt to sort the logistics items according to the analysis results includes:

[0007] Obtain a conveyor belt grayscale image, and determine a first probability that a pixel in the conveyor belt grayscale image of the current frame is a conveyor belt pixel according to a change in grayscale values ​​of pixels at the same position in the conveyor belt grayscale images of the current frame and the previous frame and a distance between the pixel and a central pixel of the conveyor belt grayscale image;

[0008] Take any gray value as the gray value to be measured; determine the distribution proximity coefficient of the gray value to be measured according to the distance between the pixels of the same gray value to be measured; determine the second probability that the gray value to be measured is the gray value of the conveyor belt background according to the first probability of the pixel of the gray value to be measured in the gray image of the conveyor belt of the current frame and the distribution proximity coefficient of the gray value to be measured;

[0009] All pixels perpendicular to the running direction of the conveyor belt and on the same straight line are regarded as a row of pixels; according to the gray value of each pixel in a row of pixels and the distribution of the second probability of the gray value of the pixel, the third probability of the existence of logistics items blocking the pixel in a row is determined;

[0010] Sort the third probability of each row of pixels, and determine the row number of the logistics item according to the sorting result; determine the conveyor belt running speed according to the change of the row number of the logistics item in the conveyor belt image of the current frame and the previous frame; determine the area index of the logistics item according to the third probability of a row of pixels corresponding to the row number of adjacent logistics items; sort the logistics items according to the area index of the logistics item, the row number of the logistics item in the current frame and the conveyor belt running speed.

[0011] Further, the first probability that a pixel point in the conveyor belt grayscale image of the current frame is a conveyor belt pixel point is determined according to the grayscale value change of the pixel point at the same position in the conveyor belt grayscale image of the current frame and the previous frame and the distance of the pixel point from the center pixel point of the conveyor belt grayscale image, including:

[0012] Calculate the absolute value of the grayscale value difference between the pixel at the same position in the current frame conveyor belt grayscale image and the previous frame conveyor belt grayscale image as the grayscale value difference of the pixel at the corresponding position in the current frame conveyor belt grayscale image;

[0013] The distance between the pixel point and the central pixel point in the grayscale image of the conveyor belt of the current frame is normalized to obtain the distance influence factor;

[0014] According to the distance influence factor and the grayscale value difference, the first probability of the pixel point in the grayscale image of the conveyor belt in the current frame is determined, wherein the grayscale value difference is negatively correlated with the first probability, the distance influence factor is positively correlated with the first probability, and the value of the first probability is a normalized value.

[0015] Furthermore, determining the distribution proximity coefficient of the grayscale value to be measured according to the distance between the pixels of the same grayscale value to be measured includes:

[0016] Taking the pixel points whose grayscale values ​​in the conveyor belt grayscale image of the current frame are grayscale values ​​to be measured as the pixel points to be measured;

[0017] The normalized value of the average distance between all the pixels to be tested and another pixel to be tested that is closest to the pixel to be tested is calculated as the distribution proximity coefficient of the grayscale value to be tested.

[0018] Further, the method of determining the second probability that the grayscale value to be measured is the grayscale value of the conveyor belt background according to the first probability of the pixel point of the grayscale value to be measured in the grayscale image of the conveyor belt of the current frame and the distribution proximity coefficient of the grayscale value to be measured includes:

[0019] Calculate the average value of the first probability of the pixel point whose gray value is to be measured in the gray image of the conveyor belt of the current frame to obtain the probability influence coefficient;

[0020] A normalized value of the ratio of the probability influence coefficient to the distribution proximity coefficient is calculated as a second probability that the grayscale value to be measured is the background grayscale value of the conveyor belt.

[0021] Furthermore, the determining, based on the grayscale value of each pixel in a row of pixels and the distribution of the second probability of the grayscale value of the pixel, a third probability of there being logistics item occlusion in a row of pixels includes:

[0022] The distance between each pixel in a row of pixels and the center pixel of the row is taken as the target distance, the normalized value of the target distance is taken as the weight, and the second probability of the gray value of each pixel in a row of pixels is weightedly added to obtain a weighted probability;

[0023] Calculate the gray value variance of all pixels in a row of pixels as the row confusion coefficient;

[0024] Calculate the absolute value of the difference between the maximum value of the second probability of the grayscale value of a row of pixels and the mean value of the second probability sum, and obtain the probability difference of the row of pixels;

[0025] A third probability that there is logistics item occlusion in a row of pixels is calculated based on the weighted probability, the row confusion coefficient and the probability difference.

[0026] Furthermore, the weighted probability is negatively correlated with the third probability, the row confusion coefficient and the probability difference are positively correlated with the third probability, and the value of the third probability is a normalized value.

[0027] Furthermore, the step of sorting the third probabilities of each row of pixels and determining the row number of the logistics item according to the sorting result includes:

[0028] A two-dimensional rectangular coordinate system is constructed with the number of pixel rows as the horizontal coordinate and the third probability as the vertical coordinate. The third probability of each row of pixels is marked in the two-dimensional rectangular coordinate system, and the points of adjacent rows are connected to obtain a probability curve.

[0029] The number of rows in the probability curve whose third probability is greater than a preset probability threshold is taken as the number of rows where the logistics item is located.

[0030] Furthermore, determining the conveyor belt running speed according to the change in the number of rows of logistics items in the conveyor belt image of the current frame and the previous frame includes:

[0031] Calculate the absolute value of the difference between the mean number of rows of all logistics items in the current frame and the mean number of rows of all logistics items in the previous frame to obtain the row moving distance;

[0032] The ratio of the row moving distance to the acquisition time interval between the current frame and the previous frame is taken as the conveyor belt running speed.

[0033] Furthermore, the determining of the area index of the logistics item according to the third probability that the number of rows where adjacent logistics items are located corresponds to a row of pixels includes:

[0034] The normalized value of the sum of the third probabilities of a row of pixels corresponding to the number of rows where the adjacent logistics items are located is used as the area indicator of the corresponding logistics item.

[0035] Furthermore, the conveyor belt further includes a sorting port, and the logistics items are sorted according to the area index of the logistics items, the number of rows where the logistics items are located in the current frame, and the conveyor belt running speed, including:

[0036] The minimum value of the number of rows of adjacent logistics items is used as the initial number of rows of logistics items, and the distance between the initial number of rows and the sorting port of the conveyor belt is calculated to obtain the distance to be transported of the items; the ratio of the distance to be transported of the items to be transported to the running speed of the conveyor belt is calculated to obtain the waiting time for transport;

[0037] Determine the category of the logistics items according to the area index of the logistics items, wherein the categories of the logistics items include first-category items and second-category items, and the logistics items whose area index is greater than a preset area threshold are regarded as first-category items; and the logistics items whose area index is less than or equal to the preset area threshold are regarded as second-category items, and different categories of logistics items correspond to different sorting ports;

[0038] After the waiting time for delivery has passed, the corresponding sorting port 4 is opened according to the category of the logistics items to sort the logistics items.

[0039] The present invention has the following beneficial effects:

[0040] The embodiment of the present invention sets a conveyor belt, an image acquisition device and an image processing chip, uses the image acquisition device to periodically capture images of the conveyor belt, uses the image processing chip to analyze the conveyor belt images, and controls the conveyor belt to sort the logistics items according to the analysis results. During the sorting analysis, since the logistics are sorted by grayscale analysis in the related art, this method cannot effectively target logistics items with complex textures and variable shapes. The present invention avoids the influence of errors caused by direct analysis of grayscale values ​​by analyzing the grayscale changes in the images of adjacent frames and the distance between the pixel points and the center pixel point of the conveyor belt grayscale image, so that the first probability can more accurately characterize the conveyor belt characteristics of the corresponding pixel points; the second probability is determined by combining the first probability and the distribution proximity coefficient, and the distribution of pixel points with the same grayscale value can be specifically analyzed. Since the grayscale value of the conveyor belt pixel points remains basically unchanged, that is, by determining the grayscale of the conveyor belt pixel points, the reverse delineation of the logistics items is achieved, and the second probability can accurately characterize the probability that a grayscale value is the background grayscale value of the conveyor belt; combined with the second probability and the grayscale value, Thus, the third probability of a row of pixels is analyzed. Since the conveyor belt and the logistics items are relatively static, that is, during the corresponding operation, the characteristics of each row remain basically unchanged. Therefore, the third probability can not only characterize whether there are pixels corresponding to logistics items in a row, but also characterize the area information of the logistics items in the row. That is, by sorting the third probabilities of the pixels in all rows, the row number of the logistics items can be determined, and the running speed of the conveyor belt is analyzed according to the changes in the conveyor belt images of two adjacent frames. The logistics items are sorted by combining the area information, running speed and row number of the logistics items themselves. The present invention can accurately locate and analyze the size of logistics items by combining the grayscale characteristics and change characteristics of logistics items on the conveyor belt, avoid the error influence of the texture and grayscale changes of the logistics items themselves on the final sorting results, improve the reliability and accuracy of logistics sorting, and enhance the logistics sorting effect. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] In order to more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the drawings required for use in the embodiments or the prior art descriptions are briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0042] Figure 1 A schematic diagram of the structure of a multifunctional logistics sorting machine provided by one embodiment of the present invention;

[0043] Figure 2 A flow chart of logistics item sorting analysis provided by one embodiment of the present invention;

[0044] Figure 3 A schematic diagram of a conveyor belt conveying logistic items provided by an embodiment of the present invention;

[0045] Figure 4 A schematic diagram of a probability curve provided by an embodiment of the present invention.

[0046] The numbers in the figure are: 1. conveyor belt; 2. image acquisition device; 3. image processing chip; 4. sorting port. DETAILED DESCRIPTION

[0047] In order to further explain the technical means and effects adopted by the present invention to achieve the predetermined invention purpose, the following is a detailed description of the specific implementation, structure, features and effects of a multifunctional logistics sorting machine proposed according to the present invention in combination with the accompanying drawings and preferred embodiments. In the following description, different "one embodiment" or "another embodiment" does not necessarily refer to the same embodiment. In addition, specific features, structures or characteristics in one or more embodiments may be combined in any suitable form.

[0048] Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.

[0049] The specific scheme of the multifunctional logistics sorting machine in the embodiment of the present invention is described in detail below with reference to the accompanying drawings.

[0050] See also Figure 1 The multifunctional logistics sorting machine of the embodiment of the present invention includes a conveyor belt 1, an image acquisition device 2 and an image processing chip 3. The image acquisition device 2 is placed above the conveyor belt 1, periodically captures the conveyor belt image, and then sends the conveyor belt image to the image processing chip 3. The image processing chip 3 receives the conveyor belt image, analyzes at least two consecutive frames of conveyor belt images, and controls the conveyor belt 1 to sort the logistics items according to the analysis results.

[0051] In the embodiment of the present invention, the image acquisition device 2 may be specifically, for example, a high-precision industrial camera, and the image processing chip 3 may be specifically, for example, an FPGA chip, or a corresponding central processing chip, without limitation.

[0052] It is understandable that since the logistics sorting technology in the related art is to determine the size of the logistics items through image analysis and then sort them, the image analysis is usually grayscale analysis. However, the texture of the outer packaging of the logistics items is relatively complex, and when the color of the outer packaging of the logistics items is similar to that of the conveyor belt, it affects the recognition and sorting effect, and the corresponding logistics item recognition and sorting reliability is poor. Therefore, the present invention combines grayscale and conveyor belt changes to achieve accurate positioning and size analysis of logistics items and achieve effective sorting.

[0053] The specific scheme for sorting and analyzing logistics items in an embodiment of the present invention is described in detail below with reference to the accompanying drawings.

[0054] See also Figure 2 , which shows a flow chart of logistics item sorting analysis provided by an embodiment of the present invention, including:

[0055] S101: Obtain a conveyor belt grayscale image, and determine a first probability that a pixel in the conveyor belt grayscale image of the current frame is a conveyor belt pixel based on a change in grayscale values ​​of pixels at the same position in the conveyor belt grayscale images of the current frame and the previous frame and a distance from the pixel to a center pixel of the conveyor belt grayscale image.

[0056] In the embodiment of the present invention, the conveyor belt grayscale image is an image obtained after the conveyor belt image is preprocessed, wherein the image preprocessing mainly includes image grayscale and image denoising processing. Image preprocessing is a technology well known in the present invention and will not be further limited or elaborated on.

[0057] In an embodiment of the present invention, the logistics items on the conveyor belt will be displaced as the conveyor belt runs. Compared with a fixed image acquisition device, the displacement of the logistics can be effectively displayed in the grayscale images of the conveyor belt in different frames. Therefore, the moving area can be determined based on the changes in the grayscale values ​​of pixels in adjacent frames.

[0058] Furthermore, in some embodiments of the present invention, a first probability that a pixel in the conveyor belt grayscale image of the current frame is a conveyor belt pixel is determined based on a change in grayscale values ​​of pixels at the same position in the conveyor belt grayscale images of the current frame and the previous frame and a distance between the pixel and the center pixel of the conveyor belt grayscale image, including: calculating an absolute value of a grayscale value difference between pixels at the same position in the conveyor belt grayscale image of the current frame and the previous frame as a grayscale value difference of pixels at corresponding positions in the conveyor belt grayscale image of the current frame; normalizing the distance between the pixel and the center pixel in the conveyor belt grayscale image of the current frame to obtain a distance influence factor; determining a first probability of the pixel in the conveyor belt grayscale image of the current frame based on the distance influence factor and the grayscale value difference, wherein the grayscale value difference is negatively correlated with the first probability, the distance influence factor is positively correlated with the first probability, and the value of the first probability is the normalized value.

[0059] Among them, a positive correlation means that the dependent variable will increase as the independent variable increases, and the dependent variable will decrease as the independent variable decreases. The specific relationship can be a multiplication relationship, an addition relationship, or the power of an exponential function, which is determined by actual application; a negative correlation means that the dependent variable will decrease as the independent variable increases, and the dependent variable will increase as the independent variable decreases. It can be a subtraction relationship, a division relationship, etc., which is determined by actual application.

[0060] In one embodiment of the present invention, the normalization processing can be specifically, for example, maximum and minimum value normalization processing, and the normalization in subsequent steps can all adopt maximum and minimum value normalization processing. In other embodiments of the present invention, other normalization methods can be selected according to the specific range of numerical values, which will not be described in detail.

[0061] In the embodiment of the present invention, the gray value difference is negatively correlated with the first probability, the distance influence factor is positively correlated with the first probability, and the value of the first probability is a normalized value. The sum of the gray value difference and the preset first hyperparameter can be used as the denominator, and the distance influence factor can be used as the numerator of the fraction, and the first probability can be calculated after normalization. Among them, the preset first hyperparameter is a safety value set to prevent the denominator from being 0. Optionally, the specific value of the preset first hyperparameter can be 0.01, and there is no limitation on this.

[0062] The grayscale value difference of a pixel is the grayscale value change of the pixel at the same position in the current frame conveyor belt grayscale image and the previous frame conveyor belt grayscale image. Since there is inevitable white noise in the image, that is, the corresponding image noise, the reliability of realizing logistics item detection only by the grayscale value change of the pixel in the image is insufficient. Therefore, the present invention combines the distance influencing factor for further analysis.

[0063] It is understandable that logistics items are usually placed in the middle of the conveyor belt, that is, the grayscale value change in the middle area of ​​the conveyor belt is more likely to be the area covered by logistics items. Therefore, the distance influencing factor is determined by the distance between the pixel point and the central pixel point of the entire conveyor belt grayscale image, and the grayscale value difference is weightedly analyzed by the distance influencing factor to obtain the first probability that the pixel point in the current frame conveyor belt grayscale image is the conveyor belt pixel point. Each pixel point corresponds to a first probability.

[0064] S102: Take any gray value as the gray value to be measured; determine the distribution proximity coefficient of the gray value to be measured based on the distance between pixels of the same gray value to be measured; determine the second probability that the gray value to be measured is the gray value of the conveyor belt background based on the first probability of the pixel of the gray value to be measured in the conveyor belt gray image of the current frame and the distribution proximity coefficient of the gray value to be measured.

[0065] Furthermore, in some embodiments of the present invention, the distribution proximity coefficient of the grayscale value to be measured is determined based on the distance between pixel points with the same grayscale value to be measured, including: taking the pixel points in the grayscale image of the current frame conveyor belt whose grayscale value is the grayscale value to be measured as the pixel points to be measured; calculating the normalized value of the mean distance between all the pixel points to be measured and another pixel point to be measured that is closest to the pixel point to be measured as the distribution proximity coefficient of the grayscale value to be measured.

[0066] It should be noted that in actual logistics transportation, the surface colors and patterns of the outer packaging or express packaging of different logistics items are different, which also leads to differences in the grayscale values ​​of different logistics items. However, in comparison, the grayscale values ​​of the conveyor belt are very close, so the distribution of the background grayscale value of the conveyor belt is in a larger area, while the distribution of the grayscale values ​​of different logistics items is often in a smaller area or scattered.

[0067] In the embodiment of the present invention, the distribution of the conveyor belts is relatively concentrated, while the distribution of the logistics items is relatively discrete. Therefore, the distribution of the corresponding grayscale value to be measured can be characterized by the distribution proximity coefficient. The embodiment of the present invention can connect the pixel point to be measured and another pixel point to be measured that is closest to it, and calculate the distance between the two pixels, and then replace the pixel point to be measured until all the pixels to be measured are traversed, and the normalized value of the distance mean is obtained as the distribution proximity coefficient of the grayscale value to be measured. The smaller the distribution proximity coefficient, the denser the distribution of the corresponding grayscale value to be measured, that is, the grayscale value to be measured is more likely to be the grayscale value corresponding to the conveyor belt.

[0068] Furthermore, in some embodiments of the present invention, the second probability that the grayscale value to be measured is the grayscale value of the conveyor belt background is determined based on the first probability of the pixel points of the grayscale value to be measured in the conveyor belt grayscale image of the current frame and the distribution proximity coefficient of the grayscale value to be measured, including: calculating the average of the first probability of the pixel points of the grayscale value to be measured in the conveyor belt grayscale image of the current frame to obtain the probability influence coefficient; calculating the normalized value of the ratio of the probability influence coefficient to the distribution proximity coefficient as the second probability that the grayscale value to be measured is the grayscale value of the conveyor belt background.

[0069] In an embodiment of the present invention, the mean of the first probabilities of pixel points with the same grayscale value is taken as the probability influence coefficient, that is, the larger the probability influence coefficient is, the more likely the corresponding grayscale value is to be the grayscale value of the conveyor belt, and the normalized value of the ratio of the probability influence coefficient to the distribution proximity coefficient is calculated as the second probability that the grayscale value to be measured is the background grayscale value of the conveyor belt, that is, the larger the probability 1 is, the denser the distribution of pixels of the corresponding grayscale value is, and the larger the first probability is, therefore, the grayscale value is more likely to represent the grayscale value of the conveyor belt itself, that is, the background grayscale value of the conveyor belt.

[0070] S103: All pixels perpendicular to the running direction of the conveyor belt and on the same straight line are regarded as a row of pixels; according to the gray value of each pixel in a row of pixels and the distribution of the second probability of the gray value of the pixel, the third probability of the existence of logistics items blocking the row of pixels is determined.

[0071] Among them, the logistics items move on the conveyor belt along the running direction of the conveyor belt, and a row of pixels is determined in the direction perpendicular to the running direction of the conveyor belt, that is, the pixels on a straight line perpendicular to the running direction of the conveyor belt are collectively referred to as a row of pixels. In a row of pixels containing logistics items, the pixels corresponding to the logistics items and the pixels of the conveyor belt itself will produce grayscale value changes. The grayscale value change can detect the possibility of occlusion of logistics items in a row of pixels, that is, the third probability.

[0072] Furthermore, in some embodiments of the present invention, a third probability that there is logistics item occlusion in a row of pixels is determined based on the grayscale value of each pixel in a row of pixels and the distribution of the second probability of the pixel grayscale value, including: taking the distance between each pixel in a row of pixels and the center pixel of the row as the target distance, taking the normalized value of the target distance as the weight, and weightedly adding the second probability of the grayscale value of each pixel in a row of pixels to obtain a weighted probability; calculating the grayscale value variance of all pixels in a row of pixels as the row confusion coefficient; calculating the absolute value of the difference between the maximum value of the second probability of the grayscale value of a row of pixels and the mean of the second probability and value to obtain the probability difference for a row of pixels; and calculating the third probability that there is logistics item occlusion in a row of pixels based on the weighted probability, the row confusion coefficient and the probability difference.

[0073] In the embodiment of the present invention, combined with the feature that the closer the pixel is to the center, the more likely it is to be the pixel corresponding to a logistics item, the distance between each pixel in a row of pixels and the center pixel of the row is calculated to obtain the target distance. The larger the target distance, the smaller the possibility of corresponding to a logistics item. The normalized value of the target distance is used as the weight, and the second probability of the grayscale value of each pixel in a row of pixels is weightedly added to obtain the weighted probability.

[0074] In an embodiment of the present invention, the degree of grayscale disorder can be determined based on the grayscale value distribution of each row of pixels. It can be understood that the grayscale value disorder of logistics items is relatively large, while the grayscale value disorder of pixels corresponding to the conveyor belt background is relatively small. Therefore, by calculating the grayscale value variance of all pixels in a row of pixels as the row disorder coefficient, the larger the row disorder coefficient, the more disordered the corresponding grayscale value distribution is.

[0075] In an embodiment of the present invention, the maximum value of the second probability of the grayscale values ​​of a row of pixels is calculated, which indicates the second probability of the pixel most likely to represent the conveyor belt in the corresponding row of pixels, and the sum average of the second probabilities of the grayscale values ​​of a row of pixels represents the overall situation of the row of pixels. Thus, the probability difference of a row of pixels is obtained by calculating the maximum value of the second probability and the sum average of the second probabilities. The smaller the probability difference, the more pixels representing the conveyor belt background in the corresponding row of pixels, and the larger the probability difference, the fewer pixels representing the conveyor belt background in the corresponding row of pixels.

[0076] Furthermore, in some embodiments of the present invention, the weighted probability is negatively correlated with the third probability, the row confusion coefficient and the probability difference are positively correlated with the third probability, and the value of the third probability is a normalized value.

[0077] In the embodiment of the present invention, the larger the row chaos coefficient is, the more chaotic the corresponding gray value distribution is, and the more likely there is a logistics item occlusion; the larger the probability difference is, the fewer pixels representing the conveyor belt background are in the corresponding row of pixels, that is, the more likely there is a logistics item occlusion, thus, the row chaos coefficient and the probability difference are positively correlated with the third probability. The larger the weighted probability is, the greater the probability that the corresponding pixel is the conveyor belt background, thus, the weighted probability is negatively correlated with the third probability.

[0078] Specifically, the embodiment of the present invention can calculate the product of the row confusion coefficient and the probability difference, and use the product value as the numerator, use the sum of the weighted probability and the preset second hyperparameter as the denominator, normalize the fraction, and calculate the third probability. The preset second hyperparameter is a safety value set to prevent the denominator from being 0. Optionally, the specific value of the preset second hyperparameter can be 0.01, which is not limited.

[0079] Among them, the third probability represents the probability that there is occlusion of logistics items in a row of pixels. The larger the third probability is, the more likely it is that there is occlusion of logistics items in the corresponding row, and the greater the severity of the occlusion is. Therefore, the state of the entire conveyor belt can be analyzed based on the third probability.

[0080] S104: Sort the third probability of each row of pixels, and determine the row number of the logistics item according to the sorting result; determine the conveyor belt running speed according to the change in the row number of the logistics item in the conveyor belt image of the current frame and the previous frame; determine the area index of the logistics item according to the third probability of a row of pixels corresponding to the row number of adjacent logistics items; sort the logistics items according to the area index of the logistics items, the row number of the logistics items in the current frame and the conveyor belt running speed.

[0081] It can be understood that since the third probability represents the probability of logistics items blocking a row of pixels, and the larger the third probability value, the more serious the degree of logistics items blocking, the third probability of each row of pixels is sorted, and the running state of the entire conveyor belt at the current moment can be analyzed according to the sorting result. It should be noted that in the embodiment of the present invention, the closer the row is to the end of the conveyor belt, the smaller the corresponding row number is.

[0082] Furthermore, in some embodiments of the present invention, the third probability of each row of pixels is sorted, and the row number of the logistics item is determined according to the sorting result, including: constructing a two-dimensional rectangular coordinate system with the row number of pixel points as the horizontal coordinate and the third probability as the vertical coordinate, marking the third probability of each row of pixels in the two-dimensional rectangular coordinate system, and connecting the points of adjacent rows to obtain a probability curve; the number of rows in the probability curve whose third probability is greater than a preset probability threshold is taken as the row number where the logistics item is located.

[0083] The probability curve can be specifically as follows: Figure 3 and Figure 4 As shown, Figure 3 A schematic diagram of a conveyor belt conveying logistic items provided by an embodiment of the present invention; Figure 4 A schematic diagram of a probability curve provided by an embodiment of the present invention. Figure 3 In the figure, "C", "D" and "E" represent three different logistics items. Figure 4 The probability curves in the graph correspond to the probability curves of the three logistics items "C", "D" and "E" respectively. It can be understood that in the normal conveyor belt area, the corresponding third probability value is small, and in the case of obstruction by logistics items, the inclination of the corresponding probability curve changes according to the placement and shape of the logistics items. Therefore, the embodiment of the present invention takes the number of rows in the probability curve whose third probability is greater than the preset probability threshold as the number of rows where the logistics items are located, thereby analyzing the number of rows where the logistics items are located.

[0084] Among them, the preset probability threshold is the threshold value of the third probability. In the embodiment of the present invention, the preset probability threshold can be specifically 0.75, that is, the number of rows where the third probability is greater than 0.75 is taken as the number of rows where the logistics item is located.

[0085] Furthermore, in some embodiments of the present invention, the conveyor belt running speed is determined based on the change in the number of rows of logistics items in the conveyor belt images of the current frame and the previous frame, including: calculating the absolute value of the difference between the mean number of rows of all logistics items in the current frame and the mean number of rows of all logistics items in the previous frame to obtain the row moving distance; and taking the ratio of the row moving distance to the acquisition time interval between the current frame and the previous frame as the conveyor belt running speed.

[0086] In the embodiment of the present invention, since the placement and shape of items in adjacent frames will not change much, the difference in the mean values ​​of the row numbers of all logistics items in adjacent frames is calculated, which can characterize the length of the corresponding conveyor belt running between frames, that is, the row moving distance, and then, the ratio of the row moving distance to the acquisition time interval between the current frame and the previous frame is used as the conveyor belt running speed.

[0087] Furthermore, in some embodiments of the present invention, the area index of the logistics items is determined according to the third probability that the number of adjacent logistics items in the row corresponds to a row of pixels, including: normalizing the sum of the third probabilities of the number of adjacent logistics items in the row corresponding to a row of pixels as the area index of the corresponding logistics item.

[0088] In the embodiment of the present invention, since the larger the item is, the more rows the corresponding logistics items are in, and the larger the sum of the third probabilities is, the logistics items cover a complete area on the conveyor belt. Therefore, the number of rows where adjacent logistics items are located can represent the number of rows corresponding to the same logistics item. The present invention can calculate the normalized sum of the third probabilities of a row of pixels corresponding to the number of rows where adjacent logistics items are located. The larger the value is, the larger the area covered by the corresponding logistics items on the conveyor belt is. Therefore, it is used as the area indicator of the logistics items.

[0089] Furthermore, in some embodiments of the present invention, the conveyor belt also includes a sorting port 4, which sorts the logistics items according to their area index, the number of rows in the current frame and the conveyor belt running speed, including: taking the minimum value of the number of rows of adjacent logistics items as the initial number of rows of the logistics items, calculating the distance between the initial number of rows and the conveyor belt sorting port to obtain the distance to be transported of the items; calculating the ratio of the distance to be transported of the items to be transported and the conveyor belt running speed to obtain the waiting time for transport; determining the category of the logistics items according to the area index of the logistics items, wherein the categories of the logistics items include first-category items and second-category items, and taking the logistics items with an area index greater than a preset area threshold as first-category items; taking the logistics items with an area index less than or equal to the preset area threshold as second-category items, and different categories of logistics items correspond to different sorting ports; after the waiting time for transport, opening the corresponding sorting port 4 according to the category of the logistics items, and sorting the logistics items.

[0090] It can be understood that the sorting port can lead to a plurality of different sorting areas, and the sorting area to which the logistics items lead can be controlled by controlling the internal machinery of the sorting port.

[0091] The preset area threshold is a threshold value of the area index. In the embodiment of the present invention, the preset area threshold may be specifically, for example, 5.5, or may be adjusted according to actual detection requirements, without limitation.

[0092] In an embodiment of the present invention, two sorting areas can be set up, namely, a sorting area corresponding to the first category of items and a sorting area corresponding to the second category of items, and then the logistics items whose area index is greater than the preset area threshold are regarded as the first category of items; and the logistics items whose area index is less than or equal to the preset area threshold are regarded as the second category of items.

[0093] In an embodiment of the present invention, each logistics item on the conveyor belt also has a corresponding waiting time for transmission, and the waiting time for transmission represents the time when the corresponding logistics item is about to be transmitted on the conveyor belt. Therefore, when the waiting time for transmission is reached, the present invention controls the sorting port to adjust to the sorting area of ​​the corresponding category according to the category of the corresponding logistics item, thereby realizing the sorting of the logistics items.

[0094] In the embodiment of the present invention, the sorting principle is to sort directly according to the area size of the logistics items. Of course, in other embodiments of the present invention, specific classification can also be carried out according to the morphological characteristics, packaging grayscale difference characteristics and other characteristics of the logistics items, and there is no limitation to this.

[0095] The embodiment of the present invention sets a conveyor belt, an image acquisition device and an image processing chip, uses the image acquisition device to periodically capture images of the conveyor belt, uses the image processing chip to analyze the conveyor belt images, and controls the conveyor belt to sort the logistics items according to the analysis results. During the sorting analysis, since the logistics are sorted by grayscale analysis in the related art, this method cannot effectively target logistics items with complex textures and variable shapes. The present invention avoids the influence of errors caused by direct analysis of grayscale values ​​by analyzing the grayscale changes in the images of adjacent frames and the distance between the pixel points and the center pixel point of the conveyor belt grayscale image, so that the first probability can more accurately characterize the conveyor belt characteristics of the corresponding pixel points; the second probability is determined by combining the first probability and the distribution proximity coefficient, and the distribution of pixel points with the same grayscale value can be specifically analyzed. Since the grayscale value of the conveyor belt pixel points remains basically unchanged, that is, by determining the grayscale of the conveyor belt pixel points, the reverse delineation of the logistics items is achieved, and the second probability can accurately characterize the probability that a grayscale value is the background grayscale value of the conveyor belt; combined with the second probability and the grayscale value, Thus, the third probability of a row of pixels is analyzed. Since the conveyor belt and the logistics items are relatively static, that is, during the corresponding operation, the characteristics of each row remain basically unchanged. Therefore, the third probability can not only characterize whether there are pixels corresponding to logistics items in a row, but also characterize the area information of the logistics items in the row. That is, by sorting the third probabilities of the pixels in all rows, the row number of the logistics items can be determined, and the running speed of the conveyor belt is analyzed according to the changes in the conveyor belt images of two adjacent frames. The logistics items are sorted by combining the area information, running speed and row number of the logistics items themselves. The present invention can accurately locate and analyze the size of logistics items by combining the grayscale characteristics and change characteristics of logistics items on the conveyor belt, avoid the error influence of the texture and grayscale changes of the logistics items themselves on the final sorting results, improve the reliability and accuracy of logistics sorting, and enhance the logistics sorting effect.

[0096] It should be noted that the sequence of the above embodiments of the present invention is only for description and does not represent the advantages and disadvantages of the embodiments. The processes depicted in the accompanying drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0097] The various embodiments in this specification are described in a progressive manner, and the same or similar parts between the various embodiments can be referenced to each other, and each embodiment focuses on the differences from other embodiments.

Claims

1. A multifunctional logistics sorting machine, comprising a conveyor belt (1), an image acquisition device (2) and an image processing chip (3), characterized in that: The image acquisition device (2) is placed above the conveyor belt (1), periodically captures images of the conveyor belt, and then sends the images of the conveyor belt to the image processing chip (3). The image processing chip (3) receives the images of the conveyor belt, analyzes at least two consecutive frames of the conveyor belt images, and controls the conveyor belt (1) to sort the logistics items according to the analysis results. The step of analyzing at least two consecutive frames of the conveyor belt images and controlling the conveyor belt to sort the logistics items according to the analysis results includes: Obtain a conveyor belt grayscale image, and determine a first probability that a pixel in the conveyor belt grayscale image of the current frame is a conveyor belt pixel according to a change in grayscale values ​​of pixels at the same position in the conveyor belt grayscale images of the current frame and the previous frame and a distance between the pixel and a central pixel of the conveyor belt grayscale image; Take any gray value as the gray value to be measured; determine the distribution proximity coefficient of the gray value to be measured according to the distance between the pixels of the same gray value to be measured; determine the second probability that the gray value to be measured is the gray value of the conveyor belt background according to the first probability of the pixel of the gray value to be measured in the gray image of the conveyor belt of the current frame and the distribution proximity coefficient of the gray value to be measured; All pixels perpendicular to the running direction of the conveyor belt and on the same straight line are regarded as a row of pixels; according to the gray value of each pixel in a row of pixels and the distribution of the second probability of the gray value of the pixel, the third probability of the existence of logistics items blocking the pixel in a row is determined; Sort the third probability of each row of pixels, and determine the row number of the logistics item according to the sorting result; determine the conveyor belt running speed according to the change of the row number of the logistics item in the conveyor belt image of the current frame and the previous frame; determine the area index of the logistics item according to the third probability of a row of pixels corresponding to the row number of adjacent logistics items; sort the logistics items according to the area index of the logistics item, the row number of the logistics item in the current frame and the conveyor belt running speed; The method of determining a first probability that a pixel point in the conveyor belt grayscale image of the current frame is a conveyor belt pixel point according to a change in the grayscale value of a pixel point at the same position in the conveyor belt grayscale image of the current frame and the previous frame and a distance between the pixel point and a central pixel point of the conveyor belt grayscale image includes: Calculate the absolute value of the grayscale value difference between the pixel at the same position in the current frame conveyor belt grayscale image and the previous frame conveyor belt grayscale image as the grayscale value difference of the pixel at the corresponding position in the current frame conveyor belt grayscale image; The distance between the pixel point and the central pixel point in the grayscale image of the conveyor belt of the current frame is normalized to obtain the distance influence factor; Determine a first probability of a pixel in the grayscale image of the conveyor belt of the current frame according to the distance influence factor and the grayscale value difference, wherein the grayscale value difference is negatively correlated with the first probability, the distance influence factor is positively correlated with the first probability, and the value of the first probability is a normalized value; The step of determining the distribution proximity coefficient of the grayscale value to be measured according to the distance between the pixels having the same grayscale value to be measured comprises: Taking the pixel points whose grayscale values ​​in the conveyor belt grayscale image of the current frame are grayscale values ​​to be measured as the pixel points to be measured; Calculate the normalized value of the average distance between all the pixels to be tested and another pixel to be tested that is closest to the pixel to be tested as the distribution proximity coefficient of the grayscale value to be tested; Determining the second probability that the grayscale value to be measured is the background grayscale value of the conveyor belt according to the first probability of the pixel point of the grayscale value to be measured in the grayscale image of the conveyor belt in the current frame and the distribution proximity coefficient of the grayscale value to be measured includes: Calculate the average value of the first probability of the pixel point whose gray value is to be measured in the gray image of the conveyor belt of the current frame to obtain the probability influence coefficient; Calculate a normalized value of the ratio of the probability influence coefficient to the distribution proximity coefficient as a second probability that the gray value to be measured is the background gray value of the conveyor belt; The determining, according to the gray value of each pixel in a row of pixels and the distribution of the second probability of the gray value of the pixel, a third probability that there is logistics item occlusion in a row of pixels comprises: The distance between each pixel in a row of pixels and the center pixel of the row is taken as the target distance, the normalized value of the target distance is taken as the weight, and the second probability of the gray value of each pixel in a row of pixels is weightedly added to obtain a weighted probability; Calculate the gray value variance of all pixels in a row of pixels as the row confusion coefficient; Calculate the absolute value of the difference between the maximum value of the second probability of the grayscale value of a row of pixels and the mean value of the second probability sum, and obtain the probability difference of the row of pixels; A third probability that there is logistics item occlusion in a row of pixels is calculated based on the weighted probability, the row confusion coefficient and the probability difference.

2. A multifunctional logistics sorting machine as claimed in claim 1, characterized in that: The weighted probability is negatively correlated with the third probability, the row confusion coefficient and the probability difference are positively correlated with the third probability, and the value of the third probability is a normalized value.

3. A multifunctional logistics sorting machine as claimed in claim 1, characterized in that: The step of sorting the third probabilities of each row of pixels and determining the row number of the logistics item according to the sorting result includes: A two-dimensional rectangular coordinate system is constructed with the number of pixel rows as the horizontal coordinate and the third probability as the vertical coordinate. The third probability of each row of pixels is marked in the two-dimensional rectangular coordinate system, and the points of adjacent rows are connected to obtain a probability curve. The number of rows in the probability curve whose third probability is greater than a preset probability threshold is taken as the number of rows where the logistics item is located.

4. A multifunctional logistics sorting machine as claimed in claim 3, characterized in that: The step of determining the conveyor belt running speed according to the change in the number of rows of the logistics items in the conveyor belt image of the current frame and the previous frame includes: Calculate the absolute value of the difference between the mean number of rows of all logistics items in the current frame and the mean number of rows of all logistics items in the previous frame to obtain the row moving distance; The ratio of the row moving distance to the acquisition time interval between the current frame and the previous frame is taken as the conveyor belt running speed.

5. A multifunctional logistics sorting machine as claimed in claim 1, characterized in that: The determining the area index of the logistics item according to the third probability that the number of rows where adjacent logistics items are located corresponds to a row of pixels includes: The normalized value of the sum of the third probabilities of a row of pixels corresponding to the number of rows where the adjacent logistics items are located is used as the area indicator of the corresponding logistics item.

6. A multifunctional logistics sorting machine as claimed in claim 1, characterized in that: The conveyor belt also includes a sorting port (4), and the logistics items are sorted according to the area index of the logistics items, the number of rows of the logistics items in the current frame and the running speed of the conveyor belt, including: The minimum value of the number of rows of adjacent logistics items is used as the initial number of rows of logistics items, and the distance between the initial number of rows and the sorting port of the conveyor belt is calculated to obtain the distance to be transported of the items; the ratio of the distance to be transported of the items to be transported to the running speed of the conveyor belt is calculated to obtain the waiting time for transport; Determine the category of the logistics items according to the area index of the logistics items, wherein the categories of the logistics items include first-category items and second-category items, and the logistics items whose area index is greater than a preset area threshold are regarded as first-category items; and the logistics items whose area index is less than or equal to the preset area threshold are regarded as second-category items, and different categories of logistics items correspond to different sorting ports; After the waiting time for transmission has passed, the corresponding sorting port (4) is opened according to the category of the logistics items to sort the logistics items.

Images