Cross-belt sorting anomaly detection method, device, computer equipment and storage medium

Abstract

The present application relates to a method, device, computer equipment and storage medium for detecting anomalies in cross-belt sorting. The method comprises: obtaining a cargo movement trajectory; determining characteristic points of the cargo movement trajectory; and judging whether there is an abnormality in the sorting of the cargo by the cross-belt by comparing the characteristic points with a floating threshold; the floating threshold is a threshold obtained by obtaining multiple cargo movement trajectory samples, determining the sample characteristic points of each cargo movement trajectory sample, and performing outlier detection on the multiple sample characteristic points. This method can detect abnormal cargo movement trajectories, determine abnormally sorted cargo, and manually correct the abnormally sorted cargo, thereby improving the accuracy of cargo sorting.

Description

Technical Field

[0001] The present application relates to the field of automation technology, and in particular to a method, device, computer equipment and storage medium for detecting anomalies in cross-belt sorting. Background Art

[0002] With the rapid development of e-commerce and logistics industries, traditional manual sorting methods are gradually unable to meet the growing demand for cargo sorting. Therefore, automated sorting equipment has emerged. Automated sorting equipment usually uses cross-belts for sorting. In cross-belt sorting equipment, the main drive belt conveyor and the trolley carrying a small belt conveyor are connected together. When the trolley moves to the specified sorting position, the goods are sorted and sent out through the rotation of the belt.

[0003] However, due to the differences in weight and shape of goods, and the fact that cross-belt sorting equipment usually sorts in a unified manner, before the goods reach the specified sorting position, the goods may fall to other sorting ports prematurely due to equipment shaking, or because the goods are too light, they may fall to other sorting ports due to inertia during the falling process, resulting in incorrect sorting of goods. The cross-belt sorting equipment will not identify the sorting errors, and thus cannot correct the sorting errors manually, resulting in reduced accuracy in sorting.

[0004] Therefore, the current cross-belt sorting technology has the problem of sorting errors, which easily leads to a decrease in the accuracy of cargo sorting. Summary of the Invention

[0005] Based on this, it is necessary to provide a cross-belt sorting anomaly detection method, device, computer equipment and storage medium that can improve the accuracy of cargo sorting in response to the above technical problems.

[0006] A method for detecting anomalies in cross-belt sorting, the method comprising:

[0007] Obtaining a cargo movement trajectory; the cargo movement trajectory is the movement trajectory of the cargo when the cargo is sorted by the cross belt;

[0008] Determine characteristic points of the cargo movement trajectory; the characteristic points include an inflection point value of the cargo movement trajectory and an intersection value between the cargo movement trajectory and a cargo sorting port;

[0009] By comparing the feature points with the floating item threshold, it is determined whether there is any abnormality in the sorting of the goods by the cross belt; the floating item threshold is a threshold obtained by obtaining multiple goods movement trajectory samples, determining the sample feature points of each of the goods movement trajectory samples, and performing outlier detection on the multiple sample feature points.

[0010] In one embodiment, the cargo movement trajectory includes a plurality of trajectory points, the plurality of trajectory points include a first trajectory point and a second trajectory point, and the first trajectory point and the second trajectory point are adjacent; and determining the characteristic points of the cargo movement trajectory includes:

[0011] Obtaining a function analytical expression of the cross band;

[0012] Substituting the first trajectory point into the function expression to obtain a first function value, and substituting the second trajectory point into the function expression to obtain a second function value;

[0013] If the first function value and the second function value have opposite signs, the intersection value is obtained according to the horizontal coordinate of the first trajectory point.

[0014] In one embodiment, the cargo movement trajectory further includes a third trajectory point, a fourth trajectory point, and a fifth trajectory point, wherein the third trajectory point, the fourth trajectory point, and the fifth trajectory point are sequentially adjacent and located before the intersection point; and determining the characteristic points of the cargo movement trajectory further includes:

[0015] respectively acquiring coordinates of the third track point, the fourth track point, and the fifth track point, wherein the coordinates include a horizontal coordinate and a vertical coordinate;

[0016] Obtaining a longitudinal coordinate distance according to the longitudinal coordinate of the third track point and the longitudinal coordinate of the fifth track point;

[0017] If the ordinate distance does not exceed a preset ordinate distance threshold, the inflection point value is obtained according to the abscissa of the fourth trajectory point.

[0018] In one embodiment, the method further comprises:

[0019] According to a preset point distance threshold, each of the sample feature points is detected to determine whether it is an outlier point, and an initial outlier point set and an initial dense point set are obtained;

[0020] Performing circle fitting on the initial dense point set to obtain a fitting circle radius;

[0021] Updating the point distance threshold according to the fitting circle radius to obtain an updated point distance threshold;

[0022] According to the updated point distance threshold, detecting whether each of the sample feature points is an outlier, and obtaining an updated dense point set;

[0023] The floating point threshold is obtained according to the updated dense point set.

[0024] In one embodiment, detecting whether each of the sample feature points is an outlier based on a preset point distance threshold to obtain an initial outlier point set and an initial dense point set includes:

[0025] Selecting a point to be detected from the plurality of sample feature points;

[0026] Counting the distances between non-detection points and the points to be detected; the non-detection points are all points other than the points to be detected in the multiple sample feature points;

[0027] If the point distance is less than the point distance threshold, the non-detection point is determined to be a close-distance point;

[0028] Counting the proportion of the close-distance points in the non-detection points;

[0029] If the ratio is less than a preset ratio threshold, the point to be detected is determined to be an outlier; all outliers in the plurality of sample feature points constitute the initial outlier set;

[0030] If the ratio is not less than the ratio threshold, the point to be detected is determined to be a cluster point; all cluster points in the multiple sample feature points constitute the initial dense point set.

[0031] In one embodiment, the floating threshold includes a first floating threshold and a second floating threshold; and obtaining the floating threshold according to the updated dense point set includes:

[0032] Obtaining the maximum value of the horizontal coordinate, the maximum value of the vertical coordinate, the minimum value of the horizontal coordinate, and the minimum value of the vertical coordinate in the updated dense point set;

[0033] A first intersection threshold is obtained according to the maximum value of the horizontal coordinate, a first inflection point threshold is obtained according to the maximum value of the vertical coordinate, a second intersection threshold is obtained according to the minimum value of the horizontal coordinate, and a second inflection point threshold is obtained according to the minimum value of the vertical coordinate;

[0034] The first floating threshold is obtained according to the first intersection threshold and the first inflection point threshold, and the second floating threshold is obtained according to the second intersection threshold and the second inflection point threshold.

[0035] In one embodiment, comparing the feature point with a floating item threshold to determine whether there is an abnormality in the sorting of the goods by the cross belt includes:

[0036] If the intersection value is greater than the first intersection threshold, and the inflection point value is greater than the first inflection point threshold, the cargo is determined to be a delayed floating item;

[0037] If the intersection value is smaller than the second intersection threshold value, and the inflection point value is smaller than the second inflection point threshold value, it is determined that the cargo is a floating item that has fallen in advance.

[0038] A cross-belt sorting anomaly detection device, comprising:

[0039] An acquisition module, configured to acquire a movement trajectory of goods; the movement trajectory of the goods is the movement trajectory of the goods when the goods are sorted by the cross belt;

[0040] A feature point determination module is used to determine the feature points of the cargo movement trajectory; the feature points include the inflection point value of the cargo movement trajectory and the intersection value between the cargo movement trajectory and the cargo sorting port;

[0041] A judgment module is configured to determine whether there is any abnormality in the sorting of the goods by the cross belt by comparing the feature points with a floating item threshold; the floating item threshold is a threshold value obtained by obtaining multiple goods movement trajectory samples, determining the sample feature points of each of the goods movement trajectory samples, and performing outlier detection on the multiple sample feature points.

[0042] A computer device includes a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, the following steps are implemented:

[0043] Obtaining a cargo movement trajectory; the cargo movement trajectory is the movement trajectory of the cargo when the cargo is sorted by the cross belt;

[0044] Determine characteristic points of the cargo movement trajectory; the characteristic points include an inflection point value of the cargo movement trajectory and an intersection value between the cargo movement trajectory and a cargo sorting port;

[0045] By comparing the feature points with the floating item threshold, it is determined whether there is any abnormality in the sorting of the goods by the cross belt; the floating item threshold is a threshold obtained by obtaining multiple goods movement trajectory samples, determining the sample feature points of each of the goods movement trajectory samples, and performing outlier detection on the multiple sample feature points.

[0046] A computer-readable storage medium stores a computer program, which, when executed by a processor, implements the following steps:

[0047] Obtaining a cargo movement trajectory; the cargo movement trajectory is the movement trajectory of the cargo when the cargo is sorted by the cross belt;

[0048] Determine characteristic points of the cargo movement trajectory; the characteristic points include an inflection point value of the cargo movement trajectory and an intersection value between the cargo movement trajectory and a cargo sorting port;

[0049] By comparing the feature points with the floating item threshold, it is determined whether there is any abnormality in the sorting of the goods by the cross belt; the floating item threshold is a threshold obtained by obtaining multiple goods movement trajectory samples, determining the sample feature points of each of the goods movement trajectory samples, and performing outlier detection on the multiple sample feature points.

[0050] The above-mentioned cross-belt sorting anomaly detection method, device, computer equipment and storage medium obtain the cargo movement trajectory and determine the characteristic points of the cargo movement trajectory. The characteristic points can be used to describe the cargo movement trajectory. By comparing the characteristic points with the floating item threshold, it is determined whether there is any abnormality in the sorting of the cargo in the cross-belt. The abnormal characteristic points can be identified, and then the abnormal cargo movement trajectory can be detected. The cargo with abnormal sorting can be determined and corrected manually, thereby improving the accuracy of cargo sorting. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] Figure 1 A schematic diagram of a flow chart of a method for detecting anomalies in cross-belt sorting according to an embodiment;

[0052] Figure 2 Schematic diagram of cargo movement trajectory in one embodiment;

[0053] Figure 3 1 is a structural block diagram of a cross-belt sorting anomaly detection device according to an embodiment;

[0054] Figure 4 Schematic diagram of the process of outlier detection step in one embodiment;

[0055] Figure 5 1 is a structural block diagram of a cross-belt sorting anomaly detection device according to an embodiment;

[0056] Figure 6 FIG. 1 is a diagram showing the internal structure of a computer device in one embodiment. DETAILED DESCRIPTION

[0057] In order to make the purpose, technical solutions and advantages of this application more clearly understood, the present application is further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain this application and are not intended to limit this application.

[0058] The cross-belt sorting anomaly detection method provided in this application can be applied to terminals or servers. The terminals include, but are not limited to, various personal computers, laptops, smartphones, tablets, and portable wearable devices, and the servers can be implemented as independent servers or server clusters consisting of multiple servers. This cross-belt sorting anomaly detection method can be applied to automated sorting in express logistics, including automated sorting within logistics transfer yards, as well as automated sorting scenarios such as airport baggage check-in and large-scale industrial operations.

[0059] In one embodiment, Figure 1 As shown, a cross-belt sorting anomaly detection method is provided, which is described by taking the application of the method to a server as an example, and includes the following steps:

[0060] Step S110, obtaining a movement trajectory of the goods; the movement trajectory of the goods is the movement trajectory of the goods when the goods are sorted by the cross belt.

[0061] In the specific implementation, when the goods are sorted by the cross-belt sorting equipment, the goods are transported to the designated sorting position through the cross-belt and then fall into the sorting port. The entire movement trajectory of the goods during the sorting process can be captured by the video acquisition equipment to form a goods movement trajectory and send it to the server.

[0062] In practical applications, multiple industrial cameras can be placed in the direction of cross-belt operation. Each industrial camera is in a real-time trigger state. Multiple industrial cameras can synchronously capture video streams. By capturing video streams during the cross-belt sorting process, the cargo movement trajectory can be generated based on the captured video streams. For example, the cargo movement trajectory can be obtained by superimposing multiple captured video streams.

[0063] Step S120, determining characteristic points of the cargo movement trajectory; the characteristic points include the inflection point value of the cargo movement trajectory and the intersection value between the cargo movement trajectory and the cargo sorting port.

[0064] In the specific implementation, after obtaining the cargo movement trajectory, the server can mark the cargo movement trajectory as a set Q of m points (q1, q2, q3, ..., q m )(m≥3), where q1 is the starting point of the cargo trajectory, which can be the center point of the cargo when it is first detected, q m The end point of the cargo trajectory can be the center point of the cargo when it was last detected. According to the set Q, the intersection value and the inflection point value of the feature points of the cargo trajectory can be determined.

[0065] The intersection value can be the horizontal coordinate of the current moving point of the goods when the goods fall into the sorting port, and the current moving point can be the center point of the goods detected when the goods fall into the sorting port. The equation of the straight line of the cross belt can be set as y = kx + t, and the function analytical expression f (x) = kx + ty can be determined according to the equation of the straight line of the cross belt. The i-th point q in the set Q is sequentially i (x i ,y i ) and the i+1th point q i+1 (x i+1 ,y i+1 ) into the function expression, we can get f(x i )=kx i +ty i and f(x i+1 )=kx i+1 +ty i+1 , if f(x i ) and f(x i+1 ) have opposite signs, i.e. (kx i +ty i )(kx i+1 +ty i+1 )<0, then x i As the intersection value a.

[0066] The inflection point value can be the horizontal coordinate of the point where the goods are running on the cross belt and rotating. After the rotation, the goods gradually approach and fall into the sorting port. The intersection of the goods movement trajectory and the cross belt can be set as q i (x i ,y i ), let k≥0, q i-k (x i-k ,y i-k ) is the first k points of the intersection, q i-k-1 (x i-k-1 ,y i-k-1 ) is the k+1 point before the intersection, q i-k-2 (x i-k-2 ,y i-k-2 ) is the first k+2 points of the intersection, and the vertical coordinate distance threshold T is set. If |y i-k -y i-k-2 |≤T, then x i-k-1 As the inflection point value b, for example, T = 5 can be taken. If |y i-k -y i-k-2 |≤5, then x i-k-1 is the inflection point value b.

[0067] After determining the intersection value and the inflection point value, the intersection value and the inflection point value can be combined into a coordinate point p(a,b), which is placed in a two-dimensional rectangular coordinate system at a position with abscissa a and ordinate b as the characteristic point of the cargo movement trajectory, and is used to characterize the characteristics of the cargo movement trajectory.

[0068] In step S130, the feature points are compared with the floating item threshold to determine whether there is any abnormality in the sorting of goods by the cross belt. The floating item threshold is obtained by obtaining multiple samples of the movement trajectory of goods, determining the sample feature points of each sample of the movement trajectory of goods, and performing outlier detection on the multiple sample feature points.

[0069] In a specific implementation, the floating threshold may include an intersection threshold T a and T a ′, and the inflection point threshold T b and T b ′, after determining the characteristic point p(a, b) of the cargo trajectory, the characteristic point can be compared with the floating threshold. If the intersection value a is greater than the intersection threshold T a , and the inflection point value b is greater than the inflection point threshold T b , or the intersection value a is less than the intersection threshold T a ', and the inflection point value b is less than the inflection point threshold T b ', it can be determined that there is an abnormality in the sorting of goods by the cross belt, and the goods are sorted incorrectly. If the intersection value a is greater than the intersection threshold T a , and the inflection point value b is greater than the inflection point threshold T b , it can be determined that the goods are delayed in falling. It may be that the goods are delayed in falling due to their light weight, and due to inertia, they fall to the sorting port behind the designated sorting port. If the intersection value a is less than the intersection threshold T a ', and the inflection point value b is less than the inflection point threshold T b ', it can be determined that the goods are floating items that have fallen in advance. This may be due to the shaking of the cross belt, which causes the goods to fall in advance and fall into the sorting port in front of the designated sorting port.

[0070] The floating item threshold can be obtained by obtaining multiple samples of cargo movement trajectories, determining the characteristic points of each cargo movement trajectory sample, obtaining multiple characteristic points, and performing outlier detection using the multiple characteristic points. The process of obtaining the floating item threshold can specifically include the following steps:

[0071] (1) Assume that m cargo movement trajectory samples are obtained, determine a feature point based on each cargo movement trajectory sample, and obtain m sample feature points. Place the m sample feature points in a two-dimensional rectangular coordinate system;

[0072] (2) Taking point p among the m sample feature points as an example, calculate the distances between point p and other sample feature points in sequence to obtain m-1 distance values;

[0073] (3) Assume that the initial value of counter s is 0 and the initial distance threshold is r0. Compare the m-1 distance values with r0 in sequence. If the distance value is less than r0, then the counter s is incremented by 1. Otherwise, if the distance value is not less than r0, then the counter s remains unchanged.

[0074] (4) Set the initial score threshold as r1, calculate the ratio of s to m-1, if s / (m-1) is less than r1, then point p is determined to be an outlier; otherwise, if s / (m-1) is not less than r1, then point p is determined to be a cluster point;

[0075] (5) According to the method of steps (2) to (4) above, the other m-1 sample feature points in the coordinate system are judged in turn as outliers or cluster points, and the initial outlier point set Q1 and the initial dense point set Q2 are obtained, wherein the initial outlier point set Q1 is the set of all outliers in the m sample feature points, and the initial dense point set Q2 is the set of all cluster points in the m sample feature points;

[0076] (6) Find the center and radius r of the fitting circle of the initial dense point set Q1, update r0 to r, keep the score threshold r1 unchanged, repeat the above (2)-(5) operations to obtain the updated dense point set Q2;

[0077] (7) Find the point with the largest horizontal coordinate in the updated dense point set Q2, and use the horizontal coordinate corresponding to this point as the intersection threshold T a Similarly, find the point with the largest ordinate in the updated Q2, and use the ordinate corresponding to this point as the inflection point threshold T b , thus obtaining the floating threshold value (T a ,T b ); Find the point with the smallest horizontal coordinate in the updated dense point set Q2, and use the horizontal coordinate of the point as the intersection threshold T a ′, similarly, find the point with the smallest ordinate in the updated Q2, and use the ordinate corresponding to this point as the inflection point threshold T b ′, and thus obtain the floating threshold value of the early falling piece (T a ′,T b ′).

[0078] When it is determined that there is an abnormality in the sorting of goods by the cross-belt, an alarm signal can be sent out through the cross-belt sorting equipment to prompt the administrator to correct the incorrectly sorted goods.

[0079] The above-mentioned cross-belt sorting anomaly detection method obtains the cargo movement trajectory and determines the characteristic points of the cargo movement trajectory. The characteristic points can be used to describe the cargo movement trajectory. By comparing the characteristic points with the floating item threshold, it is determined whether there is any abnormality in the cross-belt sorting of the cargo. The abnormal characteristic points can be identified, and then the abnormal cargo movement trajectory can be detected. The cargo with abnormal sorting can be determined and corrected manually, which can improve the accuracy of cargo sorting.

[0080] Figure 2 A schematic diagram of the cargo movement trajectory is provided. Figure 2 Cargo D is transported on a crossbelt. According to the cargo sorting plan, the crossbelt rotates cargo D at a designated location to allow it to fall into the sorting slot. This rotation point is known as inflection point B. After rotating at inflection point B, cargo D falls into the target slot specified in the cargo sorting plan, slot C. A trajectory forms from the moment cargo D begins its transport on the crossbelt to the moment it lands in slot C. This trajectory intersects the sorting slot entrance at point A.

[0081] Figure 3 A structural diagram of a cross-belt sorting anomaly detection device is provided. Figure 3 The target tracking module can be used to obtain the movement trajectories of a large number of goods; the trajectory preprocessing module can be used to obtain the intersection point of the goods and the edge of the cross belt, the turning point of the goods on the cross belt, and the corresponding compartment number of the trajectory end point in each trajectory; the trajectory analysis module can be used to find the threshold corresponding to the abnormal movement trajectory of each sorting compartment; the trajectory judgment module can be used to obtain the behavior of the express delivery. If the trajectory does not meet the threshold conditions, it is considered an abnormal trajectory. Based on the above modules, cross-belt sorting anomaly detection can specifically include the following steps:

[0082] 1. Obtain the movement trajectory of a large number of goods through the target tracking module. The starting point of the trajectory is the center point of the box where the goods are detected for the first time, and the end point of the trajectory is the center point of the box where the goods are detected for the last time. The trajectory of the goods is recorded as the point set Q (q1, q2, q3, ..., q m )(m≥3).

[0083] 2. Obtain the slot number through the trajectory preprocessing module, and save the cargo trajectory of the corresponding slot according to the slot number. Taking slot C and cargo D as an example, find the intersection point A of the trajectory of cargo D falling into C and the cross belt, and obtain the horizontal coordinate of the intersection point A, that is, the value a. Specifically, let the equation of the cross belt line be y=kx+t, and sequentially convert the i-th point q in the trajectory into i (x i ,y i ) and the i+1th point q i+1 (x i+1 ,y i+1) into the equation of the line, if (kx i +ty i )(kx i+1 +ty i+1 )<0, then x i is the value a;

[0084] According to the characteristics of the cargo sorting plan that will cause the cargo to rotate at a specific time point, find the horizontal coordinate of the inflection point B of the cargo D that falls into the grid C, that is, the value b. Specifically, let the intersection of the trajectory and the cross belt be q i (x i ,y i ), q i-k (x i-k ,y i-k ) is the first k (k ≥ 0) points of the intersection, q i-k-1 (x i-k-1 ,y i-k-1 ) is point q i-k The previous point, q i-k-2 (x i-k-2 ,y i-k-2 ) is point q i-k-1 The previous point of , if |y i-k -y i-k-2 |≤5, then x i-k-1 is the value b;

[0085] Combine the values a and b into a coordinate point p(a,b), and place point p in a two-dimensional rectangular coordinate system. Point p is the trajectory feature of the cargo D. The trajectory of the cargo in the cross-belt sorting system is as follows: Figure 2 shown.

[0086] 3. After the system has been in trial operation for a period of time, the number of goods falling into the grid C gradually increases. It can be observed that the trajectories of normal goods and floating goods are different. That is, the trajectories of normal goods are within a certain area, while the trajectories of floating goods are clearly distinguished from normal goods. The trajectory analysis module is used to obtain the trajectory characteristics of all the trajectories of goods falling into the grid C. In the rectangular coordinate system, the feature points are mainly concentrated in a certain area, and there are sporadic outliers outside the area. These outliers are the feature points obtained from the trajectories of floating goods. Through the trajectory analysis module, the initial outlier point set Q1 and the initial dense point set Q2 can be obtained based on outlier detection. Figure 4 A flowchart of the outlier detection process is provided. The outlier detection process may specifically include the following steps:

[0087] (1) Assuming that there are m feature points in the coordinate system, take point p as an example, calculate the distance between point p and non-point p in sequence, and obtain m-1 distance values;

[0088] (2) Assume that the initial value of counter s is 0 and the initial distance threshold is r0. Compare the distance value with r0 in turn. If the distance value is less than r0, then the counter s is incremented by 1. Otherwise, the counter s remains unchanged.

[0089] (3) Set the initial score threshold as r1, calculate the ratio of s to m-1, if s / (m-1) is less than r1, then point p is an outlier, otherwise point p is a cluster point;

[0090] (4) By successively judging whether the other m-1 feature points in the coordinate system are outliers, the initial outlier point set Q1 and the initial dense point set Q2 can be obtained.

[0091] After determining the initial outlier point set Q1 and the initial dense point set Q2, the outlier point set and the dense point set can be adaptively updated. Specifically, the center and radius r of the fitting circle of the initial dense point set Q1 are obtained, r0 is updated to r, the score threshold remains unchanged, and the above outlier detection operation is repeated to find the point with the largest horizontal coordinate in the updated dense point set Q2 and obtain its horizontal coordinate x. Similarly, the point with the largest vertical coordinate in the updated Q2 is obtained and obtain its vertical coordinate y. Here, x and y are the floating item threshold T for delayed drop of cargo corresponding to the slot C. a 、T b ; You can also find the point with the smallest horizontal coordinate in the updated dense point set Q2 and get its horizontal coordinate x′. Similarly, find the point with the smallest vertical coordinate in the updated Q2 and get its vertical coordinate y′. Here, x′ and y′ are the floating item thresholds T for the cargo to fall early at the slot C. a ′、T b ′.

[0092] After obtaining the floating threshold T a 、T a ′、T b 、T b ', for the goods that have just fallen into the grid C, the trajectory judgment module is used. When the horizontal coordinate of the intersection point of the trajectory of the goods and the cross belt is greater than the threshold T a And the horizontal coordinate of the inflection point is greater than T b , or the horizontal coordinate of the intersection is less than the threshold T a ′ and the horizontal coordinate of the inflection point is less than T b ', it can be determined that the goods are sorted incorrectly.

[0093] In one embodiment, the cargo movement trajectory includes multiple trajectory points, the multiple trajectory points include a first trajectory point and a second trajectory point, and the first trajectory point and the second trajectory point are adjacent; the above-mentioned step S120 may specifically include: obtaining the function expression of the cross belt; substituting the first trajectory point into the function expression to obtain the first function value, and substituting the second trajectory point into the function expression to obtain the second function value; if the first function value and the second function value have opposite signs, then obtaining the intersection value according to the horizontal coordinate of the first trajectory point.

[0094] In the specific implementation, the cargo movement trajectory may include m (m≥3) trajectory points q1, q2, q3, ..., q m , where q1 is the starting point of the cargo trajectory, which can be the center point of the cargo when it is first detected, and q m The end point of the cargo trajectory can be the center point of the cargo when it was last detected. The first trajectory point can be the i-th point q among the m trajectory points. i (x i ,y i ), the second trajectory point can be the i+1th point q among the m trajectory points i+1 (x i+1 ,y i+1 ). The equation of the cross-band line can be set to y = kx + t, and the function expression f(x) = kx + ty can be obtained according to the equation of the cross-band line. i (x i ,y i ) and the second trajectory point q i+1 (x i+1 ,y i+1 ) into the function expression, and we get the first function value f(x i )=kx i +ty i and the second function value f(x i+1 )=kx i+1 +ty i+1 , if f(x i ) and f(x i+1 ) have opposite signs, i.e. (kx i +ty i )(kx i+1 +ty i+1 )<0, then the horizontal coordinate x of the first trajectory point can be i As the intersection value a.

[0095] In this embodiment, by obtaining the function expression of the cross belt, substituting the first trajectory point into the function expression to obtain the first function value, and substituting the second trajectory point into the function expression to obtain the second function value, if the first function value and the second function value have opposite signs, then the intersection value is obtained according to the horizontal coordinate of the first trajectory point. The intersection value in the characteristic point of the cargo movement trajectory can be efficiently obtained, thereby improving the efficiency of cross-belt sorting anomaly detection.

[0096] In one embodiment, the cargo movement trajectory further includes a third trajectory point, a fourth trajectory point, and a fifth trajectory point, and the third trajectory point, the fourth trajectory point, and the fifth trajectory point are adjacent to each other and located before the intersection point. The above-mentioned step S120 may further specifically include: respectively obtaining the coordinates of the third trajectory point, the fourth trajectory point, and the fifth trajectory point, where the coordinates include a horizontal coordinate and a vertical coordinate; obtaining a vertical coordinate distance based on the vertical coordinate of the third trajectory point and the vertical coordinate of the fifth trajectory point; if the vertical coordinate distance does not exceed a preset vertical coordinate distance threshold, obtaining an inflection point value based on the horizontal coordinate of the fourth trajectory point.

[0097] In the specific implementation, the cargo movement trajectory may include m (m≥3) trajectory points q1, q2, q3, ..., q m , the third track point, the fourth track point and the fifth track point can be respectively q i-k (x i-k ,y i-k ),q i-k-1 (x i-k-1 ,y i-k-1 ) and q i-k-2 (x i-k-2 ,y i-k-2 ), where k≥0, q i-k (x i-k ,y i-k ) is the first k points of the intersection, q i-k-1 (x i-k-1 ,y i-k-1 ) is the k+1 point before the intersection, q i-k-2 (x i-k-2 ,y i-k-2 ) is the k+2 points before the intersection. Get the ordinates of the third, fourth and fifth track points, and get the third ordinate y i-k , the fourth vertical coordinate y i-k-1 and the fifth vertical coordinate y i-k-2 , if |y i-k -y i-k-2 |≤T, where T is the preset vertical coordinate distance threshold, then x i-k-1 as the inflection point value b.

[0098] In this embodiment, by respectively obtaining the coordinates of the third trajectory point, the fourth trajectory point, and the fifth trajectory point, the ordinate distance is obtained according to the ordinate of the third trajectory point and the ordinate of the fifth trajectory point. If the ordinate distance does not exceed the preset ordinate distance threshold, the inflection point value is obtained according to the abscissa of the fourth trajectory point. This can efficiently obtain the inflection point value in the characteristic point of the cargo movement trajectory, thereby improving the efficiency of cross-belt sorting anomaly detection.

[0099] In one embodiment, the above cross-belt sorting anomaly detection method may further include the following steps:

[0100] Step S101: Detect whether each sample feature point is an outlier based on a preset point distance threshold, and obtain an initial outlier point set and an initial dense point set;

[0101] Step S102, performing circle fitting on the initial dense point set to obtain the radius of the fitting circle;

[0102] Step S103, updating the point distance threshold according to the radius of the fitting circle to obtain an updated point distance threshold;

[0103] Step S104: Detect whether each sample feature point is an outlier based on the updated point distance threshold, and obtain an updated dense point set;

[0104] Step S105: Obtain a floating item threshold based on the updated dense point set.

[0105] In a specific implementation, before obtaining the cargo movement trajectory, multiple cargo movement trajectory samples can be obtained first, and the sample feature points of each cargo movement trajectory sample can be determined to obtain multiple sample feature points. Outlier detection can be performed on the multiple sample feature points to obtain a floating threshold, wherein the sample feature points are composed of inflection point values and intersection point values. The process of obtaining the inflection point value and the intersection point value is the same as the process of obtaining the inflection point value and the intersection point value in the cargo movement trajectory feature point in the aforementioned embodiment, and will not be repeated here. Performing outlier detection on multiple sample feature points can specifically include: selecting one of the multiple sample feature points as the point to be detected, and the other sample feature points as non-detection points, calculating the point distance between each non-detection point and the point to be detected, and taking the non-detection point whose point distance is less than the preset point distance threshold as a close distance point, and counting the proportion of close distance points in all non-detection points. If the proportion is less than the preset proportion threshold, the point to be detected is determined to be an outlier. Otherwise, if the proportion is not less than the proportion threshold, the point to be detected is determined to be a cluster point. All sample feature points are detected in sequence by the above-mentioned outlier detection method, and all the obtained outliers are used as the initial outlier set, and all the obtained cluster points are used as the initial dense point set. The initial dense point set can then be fitted into a circle to obtain the center and radius of the fitted circle. The radius of the fitted circle is used as the new point distance threshold, and the above-mentioned outlier detection process is executed to obtain a new dense point set. The point with the largest horizontal coordinate in the new dense point set is selected. The horizontal coordinate of this point can be used as the intersection threshold of the delayed falling part. The point with the largest vertical coordinate in the new dense point set is selected. The vertical coordinate of this point can be used as the inflection threshold of the delayed falling part. The intersection threshold and the inflection threshold of the delayed falling part can form the floating part threshold of the delayed falling part; the point with the smallest horizontal coordinate in the new dense point set is selected. The horizontal coordinate of this point can be used as the intersection threshold of the early falling part. The point with the smallest vertical coordinate in the new dense point set is selected. The vertical coordinate of this point can be used as the inflection threshold of the early falling part. The intersection threshold and the inflection threshold of the early falling part can form the floating part threshold of the early falling part.

[0106] In practical applications, assuming that there are m feature points in the coordinate system, take point p as an example, calculate the distance between point p and non-point p in turn, and obtain m-1 distance values; set the initial value of counter s to 0, the initial distance threshold to r0, and compare the distance value with r0 in turn. If the distance value is less than r0, then the counter s is increased by 1, otherwise, the counter s remains unchanged; set the initial score threshold to r1, calculate the ratio of s to m-1, if s / (m-1) is less than r1, then point p is an outlier, otherwise point p is a cluster point; determine whether the other m-1 feature points in the coordinate system are outliers in turn, and obtain the initial outlier point set Q1 and the initial dense point set Q2. After determining the initial outlier point set Q1 and the initial dense point set Q2, the outlier point set and the dense point set can be adaptively updated. Specifically, the center and radius r of the fitting circle of the initial dense point set Q1 are obtained, r0 is updated to r, the score threshold remains unchanged, and the above outlier detection operation is repeated to find the point with the largest horizontal coordinate in the updated dense point set Q2 and obtain its horizontal coordinate x. Similarly, the point with the largest vertical coordinate in the updated Q2 is obtained and obtain its vertical coordinate y. Here, x and y are the floating item threshold T for delayed drop of cargo corresponding to the slot C. a ,T b ; You can also find the point with the smallest horizontal coordinate in the updated dense point set Q2 and get its horizontal coordinate x′. Similarly, find the point with the smallest vertical coordinate in the updated Q2 and get its vertical coordinate y′. Here, x′ and y′ are the floating item thresholds T for the cargo to fall early at the slot C. a ′、T b ′.

[0107] In this embodiment, each sample feature point is tested for outlier status based on a preset point distance threshold to obtain an initial outlier point set and an initial dense point set. A circle is then fitted to the initial dense point set to obtain a fitted circle radius. The point distance threshold is then updated based on the fitted circle radius to obtain an updated point distance threshold. Each sample feature point is tested for outlier status based on the updated point distance threshold to obtain an updated dense point set. A floating item threshold is then determined based on the updated dense point set. This allows for automatic acquisition of the floating item threshold, which can then be used to perform sorting anomaly detection, improving detection efficiency. Furthermore, the floating item threshold can be adaptively determined based on cargo trajectory samples, resulting in high accuracy and, consequently, improved accuracy in sorting anomaly detection.

[0108] In one embodiment, the above-mentioned step S101 may specifically include: selecting a point to be detected from multiple sample feature points; counting the point distance between the non-detection point and the point to be detected; the non-detection point is all points other than the point to be detected in the multiple sample feature points; if the point distance is less than the point distance threshold, the non-detection point is determined to be a close-distance point; counting the proportion of close-distance points in the non-detection point; if the proportion is less than a preset proportion threshold, the point to be detected is determined to be an outlier; all outliers in the multiple sample feature points constitute an initial outlier point set; if the proportion is not less than the proportion threshold, the point to be detected is determined to be a cluster point; all cluster points in the multiple sample feature points constitute an initial dense point set.

[0109] In a specific implementation, outlier detection for multiple sample feature points can specifically include: selecting one of the multiple sample feature points as the point to be detected, and the other sample feature points as non-detection points, calculating the point distance between each non-detection point and the point to be detected, and taking the non-detection points whose point distance is less than a preset point distance threshold as close points, and counting the proportion of close points in all non-detection points. If the proportion is less than a preset proportion threshold, the point to be detected is determined to be an outlier. Otherwise, if the proportion is not less than the proportion threshold, the point to be detected is determined to be a cluster point. All sample feature points are detected in sequence using the above-mentioned outlier detection method, and all the obtained outliers are used as the initial outlier set, and all the obtained cluster points are used as the initial dense point set.

[0110] In this embodiment, by selecting a point to be detected from multiple sample feature points, counting the point distance between the non-detection point and the point to be detected, if the point distance is less than the point distance threshold, the non-detection point is determined to be a close-distance point, and counting the proportion of close-distance points in the non-detection points. If the proportion is less than a preset proportion threshold, the point to be detected is determined to be an outlier. All outliers in the multiple sample feature points constitute an initial outlier point set. If the proportion is not less than the proportion threshold, the point to be detected is determined to be a cluster point. All cluster points in the multiple sample feature points constitute an initial dense point set. Multiple sample feature points can be automatically judged as outliers or cluster points, and the outlier point set and dense point set can be quickly obtained, thereby efficiently determining the floating item threshold and improving the efficiency of cross-belt sorting anomaly detection.

[0111] In one embodiment, the floating item threshold includes a first floating item threshold and a second floating item threshold; the above-mentioned step S105 may specifically include: obtaining the maximum value of the horizontal coordinate, the maximum value of the vertical coordinate, the minimum value of the horizontal coordinate and the minimum value of the vertical coordinate in the updated dense point set; obtaining a first intersection threshold according to the maximum value of the horizontal coordinate, obtaining a first inflection point threshold according to the maximum value of the vertical coordinate, obtaining a second intersection threshold according to the minimum value of the horizontal coordinate, and obtaining a second inflection point threshold according to the minimum value of the vertical coordinate; obtaining a first floating item threshold according to the first intersection threshold and the first inflection point threshold, and obtaining a second floating item threshold according to the second intersection threshold and the second inflection point threshold.

[0112] The first intersection threshold, the first inflection point threshold, and the first floating component threshold may be the intersection threshold, the inflection point threshold, and the floating component threshold for delayed falling components, respectively.

[0113] The second intersection threshold, the second inflection point threshold, and the second floating component threshold may be the intersection threshold, the inflection point threshold, and the floating component threshold for prematurely falling components, respectively.

[0114] In the specific implementation, after obtaining the new dense point set, the point with the largest horizontal coordinate in the new dense point set can be selected. The horizontal coordinate of this point is the maximum horizontal coordinate in the updated dense point set. The maximum horizontal coordinate can be used as the first intersection threshold. The point with the largest vertical coordinate in the new dense point set can be selected. The vertical coordinate of this point is the maximum vertical coordinate in the updated dense point set. The maximum vertical coordinate can be used as the first inflection point threshold. The first intersection threshold and the first inflection point threshold can constitute the first floating threshold; the point with the smallest horizontal coordinate in the new dense point set can be selected. The horizontal coordinate of this point is the minimum horizontal coordinate in the updated dense point set. The minimum horizontal coordinate can be used as the second intersection threshold. The point with the smallest vertical coordinate in the new dense point set can be selected. The vertical coordinate of this point is the minimum vertical coordinate in the updated dense point set. The minimum vertical coordinate can be used as the second inflection threshold. The second intersection threshold and the second inflection point threshold can constitute the second floating threshold.

[0115] In this embodiment, by obtaining the maximum value of the horizontal coordinate, the maximum value of the vertical coordinate, the minimum value of the horizontal coordinate and the minimum value of the vertical coordinate in the updated dense point set, the first intersection threshold is obtained according to the maximum value of the horizontal coordinate, the first inflection point threshold is obtained according to the maximum value of the vertical coordinate, the second intersection threshold is obtained according to the minimum value of the horizontal coordinate, the second inflection point threshold is obtained according to the minimum value of the vertical coordinate, the first floating item threshold is obtained according to the first intersection threshold and the first inflection point threshold, and the second floating item threshold is obtained according to the second intersection threshold and the second inflection point threshold. The floating item threshold can be efficiently determined, thereby improving the efficiency of cross-belt sorting anomaly detection.

[0116] In one embodiment, the above step S130 may specifically include: if the intersection value is greater than a first intersection threshold value, and the inflection point value is greater than the first inflection point threshold value, then determining that the cargo is a floating item that falls in late; if the intersection value is less than a second intersection threshold value, and the inflection point value is less than the second inflection point threshold value, then determining that the cargo is a floating item that falls in early.

[0117] In the specific implementation, after determining the characteristic point p(a, b) of the cargo trajectory, the characteristic point can be compared with the floating item threshold. If the intersection value a is greater than the first intersection threshold T a , and the inflection point value b is greater than the first inflection point threshold T b , or the intersection value a is less than the second intersection threshold T a ', and the inflection point value b is less than the second inflection point threshold T b', it can be determined that there is an abnormality in the sorting of goods by the cross belt, and the goods are sorted incorrectly. Among them, if the intersection value a is greater than the first intersection threshold T a , and the inflection point value b is greater than the first inflection point threshold T b , it can be determined that the goods are delayed in falling. It may be that the goods are delayed in falling due to their light weight, and due to inertia, they fall to the sorting port behind the designated sorting port. If the intersection value a is less than the second intersection threshold T a ', and the inflection point value b is less than the second inflection point threshold T b ', it can be determined that the goods are floating items that have fallen in advance. This may be due to the shaking of the cross belt, which causes the goods to fall in advance and fall into the sorting port in front of the designated sorting port.

[0118] In this embodiment, if the intersection value is greater than the first intersection threshold and the inflection point value is greater than the first inflection point threshold, the goods are determined to be floating items that fall in late; if the intersection value is less than the second intersection threshold and the inflection point value is less than the second inflection point threshold, the goods are determined to be floating items that fall in advance. This can efficiently and accurately determine the goods with sorting abnormalities and judge the type of sorting abnormalities, making it easier to correct them manually and put the goods with sorting abnormalities into the correct sorting port, thereby improving the accuracy of sorting.

[0119] It should be understood that although Figure 1 and Figure 4 The steps in the flowchart are shown in sequence as indicated by the arrows, but these steps are not necessarily executed in the order indicated by the arrows. Unless otherwise specified in this document, there is no strict order restriction for the execution of these steps, and these steps can be executed in other orders. In addition, Figure 1 and Figure 4 At least part of the steps may include multiple steps or multiple stages. These steps or stages are not necessarily performed at the same time, but can be performed at different times. The order of execution of these steps or stages is not necessarily one by one, but can be performed in turn or alternately with other steps or at least part of the steps or stages in other steps.

[0120] In one embodiment, Figure 5 As shown, a cross-belt sorting anomaly detection device is provided, comprising: an acquisition module 510, a feature point determination module 520 and a judgment module 530, wherein:

[0121] An acquisition module 510 is configured to acquire a movement trajectory of goods; the movement trajectory of goods is the movement trajectory of the goods when the goods are sorted by the cross belt;

[0122] A feature point determination module 520 is configured to determine feature points of the cargo movement trajectory; the feature points include an inflection point value of the cargo movement trajectory and an intersection value between the cargo movement trajectory and the cargo sorting port;

[0123] The judgment module 530 is configured to determine whether there is any abnormality in the sorting of the goods by the cross belt by comparing the feature points with a floating item threshold. The floating item threshold is a threshold value obtained by obtaining multiple goods movement trajectory samples, determining the sample feature points of each of the goods movement trajectory samples, and performing outlier detection on the multiple sample feature points.

[0124] In one embodiment, the feature point determination module 520 is further used to obtain the function expression of the intersection band; substitute the first trajectory point into the function expression to obtain a first function value, and substitute the second trajectory point into the function expression to obtain a second function value; if the first function value and the second function value have opposite signs, then the intersection value is obtained according to the horizontal coordinate of the first trajectory point.

[0125] In one embodiment, the feature point determination module 520 is further configured to obtain the coordinates of the third trajectory point, the fourth trajectory point, and the fifth trajectory point, respectively, where the coordinates include a horizontal coordinate and a vertical coordinate; obtain a vertical coordinate distance based on the vertical coordinate of the third trajectory point and the vertical coordinate of the fifth trajectory point; and obtain the inflection point value based on the horizontal coordinate of the fourth trajectory point if the vertical coordinate distance does not exceed a preset vertical coordinate distance threshold.

[0126] In one embodiment, the cross-belt sorting anomaly detection device further includes:

[0127] A detection module is used to detect whether each of the sample feature points is an outlier based on a preset point distance threshold, and obtain an initial outlier point set and an initial dense point set;

[0128] A circle fitting module, configured to perform circle fitting on the initial dense point set to obtain a fitting circle radius;

[0129] A threshold updating module, configured to update the point distance threshold according to the radius of the fitting circle to obtain an updated point distance threshold;

[0130] A point set updating module, configured to detect whether each of the sample feature points is an outlier based on the updated point distance threshold, and obtain an updated dense point set;

[0131] The floating item threshold determination module is used to obtain the floating item threshold according to the updated dense point set.

[0132] In one embodiment, the above-mentioned detection module is also used to select a point to be detected from the multiple sample feature points; count the point distance between the non-detection point and the point to be detected; the non-detection point is all points other than the point to be detected in the multiple sample feature points; if the point distance is less than the point distance threshold, the non-detection point is determined to be a close-distance point; count the proportion of the close-distance points in the non-detection points; if the proportion is less than a preset proportion threshold, the point to be detected is determined to be an outlier; all outliers in the multiple sample feature points constitute the initial outlier point set; if the proportion is not less than the proportion threshold, the point to be detected is determined to be a cluster point; all cluster points in the multiple sample feature points constitute the initial dense point set.

[0133] In one embodiment, the above-mentioned floating item threshold determination module is also used to obtain the maximum horizontal coordinate, maximum vertical coordinate, minimum horizontal coordinate and minimum vertical coordinate in the updated dense point set; obtain the first intersection threshold according to the maximum horizontal coordinate, obtain the first inflection point threshold according to the maximum vertical coordinate, obtain the second intersection threshold according to the minimum horizontal coordinate, and obtain the second inflection point threshold according to the minimum vertical coordinate; obtain the first floating item threshold according to the first intersection threshold and the first inflection point threshold, and obtain the second floating item threshold according to the second intersection threshold and the second inflection point threshold.

[0134] In one embodiment, the judgment module 530 is further configured to determine that the cargo is a floating item that has been delayed if the intersection value is greater than the first intersection threshold and the inflection point value is greater than the first inflection point threshold; and to determine that the cargo is a floating item that has been dropped in advance if the intersection value is less than the second intersection threshold and the inflection point value is less than the second inflection point threshold.

[0135] The specific limitations of the cross-belt sorting anomaly detection device can be found in the aforementioned limitations of the cross-belt sorting anomaly detection method and will not be further elaborated here. Each module within the aforementioned cross-belt sorting anomaly detection device can be implemented in whole or in part via software, hardware, or a combination thereof. Each of these modules can be embedded in or independent of a processor within a computer device in hardware form, or stored in a computer device memory in software form, allowing the processor to call and execute the corresponding operations of each module.

[0136] In one embodiment, a computer device is provided. The computer device may be a server, and its internal structure diagram may be as follows: Figure 6As shown. The computer device includes a processor, a memory, and a network interface connected via a system bus. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program, and a database. The internal memory provides an environment for the operation of the operating system and computer program in the non-volatile storage medium. The database of the computer device is used to store cross-belt sorting anomaly detection data. The network interface of the computer device is used to communicate with an external terminal via a network connection. When the computer program is executed by the processor, a cross-belt sorting anomaly detection method is implemented.

[0137] Those skilled in the art will understand that Figure 6 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than shown in the figure, or combine certain components, or have a different component arrangement.

[0138] In one embodiment, a computer device is provided, including a memory and a processor. The memory stores a computer program. When executed by the processor, the computer program causes the processor to perform the steps of the above-described method for detecting anomalies in cross-belt sorting. The steps of the above-described method for detecting anomalies in cross-belt sorting may be the steps of the above-described methods for detecting anomalies in cross-belt sorting.

[0139] In one embodiment, a computer-readable storage medium is provided, storing a computer program. When executed by a processor, the computer program causes the processor to perform the steps of the aforementioned method for detecting anomalies in cross-belt sorting. The steps of the method for detecting anomalies in cross-belt sorting may be the steps of the methods for detecting anomalies in cross-belt sorting described in the various embodiments described above.

[0140] Those skilled in the art will appreciate that all or part of the processes in the above-mentioned embodiment methods can be implemented by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media used in the embodiments provided in this application may include at least one of non-volatile and volatile memory. Non-volatile memory may include read-only memory (ROM), magnetic tape, floppy disk, flash memory or optical memory, etc. Volatile memory may include random access memory (RAM) or external cache memory. As an illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM).

[0141] The technical features of the above embodiments can be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0142] The above-described embodiments merely represent several implementation methods of the present application. While the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the present invention. It should be noted that a person skilled in the art could make various modifications and improvements without departing from the spirit of the present application, all of which fall within the scope of protection of the present application. Therefore, the scope of protection of the present patent application shall be determined by the appended claims.

Claims

1. A cross-belt sorting anomaly detection method, characterized in that: The method comprises: Obtaining a cargo movement trajectory; the cargo movement trajectory is the movement trajectory of the cargo when the cargo is sorted by the cross belt; Determine characteristic points of the cargo movement trajectory; the characteristic points include an inflection point value of the cargo movement trajectory and an intersection value between the cargo movement trajectory and a cargo sorting port; By comparing the feature points with the floating item threshold, it is determined whether there is any abnormality in the sorting of the goods by the cross belt; the floating item threshold is a threshold obtained by obtaining multiple goods movement trajectory samples, determining the sample feature points of each of the goods movement trajectory samples, and performing outlier detection on the multiple sample feature points.

2. The method according to claim 1, characterized in that The cargo movement trajectory includes a plurality of trajectory points, the plurality of trajectory points include a first trajectory point and a second trajectory point, and the first trajectory point and the second trajectory point are adjacent to each other; and the feature points for determining the cargo movement trajectory include: Obtaining a function analytical expression of the cross band; Substituting the first trajectory point into the function expression to obtain a first function value, and substituting the second trajectory point into the function expression to obtain a second function value; If the first function value and the second function value have opposite signs, the intersection value is obtained according to the horizontal coordinate of the first trajectory point.

3. The method according to claim 2, characterized in that The cargo movement trajectory further includes a third trajectory point, a fourth trajectory point and a fifth trajectory point, wherein the third trajectory point, the fourth trajectory point and the fifth trajectory point are adjacent to each other and are located before the intersection point; The determining of the characteristic points of the cargo movement trajectory further includes: respectively acquiring coordinates of the third track point, the fourth track point, and the fifth track point, wherein the coordinates include a horizontal coordinate and a vertical coordinate; Obtaining a longitudinal coordinate distance according to the longitudinal coordinate of the third track point and the longitudinal coordinate of the fifth track point; If the ordinate distance does not exceed a preset ordinate distance threshold, the inflection point value is obtained according to the abscissa of the fourth trajectory point.

4. The method according to claim 1, wherein The method further comprises: According to a preset point distance threshold, each of the sample feature points is detected to determine whether it is an outlier point, and an initial outlier point set and an initial dense point set are obtained; Performing circle fitting on the initial dense point set to obtain a fitting circle radius; Updating the point distance threshold according to the fitting circle radius to obtain an updated point distance threshold; According to the updated point distance threshold, detecting whether each of the sample feature points is an outlier, and obtaining an updated dense point set; The floating point threshold is obtained according to the updated dense point set.

5. The method according to claim 4, characterized in that The method of detecting whether each of the sample feature points is an outlier based on a preset point distance threshold, and obtaining an initial outlier point set and an initial dense point set, includes: Select a point to be detected from multiple sample feature points; Counting the distances between non-detection points and the points to be detected; the non-detection points are all points other than the points to be detected in the multiple sample feature points; If the point distance is less than the point distance threshold, the non-detection point is determined to be a close-distance point; Counting the proportion of the close-distance points in the non-detection points; If the ratio is less than a preset ratio threshold, the point to be detected is determined to be an outlier; all outliers in the plurality of sample feature points constitute the initial outlier set; If the ratio is not less than the ratio threshold, the point to be detected is determined to be a cluster point; all cluster points in the multiple sample feature points constitute the initial dense point set.

6. The method according to claim 4, characterized in that The floating threshold includes a first floating threshold and a second floating threshold; obtaining the floating threshold according to the updated dense point set includes: Obtaining the maximum value of the horizontal coordinate, the maximum value of the vertical coordinate, the minimum value of the horizontal coordinate, and the minimum value of the vertical coordinate in the updated dense point set; A first intersection threshold is obtained according to the maximum value of the horizontal coordinate, a first inflection point threshold is obtained according to the maximum value of the vertical coordinate, a second intersection threshold is obtained according to the minimum value of the horizontal coordinate, and a second inflection point threshold is obtained according to the minimum value of the vertical coordinate; The first floating threshold is obtained according to the first intersection threshold and the first inflection point threshold, and the second floating threshold is obtained according to the second intersection threshold and the second inflection point threshold.

7. The method according to claim 6, characterized in that The step of comparing the feature point with a floating item threshold to determine whether there is an abnormality in the sorting of the goods by the cross belt includes: If the intersection value is greater than the first intersection threshold, and the inflection point value is greater than the first inflection point threshold, the cargo is determined to be a delayed floating item; If the intersection value is smaller than the second intersection threshold value, and the inflection point value is smaller than the second inflection point threshold value, it is determined that the cargo is a floating item that has fallen in advance.

8. A cross-belt sorting anomaly detection device, characterized in that: The device comprises: An acquisition module is used to acquire a movement trajectory of the goods; the movement trajectory of the goods is the movement trajectory of the goods when the goods are sorted by the cross belt; A feature point determination module is used to determine the feature points of the cargo movement trajectory; the feature points include the inflection point value of the cargo movement trajectory and the intersection value between the cargo movement trajectory and the cargo sorting port; A judgment module is configured to determine whether there is any abnormality in the sorting of the goods by the cross belt by comparing the feature points with a floating item threshold; the floating item threshold is a threshold value obtained by obtaining multiple goods movement trajectory samples, determining the sample feature points of each of the goods movement trajectory samples, and performing outlier detection on the multiple sample feature points.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 7 are implemented.

Images