Container safety intelligent lock penetration identification method and system based on computer vision

Abstract

The invention relates to a container safety intelligent lock penetrating identification method and system based on computer vision, and the method comprises the steps: detecting the information of two key points where a lock rod of an intelligent lock intersects with a lock hole, detecting a lock hole region of the intelligent lock, and determining a lock hole center region based on the lock hole region; and then processing is carried out according to the information of the two key points obtained through detection and the central area of the lock hole, and whether the lock rod of the intelligent lock penetrates through the lock hole or not is judged. Thus, according to the relevant information of the intelligent lock collected by a machine, whether the lock rod of the intelligent lock correctly penetrates through the lock hole or not is automatically and accurately identified, and the problems of low efficiency and tedious workload when whether the lock hook of the intelligent lock is in the locked state or not is manually checked are solved.

Description

Technical Field

[0001] This invention relates to the field of smart locks, and in particular to a method and system for identifying lock penetration in container security smart locks based on computer vision. Background Technology

[0002] In container logistics transportation, to ensure container transport safety and meet regulatory requirements, it is necessary to correctly install smart locks on the containers. During container transport, the smart locks must be kept in the locked state, meaning the lock cylinder is inserted into the keyhole, thus ensuring the container's security. Checking and confirming that the smart locks are locked is one of the key operations to ensure they function properly.

[0003] To check whether the smart locks on containers are locked, the current method is usually manual, where inspectors check whether the lock cylinder is inserted into the keyhole. If the lock cylinder is correctly inserted, the smart lock is locked; otherwise, it is unlocked. However, this manual method of checking the lock status is not only time-consuming and labor-intensive, and inefficient, but it also cannot preserve video evidence of the inspection process, making it difficult to trace the video evidence of the lock status check later.

[0004] Therefore, how to automatically and accurately identify whether the lock bar of a smart lock has correctly passed through the keyhole has become a technical problem that urgently needs to be solved in the field of smart lock security. Summary of the Invention

[0005] The first technical problem to be solved by the present invention is to provide a computer vision-based method for identifying whether the locking bar of a smart lock correctly passes through the keyhole, which is capable of automatically and accurately identifying whether the locking bar of the smart lock passes through the keyhole.

[0006] The second technical problem to be solved by the present invention is to provide a container security smart lock penetration identification system that implements the above-mentioned computer vision-based container security smart lock penetration identification method in light of the prior art.

[0007] The technical solution adopted by this invention to solve the first technical problem is: a method for identifying lock penetration using a smart container security lock based on computer vision, characterized by the following steps:

[0008] Detect information at two key points where the lock bar and keyhole of the smart lock intersect;

[0009] The keyhole area of ​​the smart lock is detected, and the center area of ​​the keyhole is determined based on this keyhole area;

[0010] Based on the two key pieces of information obtained from the detection and the central area of ​​the keyhole, the system can determine whether the lock rod of the smart lock has passed through the keyhole.

[0011] Improved, in the computer vision-based smart lock detection method for container security, the detection process of the two key point information includes the following steps a1 to a3:

[0012] Step a1: Acquire a smart lock image; wherein the smart lock image includes a lock bar image and a keyhole image;

[0013] Step a2: Extract the feature information of the acquired smart lock image to obtain the smart lock feature information;

[0014] Step a3: Based on the extracted smart lock feature information, infer the information of two key points where the lock bar intersects with the keyhole and the location information of the keyhole area; wherein, the two key point information includes the coordinate information of the first lock bar key point and the coordinate information of the second lock bar key point. The first lock bar key point is a position point located on the lock bar and exactly outside the keyhole entrance, and the second lock bar key point is a position point located on the lock bar and exactly outside the keyhole exit; the coordinate information of the first lock bar key point is marked as P1(x1,y1), and the coordinate information of the second lock bar key point is marked as P2(x2,y2).

[0015] Furthermore, in the computer vision-based smart lock penetration identification method for containers, the process of determining the central region of the keyhole includes the following steps b1 to b3:

[0016] Step b1: Based on the detected keyhole area, obtain the bounding box information of the keyhole area; wherein, the bounding box information of the keyhole area is denoted as C(x C1 ,y C1 ,x C2 ,y C2 );(x C1 ,y C1 (x) represents the coordinates of the top-left vertex of the rectangle corresponding to the keyhole area's bounding box. C2 ,y C2 (x) represents the coordinates of the lower right vertex of the rectangle corresponding to the keyhole area's bounding box; C1 =x1,y C1 =y1,x C2 =x2,y C2 =y2;

[0017] Step b2: Based on the obtained keyhole region boundary box information, obtain the coordinate information of the keyhole center point; wherein, the coordinate information of the keyhole center point is marked as C0(x C0 ,y C0 ):

[0018]

[0019] Step b3: Based on the obtained coordinates of the keyhole center point and the preset central region scaling factor, obtain the bounding box information of the keyhole center region; wherein, the bounding box information of the keyhole center region is marked as C'(x C'1 ,y C'1 ,x C'2 ,y C'2 ):

[0020]

[0021] Where α is the preset central region scaling factor, (x C'1 ,y C'1 ,x C'2 ,y C'2 ) represents the coordinate information of the four vertices on the boundary frame of the central region of the keyhole.

[0022] Further improvements are made to the computer vision-based container security smart lock penetration identification method. The process of determining whether the smart lock's locking bar has passed through the keyhole, based on the detected information from two key points and the central area of ​​the keyhole, includes the following steps c1 to c6:

[0023] Step c1: Based on the information of the two key points where the lock bar intersects with the keyhole and the boundary box information of the keyhole area, obtain the first perforation determination result of whether the lock bar passes through the keyhole;

[0024] Step c2: Based on the information of the two key points where the lock bar intersects with the lock hole, the parametric equation of the lock bar line segment is obtained; and based on the information of the bounding box of the lock hole area, the parametric equation of the four bounding box edges that form the bounding box of the lock hole area is obtained.

[0025] Step c3: Based on the obtained parametric equations of the lock bar line segment and the parametric equations of the four boundary frame edges of the keyhole area, determine whether the lock bar line segment intersects with the boundary frame of the keyhole area to determine whether the lock bar passes through the boundary frame of the keyhole area.

[0026] Step c4: Process the data based on whether the lock bar passes through the boundary box of the keyhole area.

[0027] If the locking bar passes through the boundary frame of the keyhole area, proceed to step c5; otherwise, determine that the locking bar has not passed through the keyhole.

[0028] Step c5: Based on the obtained parametric equations of the lock bar line segment and the parametric equations of the boundary frame of the central region of the lock hole, determine whether the lock bar line segment intersects with the boundary frame of the central region of the lock hole, and obtain the second perforation determination result of whether the lock bar passes through the lock hole.

[0029] Step c6: Make a judgment based on the obtained first and second perforation determination results:

[0030] If both the first and second perforation determination results indicate that the locking bar has passed through the keyhole, then the locking bar is finally determined to have passed through the keyhole; otherwise, a secondary determination operation is performed to determine whether the locking bar has passed through the keyhole.

[0031] Furthermore, in the computer vision-based container security smart lock penetration identification method, during the secondary judgment operation, when the first penetration judgment result is that the lock rod has passed through the lock hole and the second penetration judgment result is that the lock rod has not passed through the lock hole, it is determined that the lock rod is suspected of passing through the lock hole, so as to perform a human judgment confirmation operation on whether the lock rod has passed through the lock hole; when the first penetration judgment result is that the lock rod has not passed through the lock hole, it is determined that the lock rod ultimately has not passed through the lock hole.

[0032] Furthermore, in the computer vision-based smart lock for container security, the method for determining whether the lock bar passes through the keyhole in step c1 is as follows:

[0033] If both key points where the lock bar intersects the keyhole are within the bounding box of the keyhole area, then the lock bar is determined to have passed through the keyhole; otherwise, the lock bar is determined not to have passed through the keyhole. The condition for either of the two key points to be within the bounding box of the keyhole area is as follows: x C1 ≤x≤x C2 , and y C1 ≤y≤y C2 (x,y) represents the coordinates of either of the two key points.

[0034] Improvedly, in the computer vision-based smart lock penetration identification method for container security, in step c2:

[0035] The parametric equations of the locking rod line segments are marked as follows:

[0036] P(t)=P1+t·(P2-P1), 0≤t≤1;

[0037] Where t is any point on the locking bar segment; P1 is the first locking bar key point on the locking bar segment, and P2 is the second locking bar key point on the locking bar segment;

[0038] The parametric equations for the four bounding box edges that form the bounding box of the keyhole region are as follows:

[0039]

[0040] Wherein, E1(s) is the parametric equation of the upper boundary frame edge of the keyhole region, E2(s) is the parametric equation of the right boundary frame edge of the keyhole region, E3(s) is the parametric equation of the lower boundary frame edge of the keyhole region, and E4(s) is the parametric equation of the left boundary frame edge of the keyhole region.

[0041] Improvedly, in the computer vision-based smart lock penetration identification method for containers, in step c3, the condition for the lock bar segment to intersect with any boundary frame edge of the lock hole area is: there exist t and s such that P(t) = E i (s), determine that the locking bar has passed through the lock hole; otherwise, determine that the lock buckle has not passed through the lock hole; where t∈[0,1], s∈[0,1], i∈[1,2,3,4].

[0042] Furthermore, in the computer vision-based smart lock penetration identification method for container security, in step c5, the condition for the lock bar segment to intersect with any boundary frame edge of the lock hole center region is that there exist t and s such that P(t) = E O,j (s), determine if the lock bar has passed through the keyhole; otherwise, determine if the lock bar has not passed through the keyhole; where t∈[0,1], s∈[0,1], E O,j (s) represents the parametric equation of the bounding box of the keyhole center region, j∈[1,2,3,4]; E O,1 (s) represents the parametric equation of the boundary frame edge on the upper side of the boundary frame of the keyhole center region, E O,2 (s) represents the parametric equation of the right boundary frame edge of the keyhole center region boundary frame, E O,3 (s) is the parametric equation of the lower boundary frame edge of the keyhole center region boundary frame, E O,4 (s) is the parametric equation of the left boundary frame edge of the central region of the keyhole.

[0043] The technical solution adopted by this invention to solve the second technical problem is: a container security smart lock penetration identification system, which implements the aforementioned computer vision-based container security smart lock penetration identification method, characterized in that the container security smart lock penetration identification system includes:

[0044] An image acquisition device acquires an image of the smart lock; wherein the smart lock image includes an image of the lock bar and an image of the keyhole;

[0045] The image processing device detects two key points where the lock bar and the keyhole intersect, as well as the keyhole area, based on the smart lock image acquired by the image acquisition device, and determines the center area of ​​the keyhole based on the keyhole area.

[0046] The central processing unit processes the information from the two key points and the central area of ​​the keyhole to determine whether the lock rod of the smart lock has passed through the keyhole.

[0047] Improved in this invention, the container security smart lock penetration identification system further includes a memory, which is connected to the image acquisition device and the central processing unit respectively.

[0048] Compared with existing technologies, the advantages of this invention are as follows: In the computer vision-based container security smart lock penetration identification method of this invention, the two key points where the smart lock's locking rod intersects with the keyhole are detected, as well as the keyhole area, and the center area of ​​the keyhole is determined based on this area. Then, based on the detected two key points and the center area of ​​the keyhole, the method processes the data to determine whether the smart lock's locking rod has passed through the keyhole. Thus, by using the relevant information (key points and keyhole area) collected by the machine, automatic and accurate identification of whether the smart lock's locking rod has correctly passed through the keyhole is achieved, avoiding the low efficiency and tedious workload of manually checking whether the smart lock's hook is in the locked state. Attached Figure Description

[0049] Figure 1 This is a schematic diagram of the lock-penetration identification method for container security smart locks based on computer vision in an embodiment of the present invention;

[0050] Figure 2 This is a schematic diagram of the detection process for key point information;

[0051] Figure 3 A schematic diagram illustrating the process of determining the center region of the keyhole;

[0052] Figure 4 This is a schematic diagram illustrating the process of determining whether the lock bar of a smart lock passes through the keyhole. Detailed Implementation

[0053] The present invention will be further described in detail below with reference to the accompanying drawings and embodiments.

[0054] This embodiment provides a computer vision-based method for identifying lock penetration on smart container security locks. See also... Figures 1-4 As shown, the computer vision-based smart lock penetration identification method for container security in this embodiment includes the following steps 1 to 3:

[0055] Step 1: Detect information at two key points where the lock bar and keyhole of the smart lock intersect; in this embodiment, see... Figure 2 As shown, the detection process for these two key points includes the following steps a1 to a3:

[0056] Step a1: Acquire a smart lock image; wherein, the smart lock image includes a lock bar image and a keyhole image; in this embodiment, a camera device is used to acquire the smart lock image.

[0057] Step a2: Extract the feature information of the acquired smart lock image to obtain the smart lock feature information;

[0058] Step a3: Based on the extracted smart lock feature information, the key point recognition model is trained to infer the information of two key points where the lock bar intersects with the keyhole; and the keyhole area location information is obtained by training the target detection model.

[0059] Step a3: Based on the extracted smart lock feature information, train a key point recognition model to infer the information of two key points where the lock bar intersects with the keyhole; train a target detection model to infer the position information of the keyhole area; wherein, the two key point information includes the coordinate information of the first lock bar key point and the coordinate information of the second lock bar key point. The first lock bar key point is a position point located on the lock bar and exactly outside the keyhole entrance, and the second lock bar key point is a position point located on the lock bar and exactly outside the keyhole exit; the coordinate information of the first lock bar key point is labeled as P1(x1,y1), and the coordinate information of the second lock bar key point is labeled as P2(x2,y2);

[0060] Step 2: Detect the keyhole area of ​​the smart lock and determine the keyhole center area based on this area; wherein, in this embodiment, see... Figure 3 As shown, the process of determining the center region of the keyhole includes the following steps b1 to b3:

[0061] Step b1: Based on the detected keyhole area, obtain the bounding box information of the keyhole area; wherein, the bounding box information of the keyhole area is denoted as C(x C1 ,y C1 ,x C2 ,y C2 );(x C1 ,y C1 (x) represents the coordinates of the top-left vertex of the rectangle corresponding to the keyhole area's bounding box. C2 ,y C2 (x) represents the coordinates of the lower right vertex of the rectangle corresponding to the keyhole area's bounding box; C1 =x1,y C1 =y1,x C2 =x2,y C2 =y2;

[0062] Step b2: Based on the obtained keyhole region boundary box information, obtain the coordinate information of the keyhole center point; wherein, the coordinate information of the keyhole center point is marked as C0(x C0 ,y C0 ):

[0063]

[0064] Step b3: Based on the obtained coordinates of the keyhole center point and the preset central region scaling factor, obtain the bounding box information of the keyhole center region; wherein, the bounding box information of the keyhole center region is marked as C'(x C'1 ,y C'1 ,x C'2 ,y C'2 ):

[0065]

[0066] Where α is the preset central region scaling factor, (x C'1 ,y C'1 ,x C'2 ,y C'2 ) represents the coordinate information of the four vertices on the boundary box of the central region of the keyhole; for example, in this embodiment, the preset central region scale coefficient α is set to 60%;

[0067] Step 3: Based on the detected key point information of the lock rod and the center area of ​​the keyhole, process the data to determine whether the lock rod of the smart lock has passed through the keyhole. For details, see... Figure 4 As shown, the process of determining whether the lock bar of the smart lock passes through the keyhole includes the following steps c1 to c6:

[0068] Step c1: Based on the information of the two key points where the lock bar intersects with the keyhole and the boundary box information of the keyhole area, obtain the first perforation determination result of whether the lock bar passes through the keyhole;

[0069] The method for determining whether the locking bar has passed through the keyhole is as follows: if both key points are located within the keyhole area bounding box, the locking bar is determined to have passed through the keyhole; otherwise, the locking bar is determined not to have passed through the keyhole. The condition for either key point to be located within the keyhole area bounding box is as follows: x C1 ≤x≤x C2 , and y C1 ≤y≤y C2 (x,y) represents the coordinates of either of the two key points.

[0070] Step c2 involves processing the information from the two key points where the lock bar intersects with the keyhole to obtain the parametric equation of the lock bar line segment; and obtaining the parametric equation of the four boundary frame edges forming the boundary frame of the keyhole region based on the obtained boundary frame information of the keyhole region; wherein, in this embodiment:

[0071] The parametric equation of the locking bar segment is denoted as: P(t)=P1+t·(P2-P1), 0≤t≤1; where t is any point on the locking bar segment; P1 is the first locking bar key point on the locking bar segment, and P2 is the second locking bar key point on the locking bar segment;

[0072] The parametric equations for the four bounding box edges that form the bounding box of the keyhole region are as follows:

[0073]

[0074] Where E1(s) is the parametric equation of the upper boundary frame edge of the keyhole region, E2(s) is the parametric equation of the right boundary frame edge of the keyhole region, E3(s) is the parametric equation of the lower boundary frame edge of the keyhole region, and E4(s) is the parametric equation of the left boundary frame edge of the keyhole region; θ is the change in the x-axis and y-axis components of the direction vector in the parametric equations.

[0075] Step c3: Based on the obtained parametric equations of the lock bar segment and the four boundary frame edges of the keyhole region boundary frame, a judgment is made as to whether the lock bar segment intersects with the keyhole region boundary frame, so as to determine whether the lock bar passes through the keyhole region boundary frame. Specifically, in this embodiment, the condition for the lock bar segment to intersect with any boundary frame edge of the keyhole region boundary frame is: there exist t and s such that P(t) = E i (s), determine that the locking bar has passed through the boundary box of the keyhole area; otherwise, determine that the latch has not passed through the boundary box of the keyhole area; where t∈[0,1], s∈[0,1], i∈[1,2,3,4];

[0076] Step c4: Process the data based on whether the lock bar passes through the boundary box of the keyhole area.

[0077] If the locking bar passes through the boundary frame of the keyhole area, proceed to step c5; otherwise, determine that the locking bar has not passed through the keyhole.

[0078] Step c5: Based on the obtained parametric equations of the lock bar segment and the boundary frame edges of the keyhole center region, a judgment is made as to whether the lock bar segment intersects with the boundary frame of the keyhole center region, thus obtaining the second perforation determination result of whether the lock bar passes through the keyhole; specifically, in this embodiment, the condition for the lock bar segment to intersect with any boundary frame edge of the keyhole center region boundary frame is: there exist t and s such that P(t) = E O,j (s), determine if the lock bar has passed through the keyhole; otherwise, determine if the lock bar has not passed through the keyhole; where t∈[0,1], s∈[0,1], E O,j (s) represents the parametric equation of the bounding box of the keyhole center region, j∈[1,2,3,4]; E O,1 (s) represents the parametric equation of the boundary frame edge on the upper side of the boundary frame of the keyhole center region, EO,2 (s) represents the parametric equation of the right boundary frame edge of the keyhole center region boundary frame, E O,3 (s) is the parametric equation of the lower boundary frame edge of the keyhole center region boundary frame, E O,4 (s) represents the parametric equation of the left boundary frame edge of the keyhole center region bounding box. Specifically, the parametric equations of the four boundary frame edges of the keyhole center region bounding box are as follows:

[0079]

[0080] Step c6: Make a judgment based on the obtained first and second perforation determination results:

[0081] If both the first and second penetration determination results indicate that the locking bar has passed through the keyhole, then the locking bar is ultimately determined to have passed through the keyhole; otherwise, a secondary determination operation is performed to check whether the locking bar has passed through the keyhole. In this secondary determination operation, if the first penetration determination result indicates that the locking bar has passed through the keyhole and the second penetration determination result indicates that the locking bar has not passed through the keyhole, then the locking bar is suspected of having passed through the keyhole, and a manual confirmation operation is performed to check whether the locking bar has passed through the keyhole; if the first penetration determination result indicates that the locking bar has not passed through the keyhole, then the locking bar is ultimately determined not to have passed through the keyhole.

[0082] This embodiment also provides a container security smart lock penetration identification system that implements the above-described computer vision-based container security smart lock penetration identification method. The container security smart lock penetration identification system includes:

[0083] An image acquisition device acquires an image of the smart lock; wherein the smart lock image includes an image of the lock bar and an image of the keyhole;

[0084] The image processing device detects two key points where the lock bar and the keyhole intersect, as well as the keyhole area, based on the smart lock image acquired by the image acquisition device, and determines the center area of ​​the keyhole based on the keyhole area.

[0085] The central processing unit processes the information from the two key points and the central area of ​​the keyhole to determine whether the lock rod of the smart lock has passed through the keyhole.

[0086] To meet the need for image storage, in this embodiment, the container security smart lock penetration recognition system also includes a memory, which is connected to the image acquisition device and the central processing unit respectively.

[0087] Although preferred embodiments of the present invention have been described in detail above, it should be clearly understood that various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A computer vision-based method for identifying a container security intelligent lock penetration, characterized in that, The method comprises the following steps: Detecting two key point information of the intersection of the lock rod of the smart lock and the lock hole; Detecting the lock hole region of the smart lock, and determining the lock hole center region based on the lock hole region; Processing the two key point information and the lock hole center region to determine whether the lock rod of the smart lock passes through the lock hole.

2. The computer vision based container security smart lock pass identification method according to claim 1, characterized in that, The detection process of the two key point information comprises the following steps a1-a3: Step a1: collecting a smart lock image of the smart lock; wherein the smart lock image comprises a lock rod image and a lock hole image; Step a2: extracting feature information of the collected smart lock image to obtain smart lock feature information; Step a3: inferring the two key point information of the intersection of the lock rod and the lock hole and the position information of the lock hole region based on the extracted smart lock feature information; wherein the two key point information comprises first lock rod key point coordinate information and second lock rod key point coordinate information, the first lock rod key point is a position point on the lock rod and just outside the entrance of the lock hole, and the second lock rod key point is a position point on the lock rod and just outside the exit of the lock hole; the first lock rod key point coordinate information is marked as P1(x1, y1), and the second lock rod key point coordinate information is marked as P2(x2, y2).

3. The computer vision based container security smart lock pass identification method according to claim 2, characterized in that, The determination process of the lock hole center region comprises the following steps b1-b3: Step b1, according to the detected lock hole region, obtaining the lock hole region bounding box information; wherein, the lock hole region bounding box information is marked as C(x C1 ,y C1 ,x C2 ,y C2 );(x C1 ,y C1 ) is the left upper vertex coordinate information of the rectangular frame corresponding to the lock hole region bounding box, (x C2 ,y C2 ) is the right lower vertex coordinate information of the rectangular frame corresponding to the lock hole region bounding box; x C1 =x1, y C1 =y1, x C2 =x2, y C2 =y2; Step b2, obtaining the lock hole center point coordinate information according to the obtained lock hole region boundary box information; wherein the lock hole center point coordinate information is marked as C0(x C0 ,y C0 ) Step b3, obtaining the bounding box information of the lock hole center region according to the obtained lock hole center point coordinate information and a preset center region proportion coefficient; wherein the bounding box information of the lock hole center region is marked as C'(x C'1 ,y C'1 ,x C'2 ,y C'2 ) wherein a is a preset center region proportion coefficient, (x C'1 ,y C'1 ,x C'2 ,y C'2 ) is coordinate information of the four vertices of the lock hole center region bounding box.

4. The computer vision-based container security intelligent lock pass identification method according to claim 3, characterized in that, The processing of the two key point information and the lock hole center region to determine whether the lock rod of the smart lock passes through the lock hole comprises the following steps c1-c6: Step c1: obtaining a first penetration determination result of whether the lock rod passes through the lock hole based on the two key point information of the intersection of the lock rod and the lock hole and the lock hole region boundary box information; Step c2: obtaining a lock rod line segment parameter equation based on the two key point information of the intersection of the lock rod and the lock hole, and obtaining a parameter equation of four boundary box edges forming the lock hole region boundary box based on the lock hole region boundary box information; Step c3: judging whether the lock rod line segment and the lock hole region boundary box intersect based on the lock rod line segment parameter equation and the parameter equation of the four boundary box edges of the lock hole region boundary box to obtain whether the lock rod passes through the lock hole region boundary box; Step c4: processing based on the judgment result of whether the lock rod passes through the lock hole region boundary box: When the lock rod passes through the lock hole region boundary box, go to step c5; otherwise, determine that the lock rod does not pass through the lock hole; Step c5: judging whether the lock rod line segment and the lock hole center region boundary box intersect based on the lock rod line segment parameter equation and the parameter equation of the four boundary box edges of the lock hole center region boundary box to obtain a second penetration determination result of whether the lock rod passes through the lock hole; Step c6: judging based on the first penetration determination result and the second penetration determination result: When the first penetration determination result and the second penetration determination result are both that the lock rod has passed through the lock hole, finally determine that the lock rod has passed through the lock hole; Otherwise, perform a secondary judgment operation for whether the lock rod passes through the lock hole.

5. The computer vision based container security smart lock pass identification method according to claim 4, characterized in that, In step c1, the judgment of whether the lock rod passes through the lock hole is as follows: When any one of the two key point information is located in the lock hole region bounding box, it is determined that the lock rod has passed through the lock hole; otherwise, it is determined that the lock rod has not passed through the lock hole; wherein the determination condition that any one of the two key point information is located in the lock hole region bounding box is as follows: x C1 ≤x≤x C2 , and y C1 ≤y≤y C2 ; (x, y) is the coordinate of any one of the two key point information.

6. The computer vision-based container security intelligent lock pass identification method according to claim 5, characterized in that, In step c2: The parameter equation of the lock rod line segment of the lock rod is marked as: P(t) = P1 + t·(P2-P1), 0≤t≤1; wherein t is any point on the lock rod line segment; P1 is a first lock rod key point on the lock rod line segment, and P2 is a second lock rod key point on the lock rod line segment; The parameter equations of the four edge boxes of the lock hole region boundary box are as follows: E1(s) is the parameter equation of the upper edge box of the lock hole region boundary box, E2(s) is the parameter equation of the right edge box of the lock hole region boundary box, E3(s) is the parameter equation of the lower edge box of the lock hole region boundary box, and E4(s) is the parameter equation of the left edge box of the lock hole region boundary box.

7. The computer vision based container security smart lock pass identification method according to claim 6, wherein, In step c3, the condition that the lock rod line segment intersects any of the boundary box edges in the lock hole region boundary box is that there exist t and s such that P(t) = E i (s), it is determined that the lock rod has passed through the lock hole; otherwise, it is determined that the lock catch has not passed through the lock hole; wherein t∈[0,1], s∈[0,1], i∈[1,2,3,4].

8. The computer vision-based container security intelligent lock pass identification method according to claim 7, characterized in that, In step c5, the condition that the lock rod line segment intersects any of the boundary frame edges of the lock hole center region boundary frame is that there exist t and s such that P(t) = E O,j (s), it is determined that the lock rod has passed through the lock hole; otherwise, it is determined that the lock rod has not passed through the lock hole; wherein t ∈ [0, 1], s ∈ [0, 1], E O,j (s) is the parametric equation of the lock hole center region boundary frame, j ∈ [1, 2, 3, 4]; E O,1 (s) is the parametric equation of the upper side boundary frame edge of the lock hole center region boundary frame, E O,2 (s) is the parametric equation of the right side boundary frame edge of the lock hole center region boundary frame, E O,3 (s) is the parametric equation of the lower side boundary frame edge of the lock hole center region boundary frame, E O,4 (s) is the parametric equation of the left side boundary frame edge of the lock hole center region boundary frame.

9. The container security intelligent lock penetration identification system, which realizes the container security intelligent lock penetration identification method based on computer vision as claimed in claim 1, characterized in that, The container safety intelligent lock penetration recognition system comprises: An image acquisition device acquires an intelligent lock image of the intelligent lock; wherein the intelligent lock image comprises a lock rod image and a lock hole image; An image processing device detects two key point information of the intersection of the lock rod and the lock hole of the intelligent lock and a lock hole region of the intelligent lock according to the intelligent lock image acquired by the image acquisition device, and determines a lock hole center region based on the lock hole region; A central processing unit processes the two key point information and the lock hole center region obtained by detection to determine whether the lock rod of the intelligent lock penetrates the lock hole.

10. The container security smart lock pass identification system of claim 9, wherein, Further comprising a memory connected to the image acquisition device and the central processing unit.

Images