A method and system for detecting deformation of a support

Abstract

The present invention provides a detection method and system for stent deformation, constructs a stent sheath detector model based on a deep neural network, uses sample data of a monitoring video image including a stent sheath to train the stent sheath detector model; uses the stent sheath detector model to obtain the stent point position, binarizes the monitoring video image to obtain the stent point binarized image, and the stent point position is the center point of the detection frame obtained by the stent sheath detector model; projects the binarized image of the stent point to obtain the reference stent point binarized image, periodically obtains the stent point binarized image of the monitoring video image, transforms the stent binarized image based on the reference stent point binarized image to obtain the aligned stent point binarized image; superimposes the aligned stent point binarized images of adjacent periods, obtains the displacement trajectory diagrams of the aligned stent point binarized images respectively, and obtains the stent deformation position and deformation degree. The method can quickly locate the deformation position and degree.

Description

Technical Field

[0001] The present invention relates to the technical field of image processing, and in particular to a method and system for detecting deformation of a bracket. Background Art

[0002] The urban integrated pipe gallery, also known as the "common trench", is a tunnel space built underground in the city, where various municipal pipelines such as water, electricity, gas, heat, and communications are organically and intensively laid in the same tunnel, and are centrally managed as municipal infrastructure. Various types of pipelines are classified and placed on the shelves in layers, and the weight of the shelves is supported by brackets. When the weight of the pipeline exceeds the support range of the bracket, or the bracket is aged, the bracket will deform, causing the pipeline to stretch and deform, which may easily cause cable heating or pipeline leakage, causing risks such as fire or explosion. In order to ensure the long-term stable and safe operation of the pipe gallery, it is necessary to detect the deformation of the pipeline and the bracket. At present, inspections are mainly carried out through manual inspections. Since deformation is a long and slow process, subtle changes are difficult to detect, which can easily lead to omissions and pose a major safety hazard.

[0003] In order to deal with the deformation of the tunnel structure, deformation monitors are usually used at present. For example, the differential pressure settlement and inclination comprehensive measuring instrument uses the pressure change between adjacent measuring points to measure the settlement, and the inclination monitoring uses the change of static gravity acceleration to measure the inclination. For example, the distributed optical fiber strain system uses optical fiber as a sensing sensitive element and transmission signal medium, and adopts advanced OTDR technology to detect the strain changes at different positions along the optical fiber to achieve truly distributed measurement. The monitor host and workstation are arranged in the management room, and the deformation detection optical fiber is installed along the top of the tunnel, close to the inner wall.

[0004] Deformation monitors are expensive and require the installation of a large number of accessories and facilities. They are complex to construct and deploy, and are generally only used for monitoring the structure of the pipe gallery itself. Since there are many layers and pipes inside the pipe gallery, it is not suitable to use deformation monitors. Therefore, a solution with simple deployment and reasonable price is needed. Summary of the invention

[0005] In order to solve the technical problems in the prior art that deformation detectors are high in cost, require the installation of a large number of accessories and facilities, have complex construction and deployment, have many layers of shelves inside the pipe gallery, and are not suitable for deformation detection, the present invention proposes a detection method and system for bracket deformation to solve the above technical problems.

[0006] According to one aspect of the present invention, a method for detecting deformation of a stent is provided, comprising:

[0007] S1: Construct a stent sheath detector model based on a deep neural network, and train the stent sheath detector model using sample data of surveillance video images including stent sheaths;

[0008] S2: using the stent sheath detector model to obtain the stent position point, binarizing the monitoring video image to obtain a stent point binarization image, wherein the stent position point is the center point of the detection frame obtained by the stent sheath detector model;

[0009] S3: Performing a projection transformation on the support point binary image to obtain a reference support point binary image, periodically obtaining a support point binary image of a monitoring video image, and transforming the support point binary image based on the reference support point binary image to obtain an aligned support point binary image;

[0010] S4: superimposing the binary images of the aligned support points of adjacent periods, respectively obtaining the displacement trajectory diagrams of the binary images of the aligned support points, and obtaining the deformation position and deformation degree of the support.

[0011] In some specific embodiments, step S2 specifically includes using a trained stent sheath detector model to detect the input monitoring video image, obtain a detection frame of the stent sheath, use the center point of the detection frame as the stent position point, binarize the monitoring video image, and the pixel value of the stent position point is 1, and the rest is 0, to obtain a stent point binarized image. The position of the stent point can be clearly displayed through the binarization process, which is convenient for deformation analysis of the stent.

[0012] In some specific embodiments, the acquisition of the binary image of the reference support point in step S3 specifically includes: acquiring four support position points forming a quadrilateral in the same spatial plane, scaling and translating the spatial plane to obtain a target mapping plane, the focal length of the target mapping plane is the same as the focal length of the monitoring device, and the target mapping plane is parallel to the camera plane, and the support point binary image is subjected to a plane perspective transformation to obtain the initial binary image of the reference support point after the transformation. By virtue of this step, the actual distance of the sheath support point and the transformed image distance are kept in an equiproportional relationship, thereby improving the image measurement accuracy of the support deformation.

[0013] In some specific embodiments, the step S3 of aligning the support point binary image specifically includes: taking the reference support point binary image as a reference, aligning and transforming all support points of the support point binary image with a first transformation parameter, and transforming them to corresponding points of the reference support point binary image to obtain a first aligned support point binary image. Through the alignment step, the alignment transformation parameter H is recalculated each time the detection is performed, so that most of the non-deformed support target points overlap with the reference support points, so as to accurately reflect the deviation position and deviation degree of the deformed support points.

[0014] In some specific embodiments, step S3 also includes calculating the distance between the reference support point binary image and all corresponding support points in the first aligned support point binary image, removing points whose deformation deviation distance is greater than a specified distance threshold, obtaining an undeformed or slightly deformed reference support point binary image and a corresponding detected support point binary image, and calculating a second transformation parameter to transform the support point binary image to obtain a second aligned support point binary image.

[0015] In some specific embodiments, the first transformation parameter and the second transformation parameter decompose the coefficient matrix composed of the bracket point coordinate pairs through a singular value decomposition algorithm, and the first 9 parameters of the matrix are taken as the optimal solution of the first transformation parameter and the second transformation parameter.

[0016] In some specific embodiments, the displacement trajectory diagram in step S4 is specifically a line connecting the target point of the current second alignment bracket point binary image and the corresponding target point in the second alignment bracket point binary image of the previous cycle, and is superimposed on the trajectory diagram of the previous cycle. Through the trajectory diagram, the deformation position and deformation degree of the bracket can be intuitively and quickly located.

[0017] In some specific embodiments, step S4 specifically includes: in response to the deformation amount of a certain bracket point being greater than the bracket point deformation threshold, the bracket at that location is deformed; in response to the row bracket deformation coefficient being greater than the row deformation threshold, the entire row of brackets is deformed, wherein the bracket point deformation amount is the number of connected adjacent points in the trajectory graph, and the bracket deformation coefficient is the ratio of the deformation amount of the entire row of brackets to the number of bracket points.

[0018] According to a second aspect of the present invention, a computer-readable storage medium is provided, on which one or more computer programs are stored. When the one or more computer programs are executed by a computer processor, any of the above methods is implemented.

[0019] According to a third aspect of the present invention, a system for detecting deformation of a stent is provided, the system comprising:

[0020] A stent sheath detector model unit: configured to construct a stent sheath detector model based on a deep neural network, and train the stent sheath detector model using sample data of surveillance video images including stent sheaths;

[0021] Binarization processing unit: configured to obtain the position of the stent point by using the stent sheath detector model, and to obtain a stent point binary image by binarizing the monitoring video image, wherein the position of the stent point is the center point of the detection frame obtained by the stent sheath detector model;

[0022] Image transformation and alignment unit: configured to project and transform the support point binary image to obtain the reference support point binary image, regularly obtain the support point binary image of the monitoring video image, and transform the support point binary image based on the reference support point binary image to obtain the aligned support point binary image;

[0023] Deformation monitoring unit: configured to superimpose binary images of aligned support points of adjacent periods, respectively obtain displacement trajectory diagrams of the binary images of aligned support points, and obtain the deformation position and deformation degree of the support.

[0024] The present invention proposes a method and system for detecting bracket deformation, which utilizes image processing technology to solve the bracket deformation problem. Through plane perspective transformation, the image is transformed into a coordinate system parallel to the camera plane as a reference bracket point image, so as to proportionally reflect the bracket deformation offset and improve the detection accuracy. During regular measurements, the reference image is aligned to solve the judgment deviation caused by camera vibration displacement, and the alignment transformation parameters are optimized and solved by the least squares method using a large amount of coordinate point data. The alignment is further optimized by removing the deviation points and aligning again, thereby improving the bracket deformation detection accuracy. By utilizing the video surveillance cameras installed in the tunnel and through the video image analysis and processing of the present invention, the safety hazards of bracket deformation can be discovered in time to avoid safety accidents. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] The accompanying drawings are included to provide a further understanding of the embodiments and are incorporated into and constitute a part of this specification. The accompanying drawings illustrate the embodiments and together with the description are used to explain the principles of the present invention. Other embodiments and many expected advantages of the embodiments will be readily appreciated as they become better understood by reference to the following detailed description. Other features, objects and advantages of the present application will become more apparent by reading the detailed description of the non-limiting embodiments made with reference to the following drawings:

[0026] Figure 1 is a flow chart of a method for detecting deformation of a stent according to an embodiment of the present application;

[0027] Figure 2 is a flow chart of a method for detecting stent deformation according to a specific embodiment of the present application;

[0028] Figure 3 is a schematic diagram of a binary image of a detection support point according to a specific embodiment of the present application;

[0029] Figure 4 is a schematic diagram of a transformed binarized image of bracket points in a specific embodiment of the present application;

[0030] Figure 5is a schematic diagram of a stent deformation experiment of a specific embodiment of the present application;

[0031] Figure 6 is a framework diagram of a system for detecting stent deformation according to an embodiment of the present application;

[0032] Figure 7 It is a structural diagram of a computer system suitable for implementing an electronic device of an embodiment of the present application. DETAILED DESCRIPTION

[0033] The present application will be further described in detail below in conjunction with the accompanying drawings and embodiments. It is to be understood that the specific embodiments described herein are only used to explain the relevant invention, rather than to limit the invention. It should also be noted that, for ease of description, only the parts related to the relevant invention are shown in the accompanying drawings.

[0034] It should be noted that, in the absence of conflict, the embodiments and features in the embodiments of the present application can be combined with each other. The present application will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.

[0035] According to a method for detecting deformation of a stent according to an embodiment of the present application, Figure 1 FIG. 2 shows a flow chart of a method for detecting deformation of a stent according to an embodiment of the present application. Figure 1 As shown, the method includes:

[0036] S101: Construct a stent sheath detector model based on a deep neural network, and use sample data of surveillance video images including stent sheaths to train the stent sheath detector model. There are stent sheaths at the ends of the corridor stents. Collect each camera video image, mark the stent sheaths in the image with rectangular boxes, and obtain stent sheath sample data.

[0037] S102: Use the stent sheath detector model to obtain the stent position point, binarize the monitoring video image to obtain the stent point binarization image, wherein the stent position point is the center point of the detection frame obtained by the stent sheath detector model. Use the trained stent sheath detector model to detect the input monitoring video image, obtain the stent sheath detection frame, use the center point of the detection frame as the stent position point, binarize the monitoring video image, the stent position point pixel value is 1, and the rest are 0, to obtain the stent point binarization image.

[0038] S103: Project the support point binary image to obtain a reference support point binary image, periodically obtain the support point binary image of the monitoring video image, and transform the support point binary image based on the reference support point binary image to obtain the aligned support point binary image.

[0039] In a specific embodiment, four bracket position points forming a quadrilateral in the same spatial plane are obtained, and the spatial plane is scaled and translated to obtain a target mapping plane, the focal length of the target mapping plane is the same as the focal length of the monitoring device, and the target mapping plane is parallel to the camera plane, and the bracket point binary image is subjected to a plane perspective transformation to obtain the initial reference bracket point binary image after the transformation. The transformation model is simplified to an affine transformation, so that the actual distance of the sheath bracket point and the transformed image distance maintain an equiproportional relationship, thereby improving the image measurement accuracy of the bracket deformation.

[0040] In a specific embodiment, with the reference support point binary image as a reference, all support points of the support point binary image are aligned and transformed with the first transformation parameters, and transformed to the corresponding points of the reference support point binary image to obtain the first aligned support point binary image. The distance between the reference support point binary image and all corresponding support points in the first aligned support point binary image is calculated, and the points whose deformation deviation distance is greater than the specified distance threshold are removed to obtain the undeformed or slightly deformed reference support point binary image and the corresponding detection support point binary image, and the second transformation parameters are calculated to transform the support point binary image to obtain the second aligned support point binary image. Among them, the first transformation parameters and the second transformation parameters can be decomposed into a coefficient matrix composed of support point coordinate pairs by a singular value decomposition algorithm, and the first 9 parameters of the matrix are taken as the optimal solution of the first transformation parameters and the second transformation parameters.

[0041] S104: Superimpose the binary images of the aligned support points of adjacent cycles, obtain the displacement trajectory diagrams of the binary images of the aligned support points, and obtain the deformation position and deformation degree of the support. The displacement trajectory diagram is specifically a line connecting the target point of the current second binary image of the aligned support points and the corresponding target point in the second binary image of the aligned support points of the previous cycle, and is superimposed on the trajectory diagram of the previous cycle.

[0042] In a specific embodiment, in response to the deformation amount of a certain bracket point being greater than the bracket point deformation threshold, the bracket at that location is deformed; in response to the row bracket deformation coefficient being greater than the row deformation threshold, the entire row of brackets is deformed, wherein the bracket point deformation amount is the number of connected adjacent points in the trajectory graph, and the bracket deformation coefficient is the ratio of the deformation amount of the entire row of brackets to the number of bracket points.

[0043] Continue to refer Figure 2 , Figure 2 A flow chart of a method for detecting deformation of a stent according to a specific embodiment of the present application is shown, and the method specifically comprises the following steps:

[0044] Step 201: Train the stent sheath detector. There are stent sheaths at the ends of the pipe gallery stents. Collect each camera video image, mark the stent sheaths in the image with rectangular boxes, and obtain stent sheath sample data. Construct a stent sheath detector model based on a deep neural network, and train the stent sheath detector with sample data.

[0045] Step 202: Detect the stent sheath target. Obtain the camera video image, use the trained sheath detector to detect the input image, obtain the detection frame of each stent sheath, and use the center point of the detection frame as the stent position point. Binarize the image, the stent position point pixel value is 1, and the other pixel values ​​are 0, to obtain the stent point binary image. The stent point binary image is as follows Figure 3 As shown in the schematic diagram.

[0046] Step 203: Initialize the reference support point image. Select four support points that appear in the image screen and are close to the camera and are in the same physical space plane to form a quadrilateral, namely, upper left W0, upper right W1, lower right W2, and lower left W3. Take W0 as the coordinate origin and measure the spatial plane coordinates W of these four support points. i (i=0,1,2,3). Get the 4 pixel points P corresponding to the binary image of the support point i (i=0,1,2,3).

[0047] In a specific embodiment, the physical space plane coordinate W i (i=0,1,2,3) is scaled and translated to obtain the target mapping plane coordinate point W′ i =sW i +t W , so that the focal length of the target imaging plane is equal to the focal length of the camera, and the transformed image is the camera front view, that is, the camera plane is parallel to the jacket point plane, and the transformation model is simplified to an affine transformation, so that the actual distance of the jacket bracket point and the transformed image distance maintain an equal proportional relationship, thereby improving the image measurement accuracy of the bracket deformation. Among them, the vertical distance d of the selected bracket point space is calculated W =W3-W0 and the vertical distance d between the pixel P =P3-P0, and then obtain the scaling factor t W is a fixed translation amount, and the above mapping can make the spatial coordinate points fall in the same quadrant.

[0048] In a specific embodiment, the planar perspective transformation satisfies aW′ i =MP i , where M is the homography matrix a is the scaling factor. The selected 4 coordinate pairs have a total of 8 equations that form a linear equation system. The Gaussian elimination method is used to solve all parameters of the homography matrix M.

[0049] In a specific embodiment, the perspective transformation is performed on the support point binary image with the transformation parameter M to obtain the initial reference support point binary image I0 after the transformation, and the transformation correspondence of all target points and the support point information of each row in I0 are recorded. The reference support point binary image finally obtained is referenced Figure 4 The transformed binarized image of the support points is shown in FIG.

[0050] Step 204: Regular detection. At a fixed time period, such as 1 day, video images are acquired to perform stent sheath target detection to obtain a binary image I of the detected stent points. src .

[0051] Step 205: Alignment transformation of support points. During long-term operation or camera vibration, the camera position may shift slightly. Using the initial perspective transformation parameter M will cause deviations in all support target points after the transformation, affecting the judgment of support deformation. This deviation is not a simple translation deviation, so it is necessary to recalculate the alignment transformation parameter H through the alignment step for each detection, so that most of the non-deformed support target points overlap with the reference support points, so as to accurately reflect the deviation position and degree of the deformed support points.

[0052] In a specific embodiment, the reference support point binary image I0 is used as a reference, and the binary image I src All support points Transform to the corresponding point of I0 Where H is the alignment transformation homography matrix By the property that the outer product of two identical vectors is 0: Expand to get Remove the last term and simplify to the following formula: The formula is converted into a matrix equation: Ah = 0, where A is a 2N × 9 coefficient matrix consisting of N bracket point coordinate pairs:

[0053]

[0054] h is the alignment transformation parameter H to be solved, where h = [h0 h1 h2 h3 h4 h5 h6 h7 h8] T ;

[0055] Use the least squares method to optimize the solution and minimize ||Ah||2: And st||h||2=1;

[0056] Decompose the matrix A using the singular value decomposition SVD method (formula A = UDV T ), take the first 9 parameters of vector V as the optimal solution of H: h* =V.col(9).

[0057] Finally, the alignment transformation is performed with the transformation parameter H to obtain the aligned support point binary image I H .

[0058] Step 206: Remove the deviation points and align again: Calculate I0 and I H The distances of all corresponding support points are removed, and the points whose deformation deviation distance is greater than the specified distance threshold are removed to obtain the undeformed or slightly deformed reference support point I′0 and the corresponding detection support point I′ src , use the same algorithm as in step 205 to calculate the exact transformation parameter H′, and use H′ to transform I src Transform to obtain the aligned support point binary image I n Preferably, the designated distance threshold is 5. It should be appreciated that the designated threshold can be set to other different values ​​according to actual application scenarios, which can also achieve the technical effects of the present invention.

[0059] Step 207: I n The image is superimposed on the previous cycle T n-1 Trajectory image, and I n Each target point, compared with the previous cycle I n-1 The corresponding points are connected by straight lines, and the pixels through which the straight lines pass are marked as 1, so that the displacement trajectory diagram T of each bracket point is obtained. n Through this trajectory diagram, the deformation position and degree of the bracket can be located intuitively and quickly.

[0060] In a specific embodiment, the support point in the I0 image is taken as the starting point to calculate the corresponding trajectory map T n The number of connected adjacent pixels with a value of 1 is recorded as the deformation amount, and the deformation amount of each bracket point is obtained. If the bracket point deformation amount is greater than the bracket point deformation threshold (such as a value of 5), it is judged as bracket deformation. In a scene, multiple bracket point position deformations may occur due to bracket deformation. The positional relationship between them and the trajectory diagram can be used to quickly determine whether they are deformed. For each row, the deformation amounts of all bracket points in a row are accumulated and divided by the number of bracket points in a row to obtain the row bracket deformation coefficient. If the row bracket deformation coefficient is greater than the row deformation threshold (such as a value of 3), it is judged that the entire row bracket is deformed. It should be recognized that the bracket point deformation threshold and the row deformation threshold can be set to other values ​​according to different scenarios to meet the application requirements of different scenarios. Figure 5 FIG. 2 shows a schematic diagram of an experiment for detecting the deformation of a stent using the above method. Figure 5 The displacement trajectory diagram of the bracket points on the top shows that the bracket is deformed due to the pressure from the first layer of weight on the top.

[0061] Continue to refer Figure 6 , Figure 6The framework diagram of a detection system for stent deformation according to an embodiment of the present application is shown. The system specifically includes a stent sheath detector model unit 601 , a binarization processing unit 602 , an image transformation and alignment unit 603 and a deformation monitoring unit 604 . The stent sheath detector model unit 601 is configured to construct a stent sheath detector model based on a deep neural network, and train the stent sheath detector model using sample data of monitoring video images including stent sheaths; the binarization processing unit 602 is configured to obtain the stent point position using the stent sheath detector model, and binarize the monitoring video image to obtain the stent point binarization image, wherein the stent point position is the center point of the detection frame obtained by the stent sheath detector model; the image transformation and alignment unit 603 is configured to project the stent point binarization image to obtain the reference stent point binarization image, regularly obtain the stent point binarization image of the monitoring video image, and transform the stent point binarization image based on the reference stent point binarization image to obtain the aligned stent point binarization image; the deformation monitoring unit 604 is configured to superimpose the aligned stent point binarization images of adjacent periods, respectively obtain the displacement trajectory diagrams of the aligned stent point binarization images, and obtain the deformation position and deformation degree of the stent.

[0062] Reference below Figure 7 , which shows a schematic diagram of the structure of a computer system 700 suitable for implementing an electronic device of an embodiment of the present application. Figure 7 The electronic device shown is merely an example and should not bring any limitation to the functions and scope of use of the embodiments of the present application.

[0063] like Figure 7 As shown, the computer system 700 includes a central processing unit (CPU) 701, which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 702 or a program loaded from a storage part 708 into a random access memory (RAM) 703. In the RAM 703, various programs and data required for the operation of the system 700 are also stored. The CPU 701, the ROM 702, and the RAM 703 are connected to each other via a bus 704. An input / output (I / O) interface 705 is also connected to the bus 704.

[0064] The following components are connected to the I / O interface 705: an input section 706 including a keyboard, a mouse, etc.; an output section 707 including a liquid crystal display (LCD), etc. and a speaker, etc.; a storage section 708 including a hard disk, etc.; and a communication section 709 including a network interface card such as a LAN card, a modem, etc. The communication section 709 performs communication processing via a network such as the Internet. A drive 710 is also connected to the I / O interface 705 as needed. A removable medium 711, such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, etc., is installed on the drive 710 as needed, so that a computer program read therefrom is installed into the storage section 708 as needed.

[0065] In particular, according to an embodiment of the present disclosure, the process described above with reference to the flowchart can be implemented as a computer software program. For example, an embodiment of the present disclosure includes a computer program product, which includes a computer program carried on a computer-readable storage medium, and the computer program includes a program code for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from the network through the communication part 709, and / or installed from the removable medium 711. When the computer program is executed by the central processing unit (CPU) 701, the above functions defined in the method of the present application are executed. It should be noted that the computer-readable storage medium of the present application can be a computer-readable signal medium or a computer-readable storage medium or any combination of the above two. The computer-readable storage medium can be, for example, - but not limited to - an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device or device, or any combination of the above. More specific examples of computer-readable storage media may include, but are not limited to, an electrical connection with one or more conductors, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In the present application, a computer-readable storage medium may be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, device, or device. In the present application, a computer-readable signal medium may include a data signal propagated in a baseband or as part of a carrier wave, in which a computer-readable program code is carried. Such propagated data signals may take a variety of forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the above. A computer-readable signal medium may also be any computer-readable storage medium other than a computer-readable storage medium, which may send, propagate, or transmit a program for use by or in conjunction with an instruction execution system, device, or device. The program code contained on the computer-readable storage medium may be transmitted using any appropriate medium, including but not limited to: wireless, wireline, optical cable, RF, etc., or any suitable combination of the foregoing.

[0066] Computer program code for performing the operations of the present application may be written in one or more programming languages ​​or a combination thereof, including object-oriented programming languages, such as Java, Smalltalk, C++, and conventional procedural programming languages, such as "C" or similar programming languages. The program code may be executed entirely on the user's computer, partially on the user's computer, as a separate software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving a remote computer, the remote computer may be connected to the user's computer through any type of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0067] The flow chart and block diagram in the accompanying drawings illustrate the possible architecture, function and operation of the system, method and computer program product according to various embodiments of the present application. In this regard, each square box in the flow chart or block diagram can represent a module, a program segment or a part of a code, and the module, the program segment or a part of the code contains one or more executable instructions for realizing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the square box can also occur in a sequence different from that marked in the accompanying drawings. For example, two square boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each square box in the block diagram and / or flow chart, and the combination of the square boxes in the block diagram and / or flow chart can be implemented with a dedicated hardware-based system that performs a specified function or operation, or can be implemented with a combination of dedicated hardware and computer instructions.

[0068] The modules involved in the embodiments of the present application may be implemented by software or by hardware.

[0069] As another aspect, the present application also provides a computer-readable storage medium, which may be included in the electronic device described in the above embodiment; or it may exist alone and not be assembled into the electronic device. The above computer-readable storage medium carries one or more programs, and when the above one or more programs are executed by the electronic device, the electronic device: constructs a stent sheath detector model based on a deep neural network, and trains the stent sheath detector model using sample data of a monitoring video image including a stent sheath; obtains a stent position point using the stent sheath detector model, and binarizes the monitoring video image to obtain a stent point binarized image, wherein the stent position point is the center point of the detection frame obtained by the stent sheath detector model; projects the stent point binarized image to obtain a reference stent point binarized image, periodically obtains a stent point binarized image of the monitoring video image, and transforms the stent point binarized image based on the reference stent point binarized image to obtain an aligned stent point binarized image; superimposes the aligned stent point binarized images of adjacent periods, and obtains displacement trajectory diagrams of the aligned stent point binarized images respectively, and obtains the deformation position and deformation degree of the stent.

[0070] The above description is only a preferred embodiment of the present application and an explanation of the technical principles used. Those skilled in the art should understand that the scope of the invention involved in the present application is not limited to the technical solution formed by a specific combination of the above technical features, but should also cover other technical solutions formed by any combination of the above technical features or their equivalent features without departing from the above invention concept. For example, the above features are replaced with the technical features with similar functions disclosed in this application (but not limited to) by each other to form a technical solution.

Claims

1. A method for detecting deformation of a stent, characterized in that: The bracket is used to support and place various pipeline racks in the urban integrated pipe gallery, including: S1: constructing a stent sheath detector model based on a deep neural network, and training the stent sheath detector model using sample data of a surveillance video image including the stent sheath; S2: using the stent sheath detector model to obtain a stent position point, binarizing the monitoring video image to obtain a stent point binarization image, wherein the stent position point is the center point of the detection frame obtained by the stent sheath detector model; S3: Projecting the support point binary image to obtain a reference support point binary image, periodically obtaining the support point binary image of the monitoring video image, and transforming the support point binary image based on the reference support point binary image to obtain an aligned support point binary image; S4: superimposing the binary images of the aligned support points of adjacent periods, respectively obtaining displacement trajectory diagrams of the binary images of the aligned support points, and obtaining the deformation position and deformation degree of the support.

2. The method for detecting stent deformation according to claim 1, characterized in that: The step S2 specifically includes using the trained stent sheath detector model to detect the input monitoring video image, obtaining the detection frame of the stent sheath, using the center point of the detection frame as the stent position point, binarizing the monitoring video image, the pixel value of the stent position point is 1, and the rest is 0, to obtain the binarized image of the stent point.

3. The method for detecting stent deformation according to claim 1, characterized in that: The acquisition of the binary image of the reference support point in step S3 specifically includes: acquiring four support position points that form a quadrilateral in the same spatial plane, scaling and translating the spatial plane to obtain a target mapping plane, wherein the focal length of the target mapping plane is the same as the focal length of the monitoring device, and the target mapping plane is parallel to the camera plane, and the support point binary image is subjected to a planar perspective transformation to obtain the initial reference support point binary image after the transformation.

4. The method for detecting stent deformation according to claim 1, characterized in that: The aligning of the support point binary image in step S3 specifically includes: taking the reference support point binary image as a reference, aligning and transforming all the support points of the support point binary image with a first transformation parameter, and transforming them to corresponding points of the reference support point binary image to obtain a first aligned support point binary image.

5. The method for detecting stent deformation according to claim 4, characterized in that: The step S3 also includes calculating the distance between the reference support point binary image and all corresponding support points in the first aligned support point binary image, removing points whose deformation deviation distance is greater than a specified distance threshold, obtaining an undeformed or slightly deformed reference support point binary image and a corresponding detected support point binary image, and calculating a second transformation parameter to transform the support point binary image to obtain a second aligned support point binary image.

6. The method for detecting stent deformation according to claim 5, characterized in that: The first transformation parameter and the second transformation parameter decompose the coefficient matrix composed of the bracket point coordinate pairs through the singular value decomposition algorithm, and take the first 9 parameters of the matrix as the optimal solution of the first transformation parameter and the second transformation parameter.

7. The method for detecting stent deformation according to claim 5, characterized in that: The displacement trajectory diagram in step S4 is specifically a line connecting the target point of the current second alignment support point binary image and the corresponding target point in the second alignment support point binary image of the previous cycle, and is superimposed on the trajectory diagram of the previous cycle.

8. The method for detecting stent deformation according to claim 1 or 7, characterized in that: The step S4 specifically includes: in response to the deformation amount of a certain support point being greater than the support point deformation threshold, the support at that location is deformed; in response to the row support deformation coefficient being greater than the row deformation threshold, the entire row of supports is deformed, wherein the support point deformation amount is the number of connected adjacent points of the trajectory diagram, and the support deformation coefficient is the ratio of the deformation amount of the entire row of supports to the number of support points.

9. A computer-readable storage medium having one or more computer programs stored thereon, characterized in that: When the one or more computer programs are executed by a computer processor, the method according to any one of claims 1 to 8 is implemented.

10. A system for detecting deformation of a stent, characterized in that: The bracket is used to support and place various pipeline racks in the urban integrated pipe gallery. The system includes: A stent sheath detector model unit: configured to construct a stent sheath detector model based on a deep neural network, and train the stent sheath detector model using sample data of a surveillance video image including the stent sheath; A binarization processing unit: configured to obtain a stent point position using the stent sheath detector model, and binarize the monitoring video image to obtain a stent point binarization image, wherein the stent point position is a center point of a detection frame obtained by the stent sheath detector model; Image transformation and alignment unit: configured to project and transform the support point binary image to obtain a reference support point binary image, regularly obtain the support point binary image of the monitoring video image, and transform the support point binary image based on the reference support point binary image to obtain an aligned support point binary image; The deformation monitoring unit is configured to superimpose the binary images of the aligned support points of adjacent periods, respectively obtain the displacement trajectory diagrams of the binary images of the aligned support points, and obtain the deformation position and deformation degree of the support.

Images