A deformation monitoring control method and system for a non-measuring camera

Abstract

The present invention relates to a deformation monitoring control method and system for a non-metric camera, including: taking a first initial image of a target structure by a first non-metric camera, and simultaneously taking a second initial image of the target structure by a second non-metric camera; taking a first real-time image of the target structure in real time by the first non-metric camera, and simultaneously taking a second real-time image of the target structure in real time by the second non-metric camera; an image processing device performing image recognition on the first initial image and the second initial image to determine the initial three-dimensional coordinates of the feature points of the target structure; the image processing device performing image recognition on the first real-time image and the second real-time image to determine the real-time three-dimensional coordinates of the feature points of the target structure; the image processing device calculating the deformation amount of the target structure according to the initial three-dimensional coordinates and the real-time three-dimensional coordinates of the feature points of the target structure. The present invention can realize the monitoring of the overall deformation of the monitored structure.

Description

Technical Field

[0001] The present invention relates to the technical field of object deformation monitoring, and particularly relates to a deformation monitoring control method and system for a non-metric camera. Background Art

[0002] Traditional deformation monitoring mainly relies on manual operation of theodolites, levels, tape measures, etc. to periodically measure deformed objects (such as cracks, landslides, structures, etc.). Taking landslide displacement measurement as an example, the traditional landslide monitoring technology uses a theodolite for observation, and the time interval is usually once every 10 days to half a month; however, during the rainy season or when the landslide deformation is relatively severe, the observation frequency must be increased or even continuous uninterrupted observation is required. The work intensity is high and the data real-time performance is poor. For some dangerous or inaccessible areas, people often hope to complete the measurement task by remotely controlling the instrument and transmitting back the data. Thus, remote real-time performance and automation have become an important part of contemporary deformation monitoring.

[0003] Although GPS technology can automatically complete monitoring, recording, and calculation, it is necessary to deploy GPS receivers or antennas at each point, with a high cost, and it is not suitable for monitoring large bridge structures with a large number of monitoring points. Moreover, when the sampling frequency is high, the monitoring accuracy is low and cannot meet the accuracy requirements of deformation monitoring. Although the total station can achieve automated monitoring, the monitoring period is long and it cannot monitor dynamic deformation. Sensor measurement can monitor in real time, but it needs to be in direct contact with the structure body and can only monitor the local deformation of the structure body. Although three-dimensional laser scanning technology can monitor the overall deformation of large structures, the scanning period is long and it cannot monitor the dynamic deformation of the structure. Laser interferometry and deflection meter methods need to deploy prisms and optical marks at the target points respectively and cannot monitor multiple points simultaneously.

[0004] In summary, some of the currently adopted monitoring means have a low degree of automation and cannot monitor dynamic deformation; some have a high cost and are difficult to be popularized and applied in engineering; some can only monitor local deformation and cannot monitor the overall deformation of the structure body. Therefore, it is very difficult for the current monitoring means to achieve the monitoring of the overall deformation of the monitored structure within a limited economic cost and to give an early warning of the safety of the monitored structure. Summary of the Invention

[0005] The object of the present invention is to provide a deformation monitoring control method for a non-metric camera to achieve the monitoring of the overall deformation of the monitored structure within a limited economic cost.

[0006] To achieve the above object, an embodiment of the present invention provides a deformation monitoring control method for a non-metric camera, which is implemented based on a deformation monitoring control system of a non-metric camera. The system includes a first non-metric camera, a second non-metric camera, and an image processing device. The method includes the following steps:

[0007] The first non-measuring camera captures the target structure to obtain a first initial image, and simultaneously, the second non-measuring camera captures the target structure to obtain a second initial image;

[0008] The first non-measuring camera captures the target structure in real time to obtain a first real-time image, and simultaneously, the second non-measuring camera captures the target structure in real time to obtain a second real-time image;

[0009] The image processing device performs image recognition on the first initial image and the second initial image to determine the initial three-dimensional coordinates of the feature points of the target structure;

[0010] The image processing device performs image recognition on the first real-time image and the second real-time image to determine the real-time three-dimensional coordinates of the feature points of the target structure;

[0011] The image processing device calculates the deformation amount of the target structure according to the initial three-dimensional coordinates and the real-time three-dimensional coordinates of the feature points of the target structure.

[0012] Preferably, the method further includes:

[0013] The image processing device matches the feature points in the first initial image and the first real-time image, and matches the feature points in the second initial image and the second real-time image;

[0014] According to the feature point matching result, determine any one same feature point on the target structure in the first initial image, the first real-time image, the second initial image, and the second real-time image;

[0015] Wherein, the image processing device calculates the deformation amount of the target structure according to the initial three-dimensional coordinates and the real-time three-dimensional coordinates of the feature points of the target structure, including:

[0016] The image processing device obtains the initial three-dimensional coordinates and the real-time three-dimensional coordinates of any one same feature point, and calculates the deformation amount of the target structure at any one same feature point according to the initial three-dimensional coordinates and the real-time three-dimensional coordinates of any one same feature point.

[0017] Preferably, the overall deformation amount of the target structure is calculated according to the deformation amounts of multiple feature points on the target structure.

[0018] Preferably, the image processing device performs image recognition on the first initial image and the second initial image to determine the initial three-dimensional coordinates of the feature points of the target structure, including:

[0019] For any feature point on the target structure, obtain its image coordinates (x1, y1) in the first initial image and its image coordinates (x2, y2) in the second initial image;

[0020] Obtain the three-dimensional coordinates (X1, Y1, Z1) of the first non-metric camera and the three-dimensional coordinates (X2, Y2, Z2) of the second non-metric camera;

[0021] Calculate the three-dimensional coordinates (X0, Y0, Z0) of any feature point on the target structure according to the image coordinates (x1, y1), the image coordinates (x2, y2), the three-dimensional coordinates (X1, Y1, Z1) and the three-dimensional coordinates (X2, Y2, Z2).

[0022] Preferably, the image processing device performs image recognition on the first real-time image and the second real-time image to determine the real-time three-dimensional coordinates of the feature points of the target structure, including:

[0023] For any feature point on the target structure, obtain its image coordinates (x3, y3) in the first real-time image and its image coordinates (x4, y4) in the second real-time image;

[0024] Obtain the three-dimensional coordinates (X3, Y3, Z3) of the first non-metric camera and the three-dimensional coordinates (X4, Y4, Z4) of the second non-metric camera;

[0025] Calculate the three-dimensional coordinates (X S , Y S , Z S ) of any feature point on the target structure according to the image coordinates (x3, y3), the image coordinates (x4, y4), the three-dimensional coordinates (X3, Y3, Z3) and the three-dimensional coordinates (X4, Y4, Z4).

[0026] An embodiment of the present invention also provides a deformation monitoring control system for non-metric cameras, including a first non-metric camera, a second non-metric camera and an image processing device, and the image processing device includes a coordinate calculation unit and a deformation amount calculation unit;

[0027] The first non-metric camera is used to photograph the target structure to obtain a first initial image, and the second non-metric camera is used to synchronously photograph the target structure to obtain a second initial image;

[0028] The first non-metric camera is used to photograph the target structure in real time to obtain a first real-time image, and the second non-metric camera is used to synchronously photograph the target structure in real time to obtain a second real-time image;

[0029] The coordinate calculation unit is configured to perform image recognition on the first initial image and the second initial image to determine the initial three-dimensional coordinates of the feature points of the target structure; and is configured to perform image recognition on the first real-time image and the second real-time image to determine the real-time three-dimensional coordinates of the feature points of the target structure.

[0030] The deformation amount calculation unit is configured to calculate the deformation amount of the target structure according to the initial three-dimensional coordinates and the real-time three-dimensional coordinates of the feature points of the target structure.

[0031] Preferably, the image processing device further includes a feature point matching unit.

[0032] The feature point matching unit is configured to match the feature points in the first initial image and the first real-time image, and match the feature points in the second initial image and the second real-time image; and determine any one identical feature point belonging to the target structure among the first initial image, the first real-time image, the second initial image, and the second real-time image according to the feature point matching result.

[0033] The deformation amount calculation unit is specifically configured to obtain the initial three-dimensional coordinates and the real-time three-dimensional coordinates of any one identical feature point, and calculate the deformation amount of the target structure at any one identical feature point according to the initial three-dimensional coordinates and the real-time three-dimensional coordinates of any one identical feature point.

[0034] Preferably, the overall deformation amount of the target structure is calculated according to the deformation amounts of multiple feature points on the target structure.

[0035] Preferably, the coordinate calculation unit is specifically configured to:

[0036] For any one feature point on the target structure, obtain its image coordinates (x1, y1) in the first initial image and its image coordinates (x2, y2) in the second initial image.

[0037] Obtain the three-dimensional coordinates (X1, Y1, Z1) of the first non-metric camera and the three-dimensional coordinates (X2, Y2, Z2) of the second non-metric camera.

[0038] Calculate the three-dimensional coordinates (X0, Y0, Z0) of any one feature point on the target structure according to the image coordinates (x1, y1), the image coordinates (x2, y2), the three-dimensional coordinates (X1, Y1, Z1), and the three-dimensional coordinates (X2, Y2, Z2).

[0039] Preferably, the coordinate calculation unit is specifically configured to:

[0040] For any feature point on the target structure, obtain its image coordinates (x3, y3) in the first real-time image and its image coordinates (x4, y4) in the second real-time image;

[0041] Obtain the three-dimensional coordinates (X3, Y3, Z3) of the first non-metric camera and the three-dimensional coordinates (X4, Y4, Z4) of the second non-metric camera;

[0042] According to the image coordinates (x3, y3), image coordinates (x4, y4), three-dimensional coordinates (X3, Y3, Z3) and three-dimensional coordinates (X4, Y4, Z4), calculate the three-dimensional coordinates (X S , Y S , Z S ) of any feature point on the target structure.

[0043] The embodiments of the present invention have the following advantages compared with other monitoring means:

[0044] (1) High sampling frequency, capable of quickly collecting data and obtaining instant deformation information of the monitored target, which is very suitable for dynamic deformation monitoring of the measured object;

[0045] (2) The measurement data is saved in the form of digital information, which is easy to store, easy to transmit, and can obtain a large amount of geometric and physical information of the measured object, having advantages for multi-point measurement;

[0046] (3) When observing the monitored target, it does not contact the measured target and does not affect the measured target, and can complete the measurement work under conditions with poor conditions such as noise, toxicity, hypoxia, and radioactivity intensity;

[0047] (4) Only two non-metric cameras are required as sensors, and the monitoring of the overall deformation of the monitoring structure can be realized within a limited economic cost.

[0048] Other features and advantages of the embodiments of the present invention will be described in the subsequent specification. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for the description of the embodiments or the prior art. Obviously, the following drawings are only some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0050] Figure 1 It is the main flowchart of a deformation monitoring control method for a non-metric camera in an embodiment of the present invention.

[0051] Figure 2 This is a partial flowchart of a deformation monitoring control method for a non - metric camera in an embodiment of the present invention.

[0052] Figure 3 This is a flowchart of matching feature points of images before and after deformation in an embodiment of the present invention.

[0053] Figure 4 This is a schematic diagram of a deformation monitoring control system for a non - metric camera in an embodiment of the present invention. Detailed implementation manners

[0054] Various exemplary embodiments, features, and aspects of the present disclosure will be described in detail below with reference to the accompanying drawings. Additionally, for a better illustration of the present invention, numerous specific details are given in the following detailed embodiments. Those skilled in the art should understand that the present invention can be implemented without some of these specific details. In some instances, means well - known to those skilled in the art are not described in detail so as to highlight the gist of the present invention.

[0055] Refer to Figure 1 , an embodiment of the present invention provides a deformation monitoring control method for a non - metric camera, which is implemented based on a deformation monitoring control system of a non - metric camera. The system includes a first non - metric camera, a second non - metric camera, and an image processing device. The first non - metric camera and the second non - metric camera are arranged on the left and right sides of a target structure to be monitored. The method includes the following steps:

[0056] Step S10: Use the first non - metric camera to capture a first initial image of the target structure, and simultaneously use the second non - metric camera to capture a second initial image of the target structure;

[0057] Step S20: Use the first non - metric camera to continuously capture a first real - time image of the target structure, and simultaneously use the second non - metric camera to continuously capture a second real - time image of the target structure;

[0058] Step S30: The image processing device performs image recognition on the first initial image and the second initial image to determine the initial three - dimensional coordinates of the feature points of the target structure;

[0059] Step S40: The image processing device performs image recognition on the first real - time image and the second real - time image to determine the real - time three - dimensional coordinates of the feature points of the target structure;

[0060] Step S50: The image processing device calculates the deformation amount of the target structure based on the initial three - dimensional coordinates and the real - time three - dimensional coordinates of the feature points of the target structure.

[0061] Specifically, in multi-view geometry, when two non-measurement cameras on the left and right are used to observe the same spatial point, the three-dimensional spatial coordinates of the point can be obtained through the positions, postures and image observation point coordinates of the two cameras. That is, based on the principle of binocular ranging, the spatial position relationship between the point on the target structure and the camera can be known. Then, combined with the known camera position coordinates, the three-dimensional spatial coordinates of the point on the target structure can be calculated; finally, the deformation value is calculated based on the coordinate results of the point on the target structure at different times.

[0062] The implementation of the method of this embodiment requires the following preparations: (1) Camera calibration; the non-measurement camera has no frame-marked coordinate system, its internal orientation elements are unknown, and there is a certain degree of lens distortion. Therefore, the internal parameters of the non-measurement digital camera must be reliably calibrated in order to accurately solve the object space coordinates of the target to be measured based on the image point coordinates. In this embodiment, the DLT method is specifically used to calibrate the first non-measurement camera and the second non-measurement camera. (2) Manually set deformation monitoring mark points. When monitoring the slope, if it is impossible to contact the measured slope surface and the area where the mark points cannot be manually set, the method of collecting feature points without prisms can be used instead of manual marking, thereby reducing the intensity of field work. In order to ensure accuracy and reduce manual errors, feature points should avoid choosing protruding rock tips, which are prone to scattering and refraction and can easily have a greater impact on accuracy. (3) Remote data synchronization and remote control. Most existing cameras have built-in Wi-Fi SD cards, which can be connected to mobile phones or Apple devices to achieve remote control and data synchronization. As the distance between the device and the camera increases, the response time increases. When the distance between the device and the camera exceeds 40m, there is almost no response. Therefore, a device needs to be deployed on site, which can be integrated into the image processing device, with a built-in 4G card, and can realize precise synchronous shooting of multiple cameras through hardware control of the emission pulse.

[0063] See also Figure 2 , the method further comprises:

[0064] Step S60: the image processing device matches feature points in the first initial image and the first real-time image, and matches feature points in the second initial image and the second real-time image;

[0065] Step S70, determining any common feature point on the target structure in the first initial image, the first real-time image, the second initial image, and the second real-time image according to the feature point matching result;

[0066] Specifically, the deformation of the target structure in the method of this embodiment is determined by calculating the three-dimensional coordinate displacement change of the feature points on the target structure. That is to say, it is necessary to first find the same feature point in the previous and next images to achieve this goal. This embodiment uses the image feature point matching method to determine it.

[0067] In one example, preferably but not limited to, the SIFT algorithm is used to extract the same feature points of the images before and after the deformation of the target structure, obtain the sub-pixel coordinates of the two feature points, and then obtain the deformation amount of this point. The overall process of the algorithm is as Figure 3 shown;

[0068] The extraction of the feature point deformation amount using the SIFT algorithm is mainly divided into four stages:

[0069] (1) Feature point detection;

[0070] The Gaussian pyramid is constructed by performing a convolution operation between the Gaussian function and the image, and then the Difference of Gaussian (DOG) pyramid is obtained by taking the difference; in addition, the precise positioning of the key points is achieved by fitting a three-dimensional quadratic function to accurately determine the positions of the key points, reaching sub-pixel accuracy;

[0071] (2) Determination of the direction angle;

[0072] After determining the feature points in each image in the previous step, it is also necessary to calculate a direction for each feature point and perform further operations based on this direction. The principle is to assign one or more direction angles to each feature point using the gradient direction distribution characteristics of the pixels in the local neighborhood of the feature point; all subsequent operations are carried out based on the position, scale, and angle of the feature points; the direction angle of the feature point is used to construct a feature description vector using the statistical histogram of the gradient directions of the pixels within the neighborhood window of the feature point;

[0073]

[0074] In the formula, L represents the pixel gray value at the corresponding point, m(x, y) and θ(x, y) are the gradient magnitude and direction of the pixel point (x, y) respectively; after performing the above calculations on 64 points, statistics are carried out using a histogram; the horizontal axis of the histogram is the gradient direction angle, ∈0 to 360 degrees, with one column every 10 degrees, and the vertical axis is the Gaussian weighted cumulative value of the corresponding gradient value;

[0075] (3) Generation of the feature point descriptor; after obtaining the main direction and amplitude of the feature point, it is also necessary to describe the feature point to prepare for the matching between points; first, the coordinate axis is rotated to the main direction of the feature point. Only by describing the feature point with the main direction as the zero direction can it have rotational invariance; to enhance the robustness of the matching, 4×4, a total of 16 seed points can be used to describe each feature point. Since each seed point has 8 direction vectors, each feature point can generate a 128-dimensional vector. These 128 values are the SIFT feature point descriptors. At this time, the feature point description vector is no longer affected by geometric factors such as scale change and rotation; finally, the length of the SIFT feature vector is normalized to remove the influence of illumination changes;

[0076] (4) Feature point matching and deformation amount calculation; After obtaining the feature point descriptors of the two images, the Euclidean distance between the feature vectors of the two feature points is used as the similarity measurement criterion for the feature points in the two images; where xi and yi are the components of the feature vectors of the feature points to be matched in the two images respectively

[0077]

[0078] Select a certain feature point in the pre-deformation image, and traverse to obtain the top two feature points with the shortest Euclidean distance in the post-deformation image; if the ratio of the nearest distance to the second nearest distance is less than a certain threshold, then this point is considered a matching point; after finding all matching point pairs, sort them by the Euclidean distance and remove the matching point pairs with larger distances.

[0079] More specifically, the fundamental matrix is defined by the following equation:

[0080] x′ T Fx = 0

[0081] where is an arbitrary pair of matching points in the two images; since the matching of each group of points provides a linear equation for calculating the F coefficient, when at least 7 points are given (a 3×3 homogeneous matrix minus a scale, and a rank-2 constraint), the equation can calculate the unknown F; we denote the coordinates of the points as x = (x, y, 1) T , x′ = (x′, y′, 1) T , then the corresponding equation is:

[0082]

[0083] After expansion, we have:

[0084] x′xf 11 +x′yf 12 +x′f 13 +y′xf 21 +y′yf 22 +y′f 23 +xf 31 +yf 32 +f 33 = 0

[0085] Writing the matrix F in the form of a column vector, we have:

[0086] [x′x x′y x′ y′x y′y y′ x y 1]f = 0

[0087] Given a set of n groups of points, we have the following equation:

[0088]

[0089] The algorithmic process for determining the fundamental matrix is as follows:

[0090] Normalization: According to transform the image coordinates, where T and T′ are normalization transformations composed of translation and scaling;

[0091] Solve for the corresponding matching fundamental matrix

[0092] Find the linear solution: Use the coefficient matrix determined by the corresponding point set to determine the singular vector of the smallest singular value of

[0093] Singularity constraint: Use SVD to decompose, set its smallest singular value to 0, and obtain such that

[0094] Denormalization: Let The matrix F is the fundamental matrix corresponding to the data .

[0095] Among them, step S50 includes:

[0096] The image processing device acquires the initial three-dimensional coordinates and real-time three-dimensional coordinates of any one of the same feature points, and calculates the deformation amount of the target structure at any one of the same feature points according to the initial three-dimensional coordinates and real-time three-dimensional coordinates of any one of the same feature points.

[0097] Specifically, the overall deformation amount of the target structure is calculated based on the deformation amounts of multiple feature points on the target structure, and can be represented by using the front and back coordinate displacements of multiple feature points.

[0098] Specifically, step S30 includes:

[0099] Step S301: For any one feature point on the target structure, acquire its image coordinates (x1, y1) in the first initial image and its image coordinates (x2, y2) in the second initial image;

[0100] Step S302: Acquire the three-dimensional coordinates (X1, Y1, Z1) of the first non-metric camera and the three-dimensional coordinates (X2, Y2, Z2) of the second non-metric camera;

[0101] Step S303: Calculate the three-dimensional coordinates (X0, Y0, Z0) of any feature point on the target structure according to the image coordinates (x1, y1), image coordinates (x2, y2), three-dimensional coordinates (X1, Y1, Z1) and three-dimensional coordinates (X2, Y2, Z2).

[0102] Specifically, step S40 includes:

[0103] Step S401: For any feature point on the target structure, obtain its image coordinates (x3, y3) in the first real-time image and its image coordinates (x4, y4) in the second real-time image;

[0104] Step S402: Obtain the three-dimensional coordinates (X3, Y3, Z3) of the first non-metric camera and the three-dimensional coordinates (X4, Y4, Z4) of the second non-metric camera;

[0105] Step S403: Calculate the three-dimensional coordinates (X S , Y S , Z S ) of any feature point on the target structure according to the image coordinates (x3, y3), image coordinates (x4, y4), three-dimensional coordinates (X3, Y3, Z3) and three-dimensional coordinates (X4, Y4, Z4).

[0106] Referring to Figure 4 , another embodiment of the present invention further provides a deformation monitoring control system for a non-metric camera, including a first non-metric camera 1, a second non-metric camera 2 and an image processing device 3, and the image processing device 3 includes a coordinate calculation unit 31 and a deformation amount calculation unit 32;

[0107] The first non-metric camera 1 is used to photograph the target structure to obtain a first initial image, and the second non-metric camera 2 is used to synchronously photograph the target structure to obtain a second initial image;

[0108] The first non-metric camera 1 is used to photograph the target structure in real time to obtain a first real-time image, and the second non-metric camera 2 is used to synchronously photograph the target structure in real time to obtain a second real-time image;

[0109] The coordinate calculation unit 31 is used to perform image recognition on the first initial image and the second initial image to determine the initial three-dimensional coordinates of the feature points of the target structure; and is used to perform image recognition on the first real-time image and the second real-time image to determine the real-time three-dimensional coordinates of the feature points of the target structure;

[0110] The deformation amount calculation unit 32 is configured to calculate the deformation amount of the target structure according to the initial three-dimensional coordinates and the real-time three-dimensional coordinates of the feature points of the target structure.

[0111] Specifically, the image processing device 3 further includes a feature point matching unit 33.

[0112] The feature point matching unit 33 is configured to match the feature points in the first initial image and the first real-time image, and match the feature points in the second initial image and the second real-time image; and determine any same feature point belonging to the target structure in the first initial image, the first real-time image, the second initial image, and the second real-time image according to the feature point matching result.

[0113] The deformation amount calculation unit 32 is specifically configured to obtain the initial three-dimensional coordinates and the real-time three-dimensional coordinates of any same feature point, and calculate the deformation amount of the target structure at any same feature point according to the initial three-dimensional coordinates and the real-time three-dimensional coordinates of any same feature point.

[0114] Specifically, the overall deformation amount of the target structure is calculated according to the deformation amounts of multiple feature points on the target structure.

[0115] Specifically, the coordinate calculation unit 31 is specifically configured to:

[0116] For any feature point on the target structure, obtain its image coordinates (x1, y1) in the first initial image and its image coordinates (x2, y2) in the second initial image.

[0117] Obtain the three-dimensional coordinates (X1, Y1, Z1) of the first non-metric camera 1 and the three-dimensional coordinates (X2, Y2, Z2) of the second non-metric camera 2.

[0118] Calculate the three-dimensional coordinates (X0, Y0, Z0) of any feature point on the target structure according to the image coordinates (x1, y1), the image coordinates (x2, y2), the three-dimensional coordinates (X1, Y1, Z1), and the three-dimensional coordinates (X2, Y2, Z2).

[0119] Specifically, the coordinate calculation unit 31 is specifically configured to:

[0120] For any feature point on the target structure, obtain its image coordinates (x3, y3) in the first real-time image and its image coordinates (x4, y4) in the second real-time image.

[0121] Obtain the three-dimensional coordinates (X3, Y3, Z3) of the first non-measuring camera 1 and the three-dimensional coordinates (X4, Y4, Z4) of the second non-measuring camera 2;

[0122] Calculate the three-dimensional coordinates (X S , Y S , Z S ) of any feature point on the target structure according to the image coordinates (x3, y3), image coordinates (x4, y4), three-dimensional coordinates (X3, Y3, Z3) and three-dimensional coordinates (X4, Y4, Z4).

[0123] The system of this embodiment corresponds to the method of the above embodiment. Therefore, for the parts not detailed in the system of this embodiment, reference can be made to the content of the method of the above embodiment, and details will not be repeated here.

[0124] The embodiments of the present invention have the following advantages compared with other monitoring means:

[0125] (1) High sampling frequency, capable of quickly collecting data and obtaining the instantaneous deformation information of the monitored target, which is very suitable for dynamic deformation monitoring of the measured object;

[0126] (2) The measured data is saved in the form of digital information, which is easy to store, easy to transmit, and can obtain a large amount of geometric and physical information of the measured object, having advantages for multi-point measurement;

[0127] (3) When observing the monitored target, it does not contact the measured target and does not affect the measured target, and can complete the measurement work under conditions such as noise, toxicity, hypoxia, and radioactive intensity with poor conditions;

[0128] (4) Only two non-measuring cameras are required as sensors, and the monitoring of the overall deformation of the monitoring structure can be realized within a limited economic cost.

[0129] The above has described the embodiments of the present invention. The above description is exemplary, not exhaustive, and is not limited to the disclosed embodiments. Many modifications and changes are obvious to those of ordinary skill in the art in the technical field without departing from the scope and spirit of the described embodiments. The selection of the terms used herein is intended to best explain the principles of the embodiments, practical applications, or improvements to the technology in the market, or to enable other ordinary technical personnel in the technical field to understand the disclosed embodiments.

Claims

1. A deformation monitoring control method for a non - measuring camera, characterized in that, It is realized based on a deformation monitoring control system of a non-measuring camera. The system includes a first non-measuring camera, a second non-measuring camera and an image processing device. The method includes the following steps: The first non-measuring camera is used to take a first initial image of the target structure, and simultaneously the second non-measuring camera is used to take a second initial image of the target structure; The first non-measuring camera is used to take a first real-time image of the target structure in real time, and simultaneously the second non-measuring camera is used to take a second real-time image of the target structure in real time; The image processing device performs image recognition on the first initial image and the second initial image to determine the initial three-dimensional coordinates of the feature points of the target structure; The image processing device performs image recognition on the first real-time image and the second real-time image to determine the real-time three-dimensional coordinates of the feature points of the target structure; The image processing device calculates the deformation amount of the target structure according to the initial three-dimensional coordinates and the real-time three-dimensional coordinates of the feature points of the target structure; The image processing device matches the feature points in the first initial image and the first real-time image, and matches the feature points in the second initial image and the second real-time image; Determine any same feature point belonging to the target structure in the first initial image, the first real-time image, the second initial image, and the second real-time image according to the feature point matching result; Among them, the image processing device calculates the deformation amount of the target structure according to the initial three-dimensional coordinates and the real-time three-dimensional coordinates of the feature points of the target structure, including: The image processing device obtains the initial three-dimensional coordinates and the real-time three-dimensional coordinates of any same feature point, and calculates the deformation amount of the target structure at any same feature point according to the initial three-dimensional coordinates and the real-time three-dimensional coordinates of any same feature point; The image processing device performs image recognition on the first initial image and the second initial image to determine the initial three-dimensional coordinates of the feature points of the target structure, including: For any feature point on the target structure, obtain its image coordinates (x1, y1) in the first initial image and its image coordinates (x2, y2) in the second initial image; Obtain the three-dimensional coordinates (X1, Y1, Z1) of the first non-measuring camera and the three-dimensional coordinates (X2, Y2, Z2) of the second non-measuring camera; Calculate the three-dimensional coordinates (X0, Y0, Z0) of any feature point on the target structure according to the image coordinates (x1, y1), the image coordinates (x2, y2), the three-dimensional coordinates (X1, Y1, Z1) and the three-dimensional coordinates (X2, Y2, Z2); The image processing device performs image recognition on the first real-time image and the second real-time image to determine the real-time three-dimensional coordinates of the feature points of the target structure, including: For any feature point on the target structure, obtain its image coordinates (x3, y3) in the first real-time image and its image coordinates (x4, y4) in the second real-time image; Obtain the three-dimensional coordinates (X3, Y3, Z3) of the first non-metric camera and the three-dimensional coordinates (X4, Y4, Z4) of the second non-metric camera; Calculate the three-dimensional coordinates (X S , Y S , Z S ) of any feature point on the target structure according to the image coordinates (x3, y3), image coordinates (x4, y4), three-dimensional coordinates (X3, Y3, Z3) and three-dimensional coordinates (X4, Y4, Z4).

2. The method according to claim 1, characterized in that, The overall deformation amount of the target structure is calculated based on the deformation amounts of multiple feature points on the target structure.

3. A deformation monitoring control system for a non - measuring camera, characterized in that, It includes a first non-metric camera, a second non-metric camera, and an image processing device, and the image processing device includes a coordinate calculation unit and a deformation amount calculation unit; The first non-metric camera is used to capture a first initial image of the target structure, and the second non-metric camera is used to synchronously capture a second initial image of the target structure; The first non-metric camera is used to capture a first real-time image of the target structure in real time, and the second non-metric camera is used to synchronously capture a second real-time image of the target structure in real time; The coordinate calculation unit is used to perform image recognition on the first initial image and the second initial image to determine the initial three-dimensional coordinates of the feature points of the target structure; And it is used to perform image recognition on the first real-time image and the second real-time image to determine the real-time three-dimensional coordinates of the feature points of the target structure; The deformation amount calculation unit is used to calculate the deformation amount of the target structure based on the initial three-dimensional coordinates and the real-time three-dimensional coordinates of the feature points of the target structure; The image processing device further includes a feature point matching unit, The feature point matching unit is used to match the feature points in the first initial image and the first real-time image, and match the feature points in the second initial image and the second real-time image; And determine any same feature point belonging to the target structure in the first initial image, the first real-time image, the second initial image, and the second real-time image according to the feature point matching result; The deformation amount calculation unit is specifically used to obtain the initial three-dimensional coordinates and the real-time three-dimensional coordinates of any same feature point, and calculate the deformation amount of the target structure at any same feature point based on the initial three-dimensional coordinates and the real-time three-dimensional coordinates of any same feature point; The coordinate calculation unit is specifically used for: For any feature point on the target structure, obtain its image coordinates (x1, y1) in the first initial image and its image coordinates (x2, y2) in the second initial image; Obtain the three-dimensional coordinates (X1, Y1, Z1) of the first non-metric camera and the three-dimensional coordinates (X2, Y2, Z2) of the second non-metric camera; Calculate the three-dimensional coordinates (X0, Y0, Z0) of any feature point on the target structure according to the image coordinates (x1, y1), the image coordinates (x2, y2), the three-dimensional coordinates (X1, Y1, Z1), and the three-dimensional coordinates (X2, Y2, Z2); The coordinate calculation unit is specifically used for: For any feature point on the target structure, obtain its image coordinates (x3, y3) in the first real-time image and its image coordinates (x4, y4) in the second real-time image; Obtain the three-dimensional coordinates (X3, Y3, Z3) of the first non-metric camera and the three-dimensional coordinates (X4, Y4, Z4) of the second non-metric camera; Calculate the three-dimensional coordinates (X S , Y S , Z S ) of any feature point on the target structure according to the image coordinates (x3, y3), image coordinates (x4, y4), three-dimensional coordinates (X3, Y3, Z3), and three-dimensional coordinates (X4, Y4, Z4).

4. The system according to claim 3, wherein The overall deformation amount of the target structure is calculated based on the deformation amounts of multiple feature points on the target structure.

Images