Cross time-frequency skeleton separation method based on optimal pixel connection rule

Abstract

The invention discloses a cross time-frequency skeleton separation method based on an optimal pixel connection rule. The method comprises the following steps: firstly, obtaining a complete skeleton of signal time-frequency distribution through mathematical morphology operation; secondly, capturing single-layer connected skeleton points in a non-cross region by using a skeleton tracking scheme based on eight pixel neighborhoods; and finally, a new weighted direction skeleton tracking strategy is utilized to determine an optimal skeleton communication direction, so that correct separation and extraction of a crossed region skeleton are realized, the problem that frequency modulation characteristics and frequency dispersion characteristics are difficult to extract at the same time by a traditional ridge extraction method can be effectively solved, accurate separation of a frequency modulation-frequency dispersion component crossed skeleton is realized, and the accuracy of the frequency modulation-frequency dispersion component crossed skeleton is improved. And the method has good noise robustness.

Description

Technical Field

[0001] The invention belongs to the technical field of signal processing, and in particular relates to a cross-time-frequency skeleton separation method based on an optimal pixel connectivity rule. Background Art

[0002] Signals from physical systems such as radar, sonar, and mechanical equipment contain important information about their states. Signal processing techniques are fundamental tools for elucidating and exploring these physical systems. In practical applications, signals are often characterized by strong noise interference, coexistence of multiple components, and non-stationary modulation. In some cases, non-stationary signals contain both frequency modulation components and dispersion components, and they may cross each other in the time-frequency domain. Time-frequency analysis methods can simultaneously characterize frequency modulation features and dispersion features in the time-frequency domain, so they are often used to analyze complex non-stationary signals. At present, a variety of time-frequency analysis methods have been developed to improve the performance of time-frequency representation, such as synchronous compression transform, synchronous extraction transform, and parametric time-frequency transform. However, the above time-frequency analysis methods can only improve the energy concentration of the time-frequency distribution, and it is difficult to effectively separate the time-frequency features of the cross-signal components. It is worth mentioning that the peak detection method can extract component ridges from the time-frequency distribution of the signal, which is often used to estimate the instantaneous frequency or group delay in complex non-stationary signals. In particular, some cross-ridge separation methods, such as adaptive linear frequency modulation wavelet transform and ridge path recombination, can effectively separate the cross-linear frequency modulation features. However, most existing ridge separation methods can only track the maximum time-frequency coefficient along the time axis, and are therefore only suitable for estimating the cross-instantaneous frequency ridges of nonlinear frequency-modulated signals. It is difficult to separate the cross-skeleton of the frequency-modulation-dispersion component, and a method is needed to solve this problem. Summary of the invention

[0003] In order to overcome the above-mentioned shortcomings, a cross-time-frequency skeleton separation method based on the optimal pixel connectivity rule is proposed. First, the complete skeleton of the signal time-frequency distribution is obtained through mathematical morphological operations; secondly, a skeleton tracking scheme based on the eight-neighborhood of pixels is used to capture the single-layer connected skeleton points in the non-cross-region; finally, a new weighted directional skeleton tracking strategy is used to determine the optimal skeleton connectivity direction, thereby realizing the correct separation and extraction of the skeleton in the cross-region. It can effectively solve the problem that the traditional ridge extraction method is difficult to extract the frequency modulation features and the dispersion features at the same time, realize the accurate separation of the frequency modulation-dispersion component cross-skeleton, and has good noise robustness.

[0004] To achieve the above purpose, the technical solution adopted by the present invention is to provide a cross-time-frequency skeleton separation method based on the optimal pixel connectivity rule. It includes the following steps:

[0005] Step S1, obtaining the complete skeleton S(t,f) of the signal time-frequency distribution through mathematical morphological operations;

[0006] Step S2, using a skeleton tracking scheme based on eight pixel neighborhoods to capture single-layer connected skeleton points in non-intersection areas;

[0007] Step S2.1, according to the pixel connectivity definition, from a skeleton endpoint (x e ,y e ) and track the current skeleton point in the non-intersection area Update the current skeleton point with non-zero pixels in the eight-neighborhood:

[0008]

[0009] in, Indicates the tracked n+1th skeleton point, Indicates the direction coordinate of the current skeleton point tracking the only non-zero pixel in its eight-neighborhood;

[0010] Step S2.2, calculate and record the angle corresponding to the current moving direction:

[0011]

[0012] in,

[0013]

[0014] Step S2.3, the original current skeleton point The time-frequency coefficients of are set to zero to ensure the single connectivity of the current skeleton point;

[0015] Step S3, using a weighted directional skeleton tracking strategy to determine the optimal skeleton connection direction to achieve correct separation of the cross-region skeleton;

[0016] Step S3.1, obtain the weighted direction angle WD according to the path direction of the tracked skeleton point:

[0017]

[0018] Where N represents the number of tracked skeleton points, w n is the distance weight between the current point and the nth tracked skeleton point:

[0019]

[0020] Among them, e is a natural constant. The closer the nth skeleton point is to the current point, the larger e n / (N-1) The larger it is, the higher the distance weight;

[0021] Step S3.2, obtain the optimal tracking direction angle by solving the following optimization problem:

[0022]

[0023] Step S3.3, according to the optimal tracking direction angle, connect the skeleton points belonging to the current component in the intersection area, and return to step S2 to continue tracking the skeleton points until another skeleton endpoint is tracked to obtain the skeleton of a component.

[0024] According to the cross-time-frequency skeleton separation method based on the optimal pixel connectivity rule described in the present invention,

[0025] A further preferred technical solution is: the specific steps of step S1 are:

[0026] Step S1.1, binarize the signal time-frequency distribution TF(t,f) to remove the noise time-frequency coefficient:

[0027]

[0028] in, represents the binary time-frequency distribution, δ is the denoising threshold, and is obtained through the confidence interval of the signal time-frequency coefficient:

[0029] δ=mean(TF>0)-2std(TF>0)

[0030] Among them, mean(g) and std(g) represent the mean and standard deviation of the time-frequency coefficients respectively;

[0031] Step S1.2, using morphological skeleton extraction operation to obtain binary time-frequency distribution The skeleton:

[0032]

[0033] Among them, SE[g] represents the morphological skeleton extraction operation, (x s ,y s ) represents the skeleton point, (x e ,y e ) represents the skeleton endpoint.

[0034] According to the cross-time-frequency skeleton separation method based on the optimal pixel connectivity rule described in the present invention, its further preferred technical solution is: it also includes step S4: restarting step S2, extracting the skeleton of the next component, until there are no non-zero time-frequency coefficients in the binary time-frequency distribution, and completing the separation of the skeletons of all components of the complete skeleton.

[0035] Compared with the prior art, the technical solution of the present invention has the following advantages / benefits:

[0036] 1. The present invention uses connectivity rules to capture the time-frequency skeleton of components in non-intersection areas, which can effectively solve the problem that traditional ridge extraction methods are difficult to simultaneously extract frequency modulation features and dispersion features.

[0037] 2. The present invention uses a new weighted directional skeleton tracking strategy to determine the skeleton connection direction, which can achieve accurate separation of the frequency modulation-dispersion component cross skeleton and has good noise robustness. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required for use in the embodiments will be briefly introduced below. It should be understood that the following drawings only show certain embodiments of the present invention and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without paying creative work.

[0039] Figure 1 It is a schematic diagram of the cross-time-frequency skeleton separation of the present invention.

[0040] Figure 2 It is a schematic diagram of the signal time-frequency skeleton extraction result of the present invention.

[0041] Figure 3 Schematic diagram of the direction angles and coordinates of the eight-neighborhood pixel of the present invention.

[0042] Figure 4 It is a schematic diagram of the signal time-frequency skeleton separation result of the present invention.

[0043] Figure 5 It is a schematic diagram of the time-frequency skeleton extraction result of the actual pilot whale sound signal of the present invention.

[0044] Figure 6 It is a schematic diagram of the time-frequency skeleton separation results of the actual pilot whale sound signal of the present invention. DETAILED DESCRIPTION

[0045] In order to make the purpose, technical solutions and advantages of the present invention clearer, the technical solutions in the embodiments of the present invention are clearly and completely described below. Obviously, the described embodiments are part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in the field without creative work are within the scope of protection of the present invention. Therefore, the detailed description of the embodiments of the present invention provided below is not intended to limit the scope of the claimed invention, but merely represents the selected embodiments of the present invention.

[0046] Embodiment 1:

[0047] A cross-time-frequency skeleton separation method based on the optimal pixel connectivity rule, such as Figure 1 As shown, the following steps are included:

[0048] Step S1, obtain the complete skeleton of the signal time-frequency distribution through mathematical morphological operations, such as Figure 2 As shown;

[0049] Step S1.1, binarize the signal time-frequency distribution TF(t,f) to remove the noise time-frequency coefficient:

[0050]

[0051] in, represents the binary time-frequency distribution, δ is the denoising threshold, and is obtained through the confidence interval of the signal time-frequency coefficient:

[0052] δ=mean(TF>0)-2std(TF>0)

[0053] Among them, mean(g) and std(g) represent the mean and standard deviation of the time-frequency coefficients respectively;

[0054] Step S1.2, using morphological skeleton extraction operation to obtain binary time-frequency distribution The skeleton:

[0055]

[0056] Among them, SE[g] represents the morphological skeleton extraction operation, (x s ,y s ) represents the skeleton point, (x e ,y e )Skeleton endpoints;

[0057] Step S2, using a skeleton tracking scheme based on eight pixel neighborhoods to capture single-layer connected skeleton points in non-intersection areas;

[0058] Step S2.1, according to the pixel connectivity definition, from a skeleton endpoint (x e ,y e ) and track the current skeleton point in the non-intersection area Update the current skeleton point with non-zero pixels in the eight-neighborhood:

[0059]

[0060] in, Indicates the tracked n+1th skeleton point, Indicates the direction coordinate of the current skeleton point tracking the only non-zero pixel in its eight-neighborhood, such as Figure 3 As shown;

[0061] Step S2.2, calculate and record the angle corresponding to the current moving direction:

[0062]

[0063] in,

[0064]

[0065] Step S2.3, the original current skeleton point The time-frequency coefficients of are set to zero to ensure the single connectivity of the current skeleton point;

[0066] Step S3, using a new weighted directional skeleton tracking strategy to determine the optimal skeleton connection direction, thereby achieving the correct separation of the cross-region skeleton;

[0067] Step S3.1, obtain the weighted direction angle WD according to the path direction of the tracked skeleton point:

[0068]

[0069] Where N represents the number of tracked skeleton points, w n is the distance weight between the current point and the nth tracked skeleton point:

[0070]

[0071] Among them, e is a natural constant. The closer the nth skeleton point is to the current point, the larger e n / (N-1) The larger it is, the higher the distance weight;

[0072] Step S3.2, obtain the optimal tracking direction angle by solving the following optimization problem:

[0073]

[0074] Step S3.3, according to the optimal tracking direction angle, connect the skeleton points belonging to the current component in the intersection area, and return to step S2 to continue tracking the skeleton points until another skeleton endpoint is tracked to obtain the skeleton of a component.

[0075] S4, restart step S2, extract the skeleton of the next component until there are no non-zero time-frequency coefficients in the binary time-frequency distribution, such as Figure 4 As shown, the separation of the skeleton of all components of the complete skeleton is completed.

[0076] Embodiment 2:

[0077] The method provided by the present invention will be used to separate the time-frequency skeleton of the actual signal. Figure 5 The method provided by the present invention is the result of time-frequency skeleton extraction of actual pilot whale sound signals. Figure 6 The method provided by the present invention is used to separate the time-frequency skeleton of the actual pilot whale sound signal. The results show that the present invention can accurately separate the frequency modulation-dispersion component cross skeleton in complex non-stationary signals.

[0078] In the description of the present invention, it should be understood that the terms "center", "longitudinal", "lateral", "length", "width", "thickness", "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inside", "outside", "clockwise", "counterclockwise" and the like indicate orientations or positional relationships based on the orientations or positional relationships shown in the accompanying drawings, and are only for the convenience of describing the present invention and simplifying the description, and do not indicate or imply that the referred device or element must have a specific orientation, be constructed and operated in a specific orientation, and therefore should not be understood as a limitation on the present invention.

[0079] In the present invention, unless otherwise clearly specified and limited, the terms "installed", "connected", "connected", "fixed" and the like should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection or an indirect connection through an intermediate medium, it can be the internal connection of two elements or the interaction relationship between two elements. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0080] The above are only preferred embodiments of the present invention. It should be noted that the above preferred embodiments should not be regarded as limiting the present invention, and the protection scope of the present invention should be based on the scope defined by the claims. For ordinary technicians in this technical field, several improvements and modifications can be made without departing from the spirit and scope of the present invention, and these improvements and modifications should also be regarded as the protection scope of the present invention.

Claims

1. A cross-time-frequency skeleton separation method based on optimal pixel connectivity rule, characterized in that: The following steps are involved: Step S1, obtaining the complete skeleton S(t,f) of the signal time-frequency distribution through mathematical morphological operations; Step S2, using a skeleton tracking scheme based on eight pixel neighborhoods to capture single-layer connected skeleton points in non-intersection areas; Step S2.1, according to the pixel connectivity definition, from a skeleton endpoint (x e ,y e ) and track the current skeleton point in the non-intersection area Update the current skeleton point with non-zero pixels in the eight-neighborhood: in, Indicates the tracked n+1th skeleton point, Indicates the direction coordinate of the current skeleton point tracking the only non-zero pixel in its eight-neighborhood; Step S2.2, calculate and record the angle corresponding to the current moving direction: in, Step S2.3, the original current skeleton point The time-frequency coefficients of are set to zero to ensure the single connectivity of the current skeleton point; Step S3, using a weighted directional skeleton tracking strategy to determine the optimal skeleton connection direction to achieve correct separation of the cross-region skeleton; Step S3.1, obtain the weighted direction angle WD according to the path direction of the tracked skeleton point: Where N represents the number of tracked skeleton points, w n is the distance weight between the current point and the nth tracked skeleton point: Among them, e is a natural constant. The closer the nth skeleton point is to the current point, the larger e n / (N-1) The larger it is, the higher the distance weight; Step S3.2, obtain the optimal tracking direction angle by solving the following optimization problem: Step S3.3, according to the optimal tracking direction angle, connect the skeleton points belonging to the current component in the intersection area, and return to step S2 to continue tracking the skeleton points until another skeleton endpoint is tracked to obtain the skeleton of the current component.

2. The cross-time-frequency skeleton separation method based on the optimal pixel connectivity rule according to claim 1, characterized in that: The method further includes step S4: restarting step S2 to extract the skeleton of the next component until there is no non-zero time-frequency coefficient in the binary time-frequency distribution, thereby completing the separation of the skeletons of all components in the complete skeleton.

3. The cross-time-frequency skeleton separation method based on the optimal pixel connectivity rule according to claim 1, characterized in that: The specific steps of step S1 are: Step S1.1, binarize the signal time-frequency distribution TF(t,f) to remove the noise time-frequency coefficient: in, represents the binary time-frequency distribution, δ is the denoising threshold, and is obtained through the confidence interval of the signal time-frequency coefficient: δ=mean(TF>0)-2std(TF>0) Among them, mean(g) and std(g) represent the mean and standard deviation of the time-frequency coefficients respectively; Step S1.2, using morphological skeleton extraction operation to obtain binary time-frequency distribution The skeleton: Among them, SE[g] represents the morphological skeleton extraction operation, (x s ,y s ) represents the skeleton point, (x e ,y e ) represents the skeleton endpoint.

Images