Object detection method based on spatiotemporal voidness, electronic device, and storage medium

Abstract

The application provides an object detection method based on space-time gap degree, an electronic device and a storage medium. The method comprises the following steps: obtaining multiple frames of echo images with time sequence in a target region, constructing an echo pyramid corresponding to each frame of echo image to obtain multi-scale observation in the spatial dimension of each frame of echo image; based on the echo pyramid, observing under the same spatial scale along the time dimension to obtain a space-time gap degree image under multi-scale observation, and determining a target object in the target region according to the space-time gap degree image. The method provided by the application can improve the accuracy of the object detection result based on the space-time gap degree.

Description

Technical Field

[0001] This application relates to the field of active sonar technology, and in particular to an object detection method, electronic device and storage medium based on spatiotemporal gapity. Background Technology

[0002] Active sonar is a type of underwater acoustic device commonly used in underwater surveillance and port protection. It detects underwater objects, such as divers and unmanned underwater vehicles, by actively transmitting sound waves and receiving the backscattered signals. This protects critical military and commercial equipment. However, underwater topography, sediment, fish, and surface waves in shallow water environments can create high-intensity reverberation in the backscattered signals, making it difficult to distinguish the trajectories of clutter and intruding targets in the active sonar echo map. This reduces the accuracy of active sonar in detecting underwater objects. Summary of the Invention

[0003] In view of the above, it is necessary to provide an object detection method, electronic device and storage medium based on spatiotemporal porosity, which can solve the problem of not being able to effectively distinguish the reverberation of target objects and clutter scatterers in active sonar echo maps, and improve the accuracy of underwater object detection results based on spatiotemporal porosity active sonar.

[0004] The object detection method based on spatiotemporal gaps includes: acquiring multiple original echo images of a target region, each original echo image corresponding to a time label; constructing an echo pyramid corresponding to each original echo image, the echo pyramid containing multiple low-resolution echo images of the corresponding original echo image, each image in the echo pyramid corresponding to a hierarchical label; calculating the spatiotemporal gaps between images corresponding to preset hierarchical labels in multiple echo pyramids based on the time labels, obtaining a spatiotemporal gap map; and determining the target object in the target region based on the spatiotemporal gap map.

[0005] Optionally, constructing the echo pyramid corresponding to each original echo image includes: performing a preset number of resolution reduction processes on each original echo image to obtain multiple low-resolution echo images of each original echo image at multiple scales; and constructing the echo pyramid using each original echo image and the corresponding multiple low-resolution echo images.

[0006] Optionally, the resolution reduction process includes low-pass filtering and downsampling.

[0007] Optionally, the resolution reduction processing for each original echo image a preset number of times includes: performing Gaussian filtering on the original echo image I0 to obtain a low-resolution echo image I1; and performing Gaussian filtering on the low-resolution echo image I1...l-1 Gaussian filtering is performed to obtain a low-resolution echo map I. l ,in, Where, l∈{1,2,3,…,L}, and L represents the preset number of times; G(m,n) represents a Gaussian window of size (2c+1)×(2c+1) with standard deviation σ=c; * represents the convolution operation, I l (i,j) represents the low-resolution echo map I. l The echo at position (i,j), I l-1 Represents low-resolution echo map I l-1 The echo at the midpoint (2i-1-m, 2j-1-n).

[0008] Optionally, constructing the echo pyramid using each original echo image and corresponding multiple low-resolution echo images includes: using the original echo image I0 as the zeroth level of the echo pyramid, and using the low-resolution echo images I... l As the l-th level of the echo pyramid, an echo pyramid with a total of L+1 levels is constructed, and the level label corresponding to the l-th level image in the echo pyramid is l.

[0009] Optionally, the step of calculating the spatiotemporal gap between images corresponding to preset level labels in multiple echo pyramids based on the time labels includes: using a preset sliding window to slide between images in multiple l-th levels of the multiple echo pyramids based on a sliding window algorithm and the order of the time labels, to calculate the spatiotemporal gap between images corresponding to the l-th level.

[0010] Optionally, the calculation formula for the spatiotemporal porosity includes: spatiotemporal porosity Where, E(S)=∑SP(S,r) t ,r b ,r r E(S) represents the first moment of the sliding window mass distribution; E(S) 2 )=∑S 2 P(S,r t ,r b ,r r ,l), E(S 2 ) represents the second moment of the mass distribution of the sliding window; This indicates that the mass is S and the time dimension is r in stage l. t The orientation dimension is r b And distance dimension r r The number of sliding windows, N(r) t ,r b ,r r(l) indicates that the image corresponding to level l uses a time dimension of r. t The orientation dimension is r b And distance dimension r r The number of sliding windows, P(S,r) t ,r b ,r r ,l) represents the probability that the sliding window of the l-th level has a quality of S, and k represents the higher-order moments.

[0011] The computer-readable storage medium stores at least one instruction, which, when executed by a processor, implements the object detection method based on spatiotemporal gapity or the object detection method based on spatiotemporal gapity.

[0012] The electronic device includes a memory and at least one processor. The memory stores at least one instruction, which, when executed by the at least one processor, implements the object detection method based on spatiotemporal gapity.

[0013] Compared to existing technologies, the object detection method based on spatiotemporal porosity provided in this application obtains multi-frame echo maps with time series, constructs an echo pyramid corresponding to each frame echo map to obtain multi-scale observations in the spatial dimension of each frame echo map; then, based on the echo pyramid, observations at the same spatial scale are performed along the time dimension to obtain a porosity map under multi-scale observation. This method can solve the problem that single-scale observation cannot effectively distinguish between target objects and reverberation by constructing an echo pyramid. Furthermore, the echo pyramid also reduces the computational complexity of porosity features, thus reducing the computational load of multi-scale observation. In addition, extending porosity from the time dimension to the spatiotemporal dimension improves the accuracy of object detection results based on spatiotemporal porosity. Attached Figure Description

[0014] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of this application. For those skilled in the art, other drawings can be obtained based on the provided drawings without creative effort.

[0015] Figure 1 This is a flowchart of the object detection method based on spatiotemporal gaps provided in the embodiments of this application.

[0016] Figure 2 This is an example diagram of obtaining a spatiotemporal porosity map provided in an embodiment of this application.

[0017] Figure 3 This is an example diagram illustrating the detection performance of the object detection method based on spatiotemporal gaps provided in the embodiments of this application.

[0018] Figure 4 This is an architectural diagram of the electronic device provided in the embodiments of this application.

[0019] The following detailed description, in conjunction with the accompanying drawings, will further illustrate this application. Detailed Implementation

[0020] To better understand the above-mentioned objectives, features, and advantages of this application, the application will be described in detail below with reference to the accompanying drawings and specific embodiments. It should be noted that, unless otherwise specified, the embodiments and features described herein can be combined with each other.

[0021] Numerous specific details are set forth in the following description to provide a thorough understanding of this application. The described embodiments are merely some, not all, of the embodiments described herein. All other embodiments obtained by those skilled in the art based on the embodiments of this application without inventive effort are within the scope of protection of this application.

[0022] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs. The terminology used herein is for the purpose of describing particular embodiments only and is not intended to be limiting of the application.

[0023] In one embodiment, active sonar is an underwater acoustic device commonly used in underwater surveillance, port protection, and other fields. It detects underwater objects, such as divers, unmanned underwater vehicles, and other intruding targets, by actively transmitting sound wave signals, thereby protecting important military and commercial equipment. However, underwater topography, sediment, fish, surface waves, and other clutter scatterers in shallow water environments can produce high-intensity wave reverberation in the backscattering of active sonar signals. This makes it difficult to distinguish the movement trajectories of clutter scatterers from those of intruding targets in the active sonar echo map, resulting in a decrease in the accuracy of active sonar in detecting underwater objects.

[0024] Existing active sonar methods for underwater moving target detection can be categorized into three types: spatial domain-based methods, temporal domain-based methods, and spatiotemporal domain-based methods. Spatial domain-based methods include threshold-based detection methods and traditional CFAR methods; temporal domain-based methods include matrix factorization-based methods, higher-order temporal slot density methods, and clutter map-based CFAR methods; spatiotemporal domain-based methods include optical flow-based moving target detection and matrix factorization methods based on multi-frame correlation.

[0025] For example, the matrix factorization-based method described above models stable reverberation as a low-rank background matrix and moving targets as a sparse target matrix. It then decomposes the data into a low-rank background matrix and a sparse target matrix through low-rank sparse decomposition. Specifically, multiple echo frames are first vectorized and stacked into a matrix, and then iteratively decomposed using a convex optimization algorithm to obtain the low-rank background matrix and the sparse target matrix. However, the matrix factorization-based method requires multiple iterations to obtain the optimal low-rank background matrix and sparse moving target matrix. Furthermore, each iteration requires singular value decomposition (SVD) of the data matrix, which has a high time complexity and is unsuitable for real-time underwater monitoring on underwater computing platforms.

[0026] Furthermore, the aforementioned high-order temporal lacunarity method utilizes lacunarity to describe the characteristics of echo intensity variation in the time domain. The high-order temporal lacunarity method calculates a lacunarity map based on echo maps of multiple frames, with each value in the lacunarity map obtained by calculating the lacunarity of the corresponding temporal sequence. The high-order temporal lacunarity method is simple and effective for extracting the motion patterns of targets. Lacunarity is a specialized term in fractal geometry, used to measure how a pattern fills space and represent the deviation from fractal and translational invariance. Lacunarity is highly scale-dependent and can be considered a scale-dependent measure of object texture or pattern. However, the high-order temporal lacunarity method only considers the characteristic differences between the moving target and reverberation in the temporal dimension, i.e., the differences in motion patterns, lacking a description of the target and reverberation in the spatial dimension, ignoring the texture features representing the spatial dimension.

[0027] To address the aforementioned issues, the object detection method based on spatiotemporal porosity provided in this application obtains multi-frame echo maps with a time series, constructs an echo pyramid corresponding to each frame's echo map to obtain multi-scale observations in the spatial dimension of each frame's echo map, and then, based on the echo pyramid, performs observations at the same spatial scale along the time dimension to obtain a porosity map under multi-scale observations. This method can solve the problem that single-scale observations cannot effectively distinguish between target objects and reverberation by constructing an echo pyramid. Furthermore, the echo pyramid also reduces the computational complexity of porosity features, thus reducing the computational load of multi-scale observations. In addition, extending porosity from the time dimension to the spatiotemporal dimension improves the accuracy of object detection results based on spatiotemporal porosity.

[0028] See Figure 1 The diagram shown is a flowchart of a preferred embodiment of the object detection method based on spatiotemporal gaps of this application.

[0029] In this embodiment, the object detection method based on spatiotemporal porosity can be applied to electronic devices (e.g., Figure 4The electronic device 3 shown integrates the object detection function based on spatiotemporal gap provided by the method of the present application embodiment, or runs in the electronic device in the form of a software development kit (SDK). The electronic device may be a computer, server, laptop, or other device.

[0030] like Figure 1 As shown, the object detection method based on spatiotemporal gaps specifically includes the following steps. Depending on different needs, the order of the steps in this flowchart can be changed, and some steps can be omitted.

[0031] Step S1: Obtain multiple raw echo images of the target area, with each raw echo image corresponding to a time label.

[0032] In one embodiment, the target area includes any area determined by the user that requires object detection, such as a body of water that requires underwater surveillance, or the waters of a port that requires port protection.

[0033] In one embodiment, an active sonar device (e.g., a multibeam active sonar) can be installed in the target area, and the device parameters and detection parameters of the active sonar device can be set. For example, the multibeam active sonar is deployed in the target area at a distance of 1.5m from the bottom of the water, and the multibeam active sonar is fixed by a steel frame installed in the water; the active sonar is set to transmit a linear frequency modulated (LFM) acoustic signal with a frequency of 500kHz and a bandwidth of 12kHz; the sonar detection range is set to 100m and the horizontal opening angle is set to 150 degrees.

[0034] In one embodiment, an active sonar device actively transmits acoustic signals to detect the target area and generates the original echo map based on the echoes from objects in the target area. To obtain the motion trajectory of objects in the target area, multiple original echo maps can be acquired sequentially, for example, 30 original echo maps. Each original echo map has the same size and resolution.

[0035] In one embodiment, a time tag can be set for each original echo image according to the chronological order of acquisition, resulting in a time series composed of multiple time tags corresponding to the multiple original images.

[0036] In one embodiment, such as Figure 2 The image shown is an example diagram of obtaining a spatiotemporal porosity map according to an embodiment of this application. Figure 2Figure (a) in the figure is an example of multiple original echo images provided in the embodiments of this application. The time label for the first original echo image is 1, and the time label for the 30th original echo image is 30. It can be seen that the original echo images roughly show multiple motion trajectories of various objects. However, due to wave reverberation caused by underwater topography, sediment, fish, surface waves, and other clutter scattering bodies, it is impossible to determine whether the motion trajectories in the original echo images belong to underwater intruders. Therefore, it is necessary to analyze the porosity in the original echo images to determine the texture features of the motion trajectories (e.g., fractal and translational invariance) through porosity, thereby distinguishing whether any motion trajectories in the original echo images belong to underwater intruders.

[0037] Step S2: Construct an echo pyramid corresponding to each original echo image. The echo pyramid contains multiple low-resolution echo images of the corresponding original echo image, and each image in the echo pyramid corresponds to a hierarchical label.

[0038] In one embodiment, constructing the echo pyramid corresponding to each original echo image includes: performing a preset number of resolution reduction processes on each original echo image to obtain multiple low-resolution echo images of each original echo image at multiple scales; and constructing the echo pyramid using each original echo image and the corresponding multiple low-resolution echo images.

[0039] In one embodiment, the resolution reduction process includes low-pass filtering and downsampling.

[0040] The low-pass filtering process can be performed using various low-pass filters (such as Gaussian filters, Butterworth low-pass filters, etc.). Taking Gaussian filtering as an example, the resolution reduction process performed on each original echo image a preset number of times includes:

[0041] The original echo image I0 was processed by Gaussian filtering to obtain a low-resolution echo image I1.

[0042] Using a Gaussian filter to analyze low-resolution echo maps I l-1 Gaussian filtering is performed to obtain a low-resolution echo map I. l ,in,

[0043]

[0044] Where, l∈{1,2,3,…,L}, and L represents the preset number of times;

[0045] G(m,n) represents a Gaussian window of size (2c+1)×(2c+1) with standard deviation σ=c; * represents the convolution operation, I l(i,j) represents the low-resolution echo map I. l The echo at position (i,j), I l-1 Represents low-resolution echo map I l-1 The echo at the midpoint (2i-1-m, 2j-1-n).

[0046] In one embodiment, the Gaussian filter represents a two-dimensional Gaussian low-pass filter. Applying this Gaussian filter to an image (e.g., the original echo image) can suppress scale information (blurring) in the image. The standard deviation σ of the two-dimensional Gaussian low-pass filter includes the standard deviation σ of the distance dimension. r Standard deviation σ of azimuth dimension b The size and standard deviation of the Gaussian low-pass filter can be preset. For example, the size of the Gaussian low-pass filter can be 5×17, and the standard deviation can be σ. r =4 and σ b =1.

[0047] In one embodiment, the two-dimensional spatial filtering performed on the original echo image in the above method only applies spatial filtering to the original echo image to extract its spatial texture information, while the observation scale in the time dimension is only 1. The two-dimensional spatial filtering can be extended to three-dimensional spatiotemporal filtering to further obtain the observation window in the time dimension of the original echo image, acquiring motion information at different scales. Specifically, a spatiotemporal gap map can be obtained by performing three-dimensional spatiotemporal filtering on the original echo image and then calculating the time-domain gapiness. For example, multi-scale observation in the time dimension can be achieved by selecting Gaussian filters with different standard deviations and filter sizes, or by using any other arbitrary three-dimensional low-pass filter.

[0048] In one embodiment, the downsampling process includes downsampling both the distance and orientation dimensions simultaneously, and may also include downsampling only the distance dimension or only the orientation dimension. The downsampling process is used to reduce the resolution of an image (e.g., the original echo image) to generate a thumbnail of the image, thereby obtaining an image smaller in size and resolution than the original image. For example, the downsampling process may be sub-sampling.

[0049] In one embodiment, after the preset number of resolution reduction processes, multiple low-resolution echo images of each original echo image at multiple scales are obtained. The preset number of processes can be set in advance; for example, the preset number of processes can be set to 3.

[0050] In one embodiment, constructing the echo pyramid using each original echo image and corresponding multiple low-resolution echo images includes: using the original echo image I0 as the zeroth level of the echo pyramid, and using the low-resolution echo images I... l As the l-th level of the echo pyramid, an echo pyramid with a total of L+1 levels is constructed, and the level label corresponding to the l-th level image in the echo pyramid is l.

[0051] In one embodiment, the preset number of times is equal to the total number of levels of the echo pyramid minus one. For example, if the preset number of times is 3, then the total number of levels of the echo pyramid is 4.

[0052] For example Figure 2 As shown, Figure 2 Figure (b) is an example diagram of the echo pyramid provided in the embodiments of this application, wherein each original echo image corresponds to an echo pyramid, combined with... Figure 2 Figure (a) shows the original echo. Figure 1 Corresponding to echo pyramid 1, the original echo image 30 corresponds to echo pyramid 30. The echo image at level 0 in each echo pyramid is the corresponding original echo image. It can be seen that the size of the echo images at level 0 and above in each echo pyramid decreases with each level. In addition, all echo pyramids have the same number of levels, and the size and resolution of the echo images corresponding to the same level label in all echo pyramids are consistent.

[0053] In one embodiment, decreasing the resolution of the echo map in the echo pyramid relative to the original echo map can significantly reduce the computational complexity of the multi-scale spatiotemporal gap method.

[0054] Step S3: Based on the time label, calculate the spatiotemporal gap between images corresponding to preset level labels in multiple echo pyramids to obtain a spatiotemporal gap map.

[0055] In one embodiment, the temporal porosity algorithm in active sonar converts a three-dimensional echo sequence H×W×T into a porosity map H×W with the same spatial size, where H, W, and T represent the azimuth number, range number, and sequence number of the echo sequence, respectively. Each value on the porosity map is calculated at the corresponding spatial location using a sliding window of size 1 over a time series of length T. This application extends the temporal porosity algorithm from the temporal domain to the spatiotemporal domain.

[0056] In one embodiment, calculating the spatiotemporal gap between images corresponding to preset level labels in multiple echo pyramids based on the time tags includes: using a preset sliding window to slide between images in multiple l-th levels of the multiple echo pyramids based on a sliding window algorithm and the order of the time tags, and calculating the spatiotemporal gap between images corresponding to the l-th level.

[0057] In one embodiment, the formula for calculating the spatiotemporal porosity includes:

[0058] Spatiotemporal porosity

[0059] Where, E(S)=∑SP(S,r) t ,r b ,r r E(S) represents the first moment of the sliding window mass distribution; E(S) 2 )=∑S 2 P(S,r t ,r b ,r r ,l), E(S 2 ) represents the second moment of the mass distribution of the sliding window; n(S,r t ,r b ,r r ,l) represents the mass S and time dimension r in the l-th stage. t The orientation dimension is r b And distance dimension r r The number of sliding windows, N(r) t ,r b ,r r (l) indicates that the image corresponding to level l uses a time dimension of r. t The orientation dimension is r b And distance dimension r r The number of sliding windows, P(S,r) t ,r b ,r r ,l) represents the probability that the sliding window of the l-th level has a quality of S, and k represents the higher-order moments.

[0060] Specifically, using a time dimension of r t The orientation dimension is r b And distance dimension r r Using a sliding window, traverse each pixel of the echo map in the l-th order, and calculate the k-th order temporal gap value Λ corresponding to the traversed pixel position within the sliding window. STP (r t ,r b ,r r ,l).

[0061] Furthermore, when acquiring the spatiotemporal gap map, the spatiotemporal gap value Λ within the sliding window corresponding to all pixel positions is... STP (r t ,r b ,r rThe high-order feature image constructed by l) is used as the spatiotemporal gap map.

[0062] For example Figure 2 As shown, Figure 2 Figure (c) is an example of a sliding window provided in an embodiment of this application, and a three-dimensional coordinate system composed of time, orientation, and distance is shown in the figure. Arrangement along the time direction is equivalent to arranging according to the chronological order of the time tags.

[0063] In one embodiment, each level of the echo pyramid can generate a spatiotemporal porosity map. In actual use, an appropriate observation scale (such as time scale, azimuth scale, distance scale, higher-order moments) can be selected according to the characteristics of the target object. That is, only one level (such as the second level) or several levels in the echo pyramid are calculated to obtain the spatiotemporal porosity map, so as to further reduce the amount of calculation.

[0064] Step S4: Determine the target object in the target region based on the spatiotemporal gap map.

[0065] In one embodiment, determining the target object in the target region based on the spatiotemporal gap map includes: processing the spatiotemporal gap map using a connected component analysis method to determine whether the spatiotemporal gap map contains a target connected component corresponding to the target object; if the spatiotemporal gap map contains the target connected component, determining the object at the target connected component as the target object.

[0066] Specifically, regions in the spatiotemporal gapity map with image intensities higher than a preset intensity threshold are defined as connected regions. It is then determined whether a target connected region with an image size (e.g., 20 pixels × 20 pixels) corresponding to the target object (e.g., a diver) exists is identified. The Otsu method can be used to perform an automatic threshold search on the spatiotemporal gapity map to determine the intensity threshold.

[0067] For example Figure 2 As shown, Figure 2 Figure (d) in the diagram is an example of a spatiotemporal porosity map provided in an embodiment of this application. Each echo map level can generate a spatiotemporal porosity map. Combined with... Figure 2 In Figure (c), the first-level spatiotemporal gap map in Figure (d) is obtained by sliding through the first-level echo map in the order of time labels.

[0068] In one embodiment, such as Figure 3The figure shown is an example of the detection performance of the object detection method based on spatiotemporal gaps provided in the embodiments of this application. The detection performance is described using the mean Average Precision (mAP) across all categories; a higher mAP value indicates better detection performance.

[0069] STP-Lac represents the object detection method based on spatiotemporal gapity of echo pyramids provided in the embodiments of this application, T-Lac represents higher-order temporal gapity, and ST-Lac represents the spatiotemporal gapity method that directly applies a spatial Gaussian filter to the original echo map.

[0070] In the three algorithms mentioned above, a sample set consisting of the same set of original echo maps is used to calculate the porosity map, and the observation scale of the time dimension is set to 1 and the higher-order moment k value is set to 4. The sample set includes 30 frames of original echo maps with time sequence, and the original echo maps contain known sample objects.

[0071] In the ST-Lac method, the Gaussian filter size is 13×49, and the standard deviation is σ. r =12 and σ b =3. The number of echo pyramid layers used in the STP-Lac method is set to 2, the Gaussian filter size for each layer is 5×17, and the standard deviation is σ. r =4 and σ b =1.

[0072] according to Figure 3 It can be seen that the STP-Lac method provided in this application embodiment has significantly improved detection performance compared with the other two algorithms. It has the highest object detection precision and the highest mAP value under the same recall rate (within the range of 0.1 to 0.6).

[0073] Furthermore, the processing time for the gapimetric algorithms of the three algorithms mentioned above was obtained. The processing time for the STP-Lac algorithm was 0.024 seconds, for the T-Lac algorithm it was 0.129 seconds, and for the ST-Lac algorithm it was 0.135 seconds. It can be seen that the STP-Lac algorithm is nearly five times faster than the T-Lac and ST-Lac algorithms, indicating a significant reduction in computational load.

[0074] In one embodiment, the object detection method based on spatiotemporal porosity provided in this application obtains multi-frame echo maps with time series, constructs an echo pyramid corresponding to each frame echo map to obtain multi-scale observations in the spatial dimension of each frame echo map; then, based on the echo pyramid, observations at the same spatial scale are performed along the time dimension to obtain a porosity map under multi-scale observations. This method can solve the problem that single-scale observations cannot effectively distinguish between target objects and reverberation by constructing an echo pyramid. Furthermore, the echo pyramid can reduce the computational complexity of porosity features and the computational load of multi-scale observations. In addition, extending porosity from the time dimension to the spatiotemporal dimension improves the accuracy of object detection results based on spatiotemporal porosity.

[0075] The above Figure 1 This paper details the object detection method based on spatiotemporal porosity proposed in this application. The following section will combine... Figure 4 This paper introduces the functional modules of the software system for implementing the object detection method based on spatiotemporal gaps, as well as the hardware device architecture for implementing the object detection method based on spatiotemporal gaps.

[0076] It should be understood that the embodiments described are for illustrative purposes only and are not limited to this structure in the scope of the patent application.

[0077] See Figure 4 The diagram shown is a structural schematic of an electronic device provided in a preferred embodiment of this application.

[0078] In a preferred embodiment of this application, the electronic device 3 includes a memory 31 and at least one processor 32. Those skilled in the art should understand that... Figure 4 The structure of the electronic device shown does not constitute a limitation of the embodiments of this application. It can be a bus structure or a star structure. The electronic device 3 may also include more or fewer other hardware or software than shown, or different component arrangements.

[0079] In some embodiments, the electronic device 3 includes a terminal capable of automatically performing numerical calculations and / or information processing according to pre-set or stored instructions, the hardware of which includes, but is not limited to, microprocessors, application-specific integrated circuits, programmable gate arrays, digital processors, and embedded devices.

[0080] It should be noted that the electronic device 3 is merely an example. Other existing or future electronic products that are suitable for this application should also be included within the scope of protection of this application and are incorporated herein by reference.

[0081] In some embodiments, the memory 31 is used to store program code and various data. For example, the memory 31 can be used to store a spatiotemporal gap-based object detection system 30 installed in the electronic device 3, and to achieve high-speed, automatic access to programs or data during the operation of the electronic device 3. The memory 31 includes read-only memory (ROM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), one-time programmable read-only memory (OTPROM), electrically-erasable programmable read-only memory (EEPROM), compact disc read-only memory (CD-ROM) or other optical disc storage, disk storage, magnetic tape storage, or any other computer-readable storage medium capable of carrying or storing data.

[0082] In some embodiments, the at least one processor 32 may be composed of integrated circuits, such as a single-packaged integrated circuit or multiple integrated circuits packaged with the same or different functions, including combinations of one or more central processing units (CPUs), microprocessors, digital processing chips, graphics processors, and various control chips. The at least one processor 32 is the control unit of the electronic device 3, connecting various components of the entire electronic device 3 via various interfaces and lines. It executes programs or modules stored in the memory 31 and calls data stored in the memory 31 to perform various functions of the electronic device 3 and process data, such as executing... Figure 1 The object detection function based on spatiotemporal gaps is shown.

[0083] In some embodiments, the object detection system 30 based on spatiotemporal gapity operates in an electronic device 3. The object detection system 30 based on spatiotemporal gapity may include multiple functional modules composed of program code segments. The program code of each program segment in the object detection system 30 based on spatiotemporal gapity may be stored in the memory 31 of the electronic device 3 and executed by at least one processor 32 to achieve... Figure 1The object detection function based on spatiotemporal gaps is shown.

[0084] In this embodiment, the object detection system 30 based on spatiotemporal gaps can be divided into multiple functional modules according to the functions it performs. A module, as referred to in this application, is a series of computer program segments that can be executed by at least one processor and perform a fixed function, and is stored in memory.

[0085] Although not shown, the electronic device 3 may also include a power supply (such as a battery) to power the various components. Preferably, the power supply can be logically connected to the at least one processor 32 via a power management device, thereby enabling functions such as charging, discharging, and power consumption management. The power supply may also include one or more DC or AC power supplies, recharging devices, power failure testing circuits, power converters or inverters, power status indicators, and other arbitrary components. The electronic device 3 may also include various sensors, Bluetooth modules, Wi-Fi modules, etc., which will not be described in detail here.

[0086] It should be understood that the embodiments described are for illustrative purposes only and are not limited to this structure in the scope of the patent application.

[0087] The integrated unit implemented as a software functional module described above can be stored in a computer-readable storage medium. This software functional module, stored in a storage medium, includes several instructions to cause an electronic device (which may be a server, personal computer, etc.) or processor to execute portions of the methods described in the various embodiments of this application.

[0088] The memory 31 stores program code, and the at least one processor 32 can call the program code stored in the memory 31 to execute related functions. The program code stored in the memory 31 can be executed by the at least one processor 32 to realize the functions of each module to achieve the purpose of object detection based on spatiotemporal gaps.

[0089] In the several embodiments provided in this application, it should be understood that the disclosed apparatus and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of modules is only a logical functional division, and other division methods may be used in actual implementation.

[0090] The modules described as separate components may or may not be physically separate. The components shown as modules may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs.

[0091] Furthermore, the functional modules in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or in the form of hardware plus software functional modules.

[0092] It will be apparent to those skilled in the art that this application is not limited to the details of the exemplary embodiments described above, and that this application can be implemented in other specific forms without departing from the spirit or essential characteristics of this application. Therefore, the embodiments should be considered exemplary and non-limiting in all respects, and the scope of this application is defined by the appended claims rather than the foregoing description. Thus, all variations falling within the meaning and scope of equivalents of the claims are intended to be embraced within this application. No reference numerals in the claims should be construed as limiting the scope of the claims. Furthermore, it is clear that the word "comprising" does not exclude other elements or, and the singular does not exclude the plural. Multiple elements or devices recited in the apparatus claims may also be implemented by a single element or device in software or hardware. The terms "first," "second," etc., are used to indicate names and do not indicate any particular order.

[0093] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of this application and are not intended to limit it. Although this application has been described in detail with reference to the above preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of this application without departing from the spirit and scope of the technical solutions of this application.

Claims

1. A method for object detection based on spatiotemporal voidness, characterized in that, The method comprises: acquiring a plurality of original echo images of a target region, each original echo image corresponding to a time label; constructing an echo pyramid corresponding to each original echo image, the echo pyramid comprising a plurality of low-resolution echo images of the corresponding original echo image, each image in the echo pyramid corresponding to a level label; Based on the time label, the spatiotemporal gap degree between images corresponding to a preset level label in the plurality of echo pyramids is calculated to obtain a spatiotemporal gap degree map; the calculation of the spatiotemporal gap degree between images corresponding to a preset level label in the plurality of echo pyramids based on the time label comprises: based on a sliding window algorithm and the order of the time labels, a preset sliding window is used to slide between images in the plurality of first level images in the plurality of echo pyramids, and the spatiotemporal gap degree between images corresponding to the first level is calculated; wherein the calculation formula of the spatiotemporal gap degree comprises: spatiotemporal voids ; wherein, , denotes the first moment of the sliding window quality distribution; , denotes the second moment of the sliding window quality distribution; , denotes the number of sliding windows of order with quality , temporal size , azimuth size and range size , denotes the number of sliding windows of order corresponding to the image with temporal size , azimuth size and range size , denotes the probability of the sliding window of order corresponding to the image with quality , denotes the higher order moment; determining a target object in the target region according to the spatiotemporal voidness graph.

2. The spatiotemporal voidness-based object detection method of claim 1, wherein, The construction of the echo pyramid corresponding to each original echo image comprises: performing a preset number of resolution reduction processes on each original echo image to obtain a plurality of multi-scale low-resolution echo images of each original echo image; constructing the echo pyramid using each original echo image and the corresponding plurality of low-resolution echo images.

3. The spatiotemporal voidness-based object detection method of claim 2, wherein, The resolution reduction process comprises a low-pass filtering process and a down-sampling process.

4. The spatiotemporal void fraction based object detection method of claim 2, wherein, The performing of the preset number of resolution reduction processes on each original echo image comprises: Gaussian filter is used to the original echo image Gaussian filter processing is performed to obtain a low-resolution echo image ; applying a Gaussian filter to the low resolution echo map performing a Gaussian filter process to obtain a low resolution echo map wherein, ， wherein, , denotes the preset number of times; , denotes a Gaussian window of size and standard deviation ; denotes a convolution operation, denotes an echo at position in the low-resolution echo map ; denotes an echo at position in the low-resolution echo map .

5. The spatiotemporal voidness-based object detection method of claim 4, wherein, The construction of the echo pyramid using each original echo image and the corresponding plurality of low-resolution echo images comprises: constructing a low-resolution echo map as a zeroth level of the echo pyramid constructing a low-resolution echo map as a zeroth level of the echo pyramid constructing a low-resolution echo map as a zeroth level of the echo pyramid constructing a low-resolution echo map as a zeroth level of the echo pyramid constructing a low-resolution echo map as a zeroth level of the echo pyramid constructing a low-resolution echo map as a zeroth level of the echo pyramid constructing a low-resolution echo map as a zeroth level of the echo pyramid 6. An electronic device, comprising: The electronic device comprises a processor and a memory, and the processor is configured to execute a computer program stored in the memory to implement the spatiotemporal voidness-based object detection method according to any one of claims 1 to 5.

7. A computer-readable storage medium having stored thereon a computer program, characterized in that The computer program is configured to be executed by the processor to implement the spatiotemporal voidness-based object detection method according to any one of claims 1 to 5.

Images