CN111144318B - Noise reduction method for point cloud data of underwater sonar system - Google Patents
Noise reduction method for point cloud data of underwater sonar system Download PDFInfo
- Publication number
- CN111144318B CN111144318B CN201911380561.8A CN201911380561A CN111144318B CN 111144318 B CN111144318 B CN 111144318B CN 201911380561 A CN201911380561 A CN 201911380561A CN 111144318 B CN111144318 B CN 111144318B
- Authority
- CN
- China
- Prior art keywords
- point cloud
- cloud data
- data
- noise
- noise reduction
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2218/00—Aspects of pattern recognition specially adapted for signal processing
- G06F2218/02—Preprocessing
- G06F2218/04—Denoising
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02A—TECHNOLOGIES FOR ADAPTATION TO CLIMATE CHANGE
- Y02A90/00—Technologies having an indirect contribution to adaptation to climate change
- Y02A90/30—Assessment of water resources
Landscapes
- Measurement Of Velocity Or Position Using Acoustic Or Ultrasonic Waves (AREA)
Abstract
The invention discloses a point cloud data noise reduction method of an underwater sonar system, which comprises the following steps of: (1) Based on the underwater sound signal reflection principle, discrete data points without connected areas in the point cloud data are removed; (2) sidelobe effect suppression is performed on the point cloud data; (3) Removing noise based on the correlation of continuous multi-frame point cloud data; (4) Dynamically normalizing the point cloud data to process and remove noise data and background data; (5) And removing point cloud data which are not physically meaningful according to the actual physical meaning of the three-dimensional point cloud data aiming at the point cloud data of the three-dimensional data structure. Different noise reduction strategies are adopted for different noises respectively so as to realize comprehensive noise reduction on the point number data, and the noise reduction effect is improved.
Description
Technical Field
The invention belongs to the technical field of signal processing, and particularly relates to a point cloud data noise reduction method of an underwater sonar system.
Background
The underwater sonar imaging is based on the sonar point cloud data, so that the processing quality of the point cloud data directly determines the underwater sonar imaging and a series of subsequent sonar data processing processes, including extracting useful target information from the point cloud data, detecting objects in water and the like. Because the ultrasonic signal in water can generate specific underwater sound effects including side lobe effect, reverberation effect and the like in principle, and can generate certain noise signals in the processes of signal acquisition and signal transmission, the problem of data noise can not be thoroughly solved from the root. The noise interference brings certain difficulties to algorithms such as target detection, identification, tracking and the like, and the data preprocessing module needs to perform certain noise reduction processing on the point cloud data, so that the subsequent process can be greatly facilitated.
The patent application with the application publication number of CN105785349A discloses a noise removal method of a phased array three-dimensional acoustic camera sonar, and the noise removal method uniformly removes noise, and does not perform classified denoising, so that the denoising effect is poor.
The patent application with the application publication number of CN109035224A discloses a submarine pipeline detection and three-dimensional reconstruction method based on multi-beam point clouds, in the submarine pipeline detection and three-dimensional reconstruction method, a point cloud denoising filtering method based on density analysis is adopted to denoise point cloud data of a pipeline, and the denoising effect of the single denoising method is poor.
Disclosure of Invention
The invention aims to provide a point cloud data denoising method of an underwater sonar system, which adopts different denoising strategies for different noises respectively so as to comprehensively denoise point data and improve the denoising effect.
In order to achieve the purpose of the invention, the invention provides the following technical scheme:
a point cloud data noise reduction method of an underwater sonar system comprises the following steps:
(1) Based on the underwater sound signal reflection principle, discrete data points without connected areas in the point cloud data are removed;
(2) Side lobe effect suppression is carried out on the point cloud data;
(3) Removing noise based on the correlation of continuous multi-frame point cloud data;
(4) Dynamically normalizing the point cloud data to process and remove noise data and background data;
(5) And removing point cloud data which are not physically meaningful according to the actual physical meaning of the three-dimensional point cloud data aiming at the point cloud data of the three-dimensional data structure.
Compared with the prior art, the invention has the following beneficial technical effects:
(1) The point cloud data denoising method for the underwater sonar system can remarkably improve signal effectiveness, and can effectively inhibit noise signals in data, so that the denoising method can provide effective data support for a subsequent data processing flow and has important significance for the underwater sonar system.
(2) The noise reduction method for the point cloud data of the underwater sonar system is simple to realize, the existing simple filtering technology is used, noise reduction processing of specific scenes is conducted according to noise characteristics in the point cloud data of the underwater sonar system, the flow of the underwater sonar system and the like, and the method is high in pertinence and good in practical effect.
(3) The noise reduction method for the point cloud data of the underwater sonar system can be further finely adjusted according to the actual application scene so as to improve the noise removal effect of the point cloud data in the actual application scene.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings that are required in the embodiments or the description of the prior art will be briefly described, and it is obvious that the drawings in the following description are only some embodiments of the present invention, and other drawings may be obtained according to these drawings without inventive effort for a person skilled in the art.
FIG. 1 is a flow chart of the steps of processing data by an underwater sonar system provided by an embodiment;
fig. 2 is a schematic diagram of classification of point cloud data noise of an underwater sonar system according to an embodiment;
FIG. 3 is a flowchart of a method for noise reduction of point cloud data of an underwater sonar system provided by an embodiment;
fig. 4 is a flowchart of a sidelobe effect suppression step in the point cloud data denoising method of the underwater sonar system provided by the embodiment;
fig. 5 is a flowchart of a dynamic normalization step in the point cloud data denoising method of the underwater sonar system according to the embodiment.
Detailed Description
The present invention will be described in further detail with reference to the drawings and examples, in order to make the objects, technical solutions and advantages of the present invention more apparent. It should be understood that the detailed description is presented by way of example only and is not intended to limit the scope of the invention.
FIG. 1 is a flow chart of the steps of processing data by an underwater sonar system provided by an embodiment. As shown in fig. 1, the main processing flow of the underwater sonar system is that an a/D chip collects an original signal, then performs beam forming processing to obtain a point cloud data type required to be processed by the noise reduction method, then performs noise reduction processing on the point cloud data by using the noise reduction method, and performs obstacle detection, target recognition or real-time image display on the processed data according to an actual application scene.
Fig. 2 is a schematic diagram of point cloud data noise classification of an underwater sonar system according to an embodiment. As shown in fig. 2, the main noise in the point cloud data of the underwater sonar system is: due to side lobe effect caused by acoustic principle, gaussian noise caused by line transmission, signal sampling and the like, signal reverberation caused by acoustic reflection diffraction and the like. Some of these signals are regular noise with a certain regularity, such as side lobe effects. Some are random noise, such as signal reflection, diffraction noise, transmission noise, which are randomly distributed.
Fig. 3 is a flowchart of a method for denoising point cloud data of an underwater sonar system according to an embodiment. As shown in fig. 3, the method for denoising the point cloud data of the underwater sonar system comprises the following steps:
s101, removing discrete data points without connected areas in the point cloud data based on the underwater sound signal reflection principle.
Based on the underwater acoustic signal reflection principle, certain diffraction phenomenon can be generated around the object after the underwater acoustic signal irradiates the target object, and based on the characteristic, points which have high single-point signal intensity but are not related to surrounding array element signals are considered as noise points and are removed. Most of discrete noise signals in the point cloud data can be removed by removing discrete points without connected areas, and an ideal primary denoising effect can be achieved through primary processing on underwater sonar point cloud signals.
The noise of point cloud data of a sonar system is mainly Gaussian noise, and on the basis of S101, the noise of the point cloud data is reduced by adopting a mean value filtering method, so that random noise points in the point cloud data are reduced. The method can perform a filtering operation on a small portion of noise still present in S101.
S102, side lobe effect suppression is conducted on point cloud data.
For the side lobe effect in the acoustic effect, according to the side lobe characteristics, the first side lobe value is far smaller than the main lobe value, the side lobe value is removed or reduced, so that noise pollution caused by the side lobe effect is removed, and almost all strong target data points have the side lobe effect, so that the step is necessary and relatively effective.
For target data, the data of the target main lobe are consistent in relative order, while for background non-target data, the data of the target main lobe are relatively low in order. The side lobe suppression can be effectively performed by removing all data points smaller than a certain threshold value of the main lobe signal in the same level. The side lobe is suppressed, and other useful data is not suppressed, so that the method has good practical operability.
Fig. 4 is a flowchart of a sidelobe effect suppression step in the point cloud data denoising method of the underwater sonar system according to the embodiment, where as shown in fig. 4, the sidelobe effect suppression step includes:
taking out point cloud data with the same physical distance according to the point cloud data collected by each layer, finding out the point cloud data with the maximum value from the point cloud data, traversing all the point cloud data of each layer by taking the preset percentage of the maximum value as a screening threshold, and removing the point cloud data exceeding the corresponding screening threshold of each layer to remove noise points caused by side lobe effect;
besides removing noise points caused by side effects, the sidelobe effect suppression step further comprises the following steps: and multiplying the point cloud data smaller than the maximum preset percentage by a certain coefficient to reduce noise caused by side lobe effect.
And S103, removing noise based on the correlation of the continuous multi-frame point cloud data.
Because the point cloud data with the target exist in continuous multiframes, and the data have relative randomness in the continuous multiframes for noise, the random noise without correlation can be removed by adopting a multiframe filtering method, and the specific process is as follows:
aiming at point cloud data with lower correlation in continuous multi-frame data, directly removing the point cloud data;
and removing noise points in the image by a median filtering method of the same position of the continuous multiframe aiming at point cloud data with higher relativity in the continuous multiframe data, namely replacing the data value of the same position of each frame by using the data average value of the same position of the continuous multiframe.
In an embodiment, point cloud data with occurrence times smaller than a preset frequency in each group of continuous n-frame data is considered as point cloud data with lower correlation, and the rest point cloud data are data with higher correlation, wherein the preset frequency is generally smaller than 1/2n.
When the data source is dynamic, namely when the collected sonar is of a motion type, namely when the detection sensor moves relatively to the target water area, the sonar data corresponding to the same target can be generated at different positions possibly due to the relative movement of the detection sensor in continuous multi-frame data, so that misjudgment can be caused on the target and noise, and therefore, the collected dynamic point cloud data is corrected according to the posture correction information of the detection sensor before noise is removed based on the correlation of the continuous multi-frame point cloud data. The sonar position offset is added into the sonar point cloud data through the attitude sensor, so that the dynamic point cloud data is corrected, and the data among multiple frames of data are guaranteed to have correlation comparability.
When the continuous multi-frame correlation processing is performed, for each group of continuous multi-frames, the data of the part, which is not overlapped with the previous frame after the offset value is added in the point cloud data, still remains the original data, because the data may be used in the correlation processing process of each subsequent group of continuous multi-frames.
Based on the sonar detection principle, the detection range can be hundreds of meters away, and the offset between continuous multiframes is often only a few meters, so that the overlapping rate of the point cloud data between the continuous multiframes is relatively high, and the noise removal availability based on the correlation of the point cloud data of the continuous multiframes is high.
S104, carrying out dynamic normalization on the point cloud data to process and remove noise data and background data.
Specifically, as shown in fig. 5, statistics of data intensity values is performed on the point cloud data after the processing of S101 to S103 is completed, a preset percentage is multiplied by a mean value of the statistical data intensity values to be used as a normalization screening threshold value, all the point cloud data are facilitated, and the point cloud data exceeding the normalization screening threshold value are removed, so that dynamic normalization operation is performed on the point cloud data to remove noise data and background data.
In the embodiment, through the dynamic normalization operation of the point cloud data, noise data with lower relative intensity values and background data without targets are further filtered, so that the target data in the point cloud data can be better highlighted.
S105, removing point cloud data which are not physically meaningful according to the actual physical meaning of the three-dimensional point cloud data aiming at the point cloud data of the three-dimensional data structure.
For point cloud data with a three-dimensional data structure, further processing can be performed according to the physical meaning of the point cloud data. Compared with two-dimensional point cloud data, the three-dimensional point cloud data is added with actual physical depth information, and points which are not physically meaningful can be removed based on the characteristics. In particular, point cloud data that exceeds the water surface and falls below the water bottom in the depth direction may be removed.
The point cloud data denoising method for the underwater sonar system can remarkably improve signal effectiveness, and can effectively inhibit noise signals in data, so that the denoising method can provide effective data support for a subsequent data processing flow and has important significance for the underwater sonar system.
The noise reduction method for the point cloud data of the underwater sonar system is simple to realize, the existing simple filtering technology is used, noise reduction processing of specific scenes is conducted according to noise characteristics in the point cloud data of the underwater sonar system, the flow of the underwater sonar system and the like, and the method is high in pertinence and good in practical effect.
According to the point cloud data noise reduction method of the underwater sonar system, the screening threshold value can be finely adjusted according to the actual application scene, so that the noise removal effect of the point cloud data in the actual application scene is improved.
The foregoing detailed description of the preferred embodiments and advantages of the invention will be appreciated that the foregoing description is merely illustrative of the presently preferred embodiments of the invention, and that no changes, additions, substitutions and equivalents of those embodiments are intended to be included within the scope of the invention.
Claims (4)
1. The point cloud data noise reduction method for the underwater sonar system is characterized by comprising the following steps of:
(1) Based on the underwater sound signal reflection principle, discrete data points without connected areas in the point cloud data are removed;
(2) Sidelobe effect suppression is performed on point cloud data, and the sidelobe effect suppression method comprises the following steps: taking out point cloud data with the same physical distance according to the point cloud data collected by each layer, finding out the point cloud data with the maximum value from the point cloud data, traversing all the point cloud data of each layer by taking the preset percentage of the maximum value as a screening threshold, removing the point cloud data exceeding the corresponding screening threshold of each layer to remove noise caused by a side lobe effect, and multiplying the point cloud data with the preset percentage of the maximum value by a certain coefficient to reduce the noise caused by the side lobe effect;
(3) Removing noise based on the correlation of continuous multi-frame point cloud data, and directly removing the point cloud data with lower correlation in the continuous multi-frame data; for point cloud data with higher relativity in continuous multi-frame data, removing noise points in the image by a median filtering method of the same position of the continuous multi-frame, namely replacing the data value of the same position of each frame by using the data average value of the same position of the continuous multi-frame;
when continuous multi-frame correlation processing is carried out, aiming at each group of continuous multi-frames, adding an offset value into the point cloud data, and keeping original data of a part which is not overlapped with the previous frame;
(4) Dynamically normalizing the point cloud data to process and remove noise data and background data;
(5) And removing point cloud data which are not physically meaningful according to the actual physical meaning of the three-dimensional point cloud data aiming at the point cloud data of the three-dimensional data structure.
2. The noise reduction method for the point cloud data of the underwater sonar system according to claim 1, wherein on the basis of the step (1), a mean value filtering method is adopted to perform noise reduction treatment on the point cloud data so as to reduce random noise points in the point cloud data.
3. The method for noise reduction of point cloud data of an underwater sonar system according to claim 1, wherein the dynamic point cloud data is corrected according to posture correction information of the detection sensor before noise is removed based on correlation of continuous multi-frame point cloud data with respect to the collected dynamic point cloud data.
4. The method for noise reduction of point cloud data of an underwater sonar system according to claim 1, wherein in the step (4), the statistical of the data intensity values is performed on the point cloud data processed in the steps (1) to (3), the average value of the statistical data intensity values is multiplied by a preset percentage to be used as a normalized screening threshold value, all the point cloud data are facilitated, and the point cloud data exceeding the normalized screening threshold value are removed, so that dynamic normalization operation is performed on the point cloud data to remove noise data and background data.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911380561.8A CN111144318B (en) | 2019-12-27 | 2019-12-27 | Noise reduction method for point cloud data of underwater sonar system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911380561.8A CN111144318B (en) | 2019-12-27 | 2019-12-27 | Noise reduction method for point cloud data of underwater sonar system |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN111144318A CN111144318A (en) | 2020-05-12 |
| CN111144318B true CN111144318B (en) | 2023-08-29 |
Family
ID=70521132
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201911380561.8A Active CN111144318B (en) | 2019-12-27 | 2019-12-27 | Noise reduction method for point cloud data of underwater sonar system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111144318B (en) |
Families Citing this family (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111681667B (en) * | 2020-06-23 | 2021-05-04 | 青岛科技大学 | Underwater sound signal denoising method based on adaptive window filtering and wavelet threshold optimization |
| WO2022186103A1 (en) * | 2021-03-02 | 2022-09-09 | パイオニア株式会社 | Information processing device, information processing method, program, and storage medium |
| CN117538881B (en) * | 2024-01-10 | 2024-05-07 | 海底鹰深海科技股份有限公司 | Sonar water imaging beam forming method, system, equipment and medium |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106643671A (en) * | 2016-12-01 | 2017-05-10 | 江苏省测绘工程院 | Underwater point cloud denoising method based on airborne LiDAR depth sounding system |
| CN109035224A (en) * | 2018-07-11 | 2018-12-18 | 哈尔滨工程大学 | A kind of Technique of Subsea Pipeline Inspection and three-dimensional rebuilding method based on multi-beam point cloud |
-
2019
- 2019-12-27 CN CN201911380561.8A patent/CN111144318B/en active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106643671A (en) * | 2016-12-01 | 2017-05-10 | 江苏省测绘工程院 | Underwater point cloud denoising method based on airborne LiDAR depth sounding system |
| CN109035224A (en) * | 2018-07-11 | 2018-12-18 | 哈尔滨工程大学 | A kind of Technique of Subsea Pipeline Inspection and three-dimensional rebuilding method based on multi-beam point cloud |
Non-Patent Citations (1)
| Title |
|---|
| 牛蕾.SAR图像舰船目标去旁瓣处理.《雷达科学与技术》.2018,第197-200页. * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN111144318A (en) | 2020-05-12 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111144318B (en) | Noise reduction method for point cloud data of underwater sonar system | |
| CN108776336B (en) | EMD-based adaptive through-wall radar static human body target positioning method | |
| CN104103041B (en) | Ultrasonoscopy mixed noise Adaptive Suppression method | |
| CN107789008B (en) | Self-adaptive ultrasonic beam synthesis method and system based on channel data | |
| CN108354627B (en) | Ultrasonic beam forming method for improving frame frequency | |
| CN112750090A (en) | Underwater image denoising method and system for improving wavelet threshold | |
| CN115453497A (en) | Filtering method and system for low signal-to-noise ratio sparse radar signal | |
| Zhang et al. | A reverberation noise suppression method of sonar image based on shearlet transform | |
| CN104166975A (en) | Low-altitude infrared target detection algorithm in complex environment | |
| CN111766306A (en) | Full-focus imaging method and device | |
| CN116452439A (en) | Noise reduction method and device for laser radar point cloud intensity image | |
| CN116612032A (en) | Sonar image denoising method and device based on self-adaptive wiener filtering and 2D-VMD | |
| CN105487056A (en) | S-wave band Doppler radar breaking wave interference suppression method and device | |
| CN114170239A (en) | Large-view-field self-adaptive wall filtering method and system based on hierarchical search | |
| Xiao et al. | Underwater image classification based on image enhancement and information quality evaluation | |
| CN101396278A (en) | Method of removing real-time ultrasound pattern speckle noise | |
| CN103584886B (en) | Self-adaption apodization method based on phase coherent information | |
| CN113625286A (en) | Strong robustness truncation coherence coefficient ultrasonic beam forming method based on coherence features | |
| JP3366872B2 (en) | Ultrasound diagnostic equipment | |
| CN112634222A (en) | SAR image ship target detection method | |
| CN112147226A (en) | Optimal decomposition layer number determining method based on laser ultrasonic signal wavelet denoising | |
| CN112230200A (en) | Improved combined noise reduction method based on laser radar echo signals | |
| CN117607876B (en) | Method and system for detecting passive sonar multi-beam narrowband signals | |
| CN117726561B (en) | Intravascular ultrasound image processing method, related device and storage medium | |
| CN113379906B (en) | Single-photon three-dimensional image reconstruction method and system based on super-pixel filtering |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |