71 results about "Sonar imagery" patented technology

A vessel-mountable integrated sonar system is provided. The vessel-mountable integrated sonar system comprises at least one imaging sonar data acquisition device and at least one processing system electronically and removably connected to the at least one imaging sonar data acquisition device, wherein the sonar data acquisition device preferably provides acoustic data to the processing system, producing sonar imageries utilizing the acoustic data, and wherein the system provides digital tilt and azimuth direction feedback for accurate geo-referencing of data to localize targets of interest.

The invention provides a deep learning-based adaptive weight convolutional neural network underwater sonar image classification method. The method comprises the steps of (1) according to characteristics of underwater sonar images in a data set, generating a DBN two-dimensional parameter matrix; (2) adaptively adjusting distribution of a CNN filter weight matrix; and (3) realizing deep learning-based adaptive weight CNN underwater sonar image classification. According to the deep learning-based adaptive weight convolutional neural network underwater sonar image classification method provided bythe invention, the problem of randomness of filter weight initialization in a CNN can be solved, so that the situation of falling into local optimum is avoided, the classification correctness can bebetter improved, and certain validity is achieved.

A sonar system is provided including a display screen, a transducer assembly including at least one transducer that is configured to emit one or more sonar signals into an underwater environment and receive sonar return data reflected from one or more objects, and a sonar module configured to generate a 3D matrix based on the sonar return data including a plurality of sonar returns that are each defined by a 3D positional value, determine a group of the plurality of sonar returns associated with an object, assign a predetermined icon to the determined group, and generate and display 3D image of the sonar return data. The predetermined icon is positioned within the 3D image at a position that corresponds to the position of the determined group such that the position of the predetermined icon corresponds to the position of the object.

The invention provides a side-scan sonar image feature extraction method based on a full convolutional neural network. The method comprises the following steps: carrying out the data augmentation through an original sonar image, and set needed by model training and testing; manually labeling the edge area of the submarine topography of each image in the sample set, distinguishing a target and a background, and obtaining a model training and testing label graph; an FCNs model is constructed; inputting the submarine topographic map and the corresponding label map into a network, training the network by adopting a small-batch gradient descent method of a driving quantity item, and storing an optimal network model; comparing convergence and stability of the network under a random gradient descent method and a small-batch gradient descent method; and extracting terrain edge contour features, outputting a feature extraction result, and carrying out qualitative evaluation on the result. According to the method, complex preprocessing is not needed, and the sonar feature extraction method is high in speed, high in efficiency and high in speckle noise resistance; the performance of the network is improved, and the convergence and stability of each network model of the FCNs are ensured.

The invention provides an underwater sonar image unsupervised classification method based on class consciousness domain self-adaptation. The method comprises the steps of (1) using a generative adversarial network for constructing and generating a data set; (2) proposing an improved method based on an adversarial auto-encoder to construct a domain-adaptive source domain; (3) proposing to constructa domain-adaptive target domain based on an adversarial learning method; and (4) training a target domain, and completing unsupervised classification of the underwater sonar images on the balanced native data set and the non-balanced native data set. According to the invention, two GANs of CGAN and DCGAN are used to generate images so as to construct an underwater sonar image generation data set,and an unsupervised domain adaptive method is introduced into unsupervised classification of underwater sonar images according to the situation of label missing. And a balanced underwater sonar imagenative data set and an unbalanced underwater sonar image native data set are simultaneously constructed to verify the adaptability of the method provided by the invention.

Sonar imagery captured during a trip on a body of water can be played back to users, forming a trip replay. However, since the depth of the body of water varies drastically over the course of the trip, the resulting sonar image captured and played over the trip replay may result in loss of detail near shallow depths (even though such detail was captured and is available). Embodiments of the present invention seek to provide the ability to zoom in on portions of the sonar image during trip replay. Additionally, further zoom features, such as bottom lock and canopy lock, provide additional functionality.

The invention relates to an underwater target tracking method based on a forward-looking sonar image, and the method can accurately track the position of an underwater target, and includes carrying out the real-time quick detection of the underwater target, and completing the installation and debugging of a GPS, an attitude sensor, a holder, a PC and forward-looking sonar equipment; controlling the pan-tilt to move in a detection range through the PC, and giving a tracking target; acquiring information data; recovering the sonar information into a gray level image of the sonar according to theintensity of the forward-looking sonar callback signal; preprocessing the images, extracting two groups of feature points to obtain two groups of feature point sets, matching the feature point sets to separate correct matching from error matching, and calculating a target image, transformation coordinates and a transformation matrix; and performing repeated matching and repeated transformation toobtain a transformation matrix of the current target position, thereby realizing underwater target tracking based on forward-looking sonar image matching. According to the underwater target trackingmethod, the position of the target is accurately tracked, the detection speed is high, and the method has the characteristic of real-time performance.

A method of tracking a known object is presented, wherein a sonar image of an object which is distorted by an artifact associated with sonar imaging is compared with an image generated from a model of the object, and at least one of the two images is modified to reduce differences between them.

The invention discloses an underwater fishing net detection method based on a forward-looking sonar image and an AUV platform. The method comprises the following steps: collecting and acquiring a forward-looking sonar original data packet through a multi-beam forward-looking sonar carried on an AUV; analyzing the obtained sonar original data packet in real time to obtain an original sonar data matrix, and then interpolating sonar data to obtain a complete fan-shaped forward-looking sonar image; transmitting the obtained forward-looking sonar image to the AUV platform, and inputting the forward-looking sonar image into a target detection model trained in advance to carry out fishing net detection in real time; and converting the detection result into a corresponding system instruction and feeding back to an AUV main control module, and allowing the AUV to make a corresponding obstacle avoidance decision according to the detection result. According to the scheme, the problems of complexalgorithm implementation, low detection accuracy, poor real-time performance and the like in the prior art are solved, the autonomous survival ability of the AUV is improved, the AUV can master the information of the sea area ahead in real time, and an underwater fishing net can be autonomously detected and avoided.

It is an object to maintain a processing speed while increasing accuracy of a sonar image. Disclosed is a phased-array synthetic-aperture sonar system, including: a plurality of phased array processors (9) which perform, in parallel, phased array processing for a plurality of incoming data; a phased array distribution device (8) which distributes each simultaneously incoming data to the plurality of phased array processors (9), the incoming data being input from a plurality of receivers disposed in a platform; a synthetic aperture processor (11) which performs synthetic aperture processing by using a plurality of phased array processing results; and a data-shaping buffer device (10) which transfers the phased array processing results output from the plurality of phased array processors (9) to the synthetic aperture processor (11) in the order of receipt of the incoming data.

The invention discloses a submarine topography data verification method based on more than three pieces of uniformly distributed multi-beam data. The method comprises the following steps: transmittingand receiving wave beams through a transducer array based on an acoustic system; performing beamforming and conversion between an electric signal and a sound signal based on an acquisition system, and performing the classified storage of acquired water depth information, time information and sound intensity information as much as possible; based on an auxiliary sensor, combining and processing various sensor data used for navigation positioning and attitude calibration sound velocity tracking measurement; based on a data processing system, summarizing the measured data; integrating the data to drawing multi-beam sonar images reflecting the submarine topography and landform; comparing water depth data, sound intensity data and water column data in a measured area with measured data in a traditional mode based on a multi-beam wave sounding system, and then carrying out the regularization comparison processing on the traditional measured data, thereby improving the measurement precisionof the submarine topography.

The invention discloses a target guide approaching control method and system for an underwater autonomous vehicle, relates to the technical field of underwater autonomous vehicle control, and the method and the system can improve the online recognition capability of the vehicle, perform task re-planning after finding a suspected target, and ensure the underwater autonomous path planning and target guide approaching functions of the underwater vehicle. According to the technical scheme of the invention, the method and the system are used for achieving the recognition, processing and guided approaching of an underwater target by the autonomous underwater vehicle. The method comprises the following steps: the autonomous underwater vehicle autonomously sails underwater according to a pre-planned route; a forward-looking sonar detects the navigation area of the autonomous underwater vehicle to obtain a sonar image; and the image processor performs online processing on the sonar image and outputs the position information of the underwater target to the control unit in real time; the control unit carries out task re-planning according to the position information of the underwater target, and the autonomous underwater vehicle sails according to an air route obtained through re-planning and is guided to approach the target.

The invention discloses an underwater acoustic dataset augmentation method and system based on a conditional generative adversarial network, and the method comprises the steps: 1, obtaining the slices of a target object in a sonar dataset and a remote sensing dataset, carrying out the contour enhancement and pixel filling of a target object image in the slices, retaining the class information of the target object in the qualified slices, and carrying out the recognition of the target object; classifying and storing as samples; step 2, through a conditional generative adversarial network, augmenting the samples which do not reach a preset number threshold in the step 1; and step 3, pasting the samples reaching a preset number in the step 1 and the samples augmented in the step 2 to the seabed reverberation background image in the sonar data set, performing optimization processing on the pasted image, simulating the image into a sonar image, and forming the sonar data set. According to the method, the underwater acoustic image data set is enriched, and the development and application of the deep learning method in underwater target detection and segmentation tasks are promoted on the basis, so that the accuracy of environmental perception target detection is improved.

The embodiment of the invention provides a water bottom position detection method and device, and the method comprises the steps: generating a water bottom salient image of a target water area based on a water bottom sonar image through a visual attention mechanism after obtaining the water bottom sonar image of the target water area; and determining real water bottom position information of the target water area based on the water bottom salient image by using a preset water bottom detection method. A visual attention mechanism is utilized to perform salient feature extraction on the water bottom sonar image to obtain the water bottom salient image, and the water bottom salient image can represent salient contrast features between the water bottom and a non-water body object in the water area, so that preliminary area delineation can be performed on the position of the water bottom based on the water bottom salient image, the pre-estimated position information of the water bottom is determined, and then the position of the water bottom is accurately positioned based on the pre-estimated position information by using a preset water bottom detection method, and thus the positioning accuracy of the multi-beam probing sonar system on the real water bottom position of the target water area is improved.

The invention relates to the field of sonar image processing and specifically relates to a sonar image fusion method based on an extended Piella framework. The sonar image fusion method based on the extended Piella framework comprises the steps of (1) initializing parameters and inputting two groups of sonar images to be fused, (2) performing multiresolution conversion on the sonar images to be fused and establishing the respective multiresolution image sequences of the sonar images to be fused, and (3) performing multiresolution inverse transformation on the high frequency sub-bands and low frequency sub-bands of the images to be fused to obtain the final fusion result. The invention firstly provides a fusion method for sonar continuous frame images. Compared with an existing image fusion method, the sonar image fusion method based on the extended Piella framework is remarkably improved in both definition and detail expression of fusion.

The invention discloses a foresight sonar beam domain image splicing method, which comprises the following steps of: constructing an expression domain mapping constellation diagram, and forming a data conversion relationship between a beam domain and an image domain; according to an expression domain mapping constellation diagram, splitting actual sonar motion, and constructing an expression domain mapping model in a rotation motion mode and a displacement motion mode; the method comprises the following steps of: adopting an actually measured near-frame sonar image, decomposing the actual sonar motion condition according to rotation motion and displacement motion by taking a previous frame image as a benchmark, and performing image deformation operation on a next frame image by referring to an expression domain mapping model to realize target superposition in the two images; carrying out image fusion on the two beam domain images of which the targets are overlapped by adopting a Laplace pyramid hierarchical fusion strategy; a forward-looking sonar sequence image is adopted, continuous multi-frame data beam domain splicing operation is carried out, and the image is expanded to an image domain through interpolation. According to the method, edge seams generated in the image and splicing are well removed, and a false target value occurring in the fan-shaped expansion interpolation operation splicing process is eliminated.