32 results about "Benthos" patented technology

The invention relates to the field of underwater benthos detection, in particular to an underwater benthos detection method based on an improved DETR model. The method comprises the following steps: constructing an improved DETR model, and training by using a marked underwater image set to obtain a trained improved DETR model; using the trained improved DETR model to detect benthic organisms in the underwater image; wherein the improved DETR model comprises a backbone network, an encoder and a decoder with an auxiliary prediction head; in the backbone network, extracting image features by adopting a convolution gating linear unit which introduces a compact inverted bottleneck module and a channel shuffling module; in an encoder, a HIFI module combining a HiLo high and low frequency attention mechanism and multi-scale expansion convolution is adopted to enhance local detail preservation and global environment understanding, and an EDCM module is adopted to enhance feature representation and calculation efficiency of an underwater scene. According to the invention, underwater benthos detection can be realized with high precision and high efficiency, and the method is suitable for an underwater environment with limited resources.

The invention discloses a physical simulation method and device for the transportation effect of submarine organisms in a nuclear power plant, and the method comprises the steps: obtaining the motion parameters of target submarine organisms in the nuclear power plant, the motion parameters comprise one or more of individual mass, volume, body length, settling velocity, critical starting friction flow velocity, critical starting vertical line average flow velocity, starting wave height and critical starting flow velocity of target benthos of the nuclear power plant; establishing a physical model for simulating the transportation effect of the submarine organisms of the nuclear power plant; determining a simulant of the target benthos based on the motion parameters of the target benthos and a physical model; and simulating the amount of motion of the target benthos of the nuclear power plant by using the simulant and the physical model. In this way, it can be guaranteed that the motion process of the simulant should be kept similar to the actual motion process of the target benthic organism, the scale effect is avoided, and the premise is provided for accurately evaluating the entrainment and entrainment risk of the subbottom organism.

A marine benthic organism detection, identification and tracking method comprises the following steps: collecting, screening and constructing a marine organism benthic data set, performing small target data enhancement, and performing detection and preliminary classification on a target by using a P2-YOLOv11 deep learning model; then, a lightweight ShuffleDeepsort tracking algorithm is used for carrying out tracking; and finally, performing dynamic frame skipping identification in combination with track information obtained by tracking and preliminary classification information, and realizing fine classification under different order element systems by using a multi-level classification network to obtain a multi-level identification result of marine organisms. According to the invention, the processing effect of the small-target marine benthic organism image and the submarine multi-target tracking image actually shot by the underwater vehicle can be obviously improved.

The utility model discloses a river benthos sampling and analyzing device, and relates to the technical field of sampling devices. The sampling and analyzing assembly is used for sampling excreted sediments of river benthos, and the sampling and analyzing assembly comprises a sampling cylinder and a balancing weight fixedly arranged at the bottom end of the sampling cylinder; and the lifting assembly is used for automatically lifting the sampling analysis assembly. According to the river benthic organism sampling and analyzing device disclosed by the utility model, the sampling and analyzing component is arranged, so that when the device is used for sampling excrement of river benthic organisms, the aim of automatically sampling the excrement of the river benthic organisms is fulfilled, and in the sampling process, the excrement of the river benthic organisms can be automatically sampled. The rotation of the small motor can drive three L-shaped rods and bristles to stir the river bottom, so that excrement around the bottom end of the sampling barrel can be raised and is sucked into the sampling barrel, the purpose of sampling the periphery of the sampling barrel in a large range is achieved, and the sampling sufficiency is improved.

The utility model discloses a ship submarine organism cleaning robot, including main body frame, mobile system, cleaning execution system, adsorption balance system and identification system. Mobile system is by multi -freedom degree mechanical leg and the constitution of the Mecanum wheel, realizes the omni -directional movement on complex surface. The cleaning execution system is by rough cleaning wheel and fine cleaning wheel composition, through the gear set and transmission chain synchronous drive, utilizes the speed difference and spring pressure adjusting assembly to realize grading high -efficient cleaning. Adsorption balance system is by electromagnet, counter -thrust rudder and compressed air bag constitute, and cooperatively provide stable adhesion and attitude control. The identification system includes camera, LED lamp and main control unit, can intelligent identification ship body attachment and dynamic adjustment cleaning strategy. The utility model can be stable operation in underwater complex environment, realizes the efficient, accurate cleaning of the attachment such as acorn barnacle, algae, improves the ship operation efficiency and prolongs the service life.

The invention discloses a benthic organism distribution measurement method and device influencing the conductivity of a water body, and relates to the technical field of aquaculture, and the method comprises the steps: dividing a target water area into measuring point grids; setting background conductivity in the non-target organism area; synchronously measuring the conductivity, the temperature and the water depth of the water body at the measuring point grid, and collecting the conductivity gradient in a vertical set range; correcting the original conductivity into normalized conductivity by using the temperature coefficient and the water depth coefficient; calculating a biological conductivity difference value; when the biological conductivity difference value exceeds a preset threshold value and lasts for a certain time, biological negative anomaly is judged; inverting the position of the target organism by taking the biological conductivity difference value as input; after all the positioning points in the grid are clustered, estimating a target biomass quantity; splicing each grid result, and outputting a target biological density distribution diagram and a total amount; according to the method, the accuracy of benthic organism distribution measurement is improved through conductivity space-time mapping and a spherical intersection algorithm.

The application belongs to the technical field of marine benthic organism detection, and provides a Faster R-CNN marine benthic organism detection method based on an improved attention mechanism. In view of the problems of complex environmental noise, diverse biological scales, insufficient target detection frame positioning and classification accuracy and the like of marine benthic organisms, an improved multi-scale channel attention mechanism is used to optimize a ResNet50 network, and the ResNet50 network is used as a backbone network of a Faster R-CNN model to enhance the feature extraction capability. Then, a RoIAlign layer with a bilinear interpolation technology is used to replace a traditional RoIPooling layer, so that the position accuracy of the detection frame is improved. In the final classification stage, an extreme learning machine (ELM) feedforward classification network is introduced to replace a full connection layer, so that the classification accuracy is improved. The average precision mean (mAP) of the model is as high as 82.5% and 81.36% respectively, and these results fully prove the effectiveness of the algorithm.

The invention relates to the technical field of benthos ecological research, in particular to a large benthos inhabiting device which comprises a floating ring, the bottom of the floating ring is fixedly connected with a net bag with an upward opening, the bottom of the inner side of the net bag is connected with a plurality of sets of bottom nets distributed in the vertical direction, and the lower end of the net bag is fixedly connected with a fixing ring; when the amount of feed in the feeding box reaches a certain amount, balls are separated from extrusion rods by means of the weight of the feeding box, when the feeding box moves downwards to a certain position, a piston plate and a piston rod are driven to move downwards by means of pressure generated by water, and when the piston rod moves downwards, a swing plate is extruded, so that a magnet on the swing plate is not magnetically attracted to the feeding box any more; meanwhile, when the swing plate swings, the feed in the feeding box is discharged into the water body, so that the benthic organisms in the area can be conveniently fed, and the feeding convenience is improved.

The invention provides a benthic organism induction isolation protection device and method, and belongs to the technical field of equipment for underwater operation, the benthic organism induction isolation protection device comprises two seat bodies, the upper surfaces of the seat bodies are provided with insertion holes, and the opposite sides of the two seat bodies are provided with notches; the reel motor is arranged at the bottom of the insertion hole, and a motor shaft of the reel motor and the insertion hole are coaxially arranged; one end of the reel is inserted into the insertion hole, and the reel is connected with a motor shaft of the reel motor; two ends of the separation net are respectively connected with the reels on the two seat bodies; the sound wave generator is fixed on the seat body; the propellers are arranged at the four corners of the seat body respectively and used for adjusting the posture and the position of the seat body. According to the benthos induction isolation protection device, a target work area and an isolation protection area can be isolated according to a set area. The sound wave generator is arranged to guide migration, it is guaranteed that the organisms cannot be affected by mining and die, and it can be avoided that the repelled organisms are prone to being affected by natural enemies or the environment and die.

The invention belongs to the field of ecological risk assessment, and particularly relates to an ecological risk assessment method for heavy metal in sediment based on benthic organism heavy metal tissue residues. The method comprises the following steps: (1) extracting and testing the total amount of heavy metals and the occurrence form of the heavy metals in a surface sediment sample of a target drainage basin, and measuring the content Ci of the ith heavy metal in an effective indication organism; (2) constructing a biological residue safety threshold Pi of the ith heavy metal in the sediment of the target drainage basin for the effective indicator organism; and (3) based on the improved single-factor tissue residual index TI and the comprehensive tissue residual index NTI, dividing the ecological risk of a single heavy metal element to the target drainage basin and the comprehensive ecological risk level of multiple heavy metal elements to the target drainage basin in the sediment. Important reference and basis are provided for evaluating the implementation effect of developed treatment measures and formulating scientific sediment heavy metal pollution treatment plans in the later period.

The utility model provides a sampling support and a sampling device for collecting large benthonic animals, which are used for solving the problem that a net opening of a triangular trawl net does not sink to the bottom of a river when the large benthonic animals are collected in the prior art, and comprise a sleeve type sampling rod and a triangular net opening support enclosed by three connecting rods, a dragging rope fixing ring is arranged at the joint of every two connecting rods of three connecting rods of a triangular net opening support, a fixing piece used for being connected with a sleeve type sampling rod is arranged in the center of the bottom connecting rod of the triangular net opening support, and the sleeve type sampling rod is arranged. According to the sampling device for the large benthonic animals, the sleeve type sampling rod can be pressed to ensure that all the sampling devices sink to the bottoms of rivers which are different in water depth and cannot wade, and then it is ensured that ideal samples can be collected when the large benthonic animals are collected.

The present invention discloses a clamping type biological sampler for a deep - sea submersible manipulator, comprising: a sampler support body; an electrical control cabin, the limit pin assembly includes a limit pin, and the controller can control and drive the limit pin to perform telescopic actions; a handle; a net bag, which has an opening at the bottom; a self - shrinking closing part, which is a closed structure; an outer traction part, one end of which is connected to the self - shrinking closing part, and the other end is a free end for detachably connecting with the limit pin; an inner traction part, one end of the inner traction part is connected to the self - shrinking closing part, and the other end is fixed on the inner wall of the sampler support body. The clamping type biological sampler for a deep - sea submersible manipulator of the present invention solves the problem that it is difficult to directly capture organisms with strong activity abilities such as fish, shrimps, and crabs at present. It has a high capture rate, convenient sampling, and can greatly increase the types and quantities of seabed biological samples.

The invention discloses a tidal flat benthonic animal distribution prediction method based on remote sensing satellite data, and relates to the technical field of tidal flat benthonic animal distribution research, and the method comprises the steps: obtaining habitat feature data through inversion according to the remote sensing satellite data, and carrying out the preprocessing; sea surface temperature data are obtained according to remote sensing satellite data and preprocessed; sampling a target area to obtain samples and benthonic animal sampling data, and preprocessing the samples and the benthonic animal sampling data; dividing the preprocessed habitat feature data, sea surface temperature data and benthonic animal sampling data into a training set and a verification set, establishing a machine learning model, and pre-training the machine learning model; selecting an optimal hyper-parameter of the machine learning model based on a particle swarm optimization algorithm; and verifying the fitting degree of the machine learning model based on the verification set to screen out an optimal model, and obtaining a benthonic animal prediction distribution diagram by using the optimal model. According to the method, the benthonic organism data is fused with the optimization algorithm and the machine learning model, so that the prediction precision of the benthonic animal distribution of the tidal flat is improved.

The invention relates to the technical field of ecological monitoring, and discloses a key habitat element identification method and device for benthonic animals in a diversion lake, and the method comprises the steps: normalizing the multi-source environment and biological data of a target diversion lake, and carrying out the quality control fusion according to ecological monitoring specifications, and forming a standardized environment-biological data set; constructing a data network topological structure by taking a preset ecological mechanism library as a base, taking habitat elements as a middle layer and taking benthonic animal states as a top layer; fitting topological structure nodes and key edge parameters to obtain a parameterized causal relationship network; a habitat element node intervention value is set, anti-fact intervention simulation is carried out, and a benthonic animal state continuous response value is obtained; discretizing the response value according to a preset ecological grade standard to obtain benthonic animal state simulation distribution; quantitatively simulating the distribution causal effect intensity, and sorting the key habitat elements to form a key habitat element list; according to the invention, the key habitat element identification efficiency can be improved.

The application discloses a kind of based on improved YOLO11n's marine ranching benthos detection method, specifically includes the following steps: (1) establishes marine ranching benthos image dataset;(2) add annotation information for the image in dataset;(3) construct based on improved YOLO11n's marine ranching benthos detection model;(4) training set and verification set are used to train model and save the model trained;(5) test set is used to test model, meet the requirement of accuracy and generalization, that is, obtain the final benthos detection model.Compared with prior art, the marine ranching benthos detection method based on improved YOLO11n disclosed in the application can enhance the detection capability of the occluded target, enhance the discrimination degree of benthos and background, and improve the detection capability of small target benthos.

The invention belongs to the field of computer vision, and particularly relates to an ocean particle image velocity measurement method based on unsupervised learning and multi-scale deep convolution. Based on a classical optical flow network LiteFlowNet, in combination with a multi-scale Inception deep convolution module, the ability of the network to extract complex flow field features is enhanced. Meanwhile, a sub-pixel up-sampling layer is adopted to replace a traditional bilinear interpolation method, so that the reconstructed flow field is finer. Besides, the network is trained through unsupervised loss, dependence on a large-scale annotation data set is eliminated, structural similarity loss is fused in an unsupervised loss function, so that image structure information is more effectively reserved, and the precision of optical flow field estimation is further improved. The device can be used for detecting a seabed flow field, and plays a key role in ocean development and benthic organism research.

The present invention proposes a device and method for inducing isolation and protection of seabed organisms, which belongs to the technical field of equipment for underwater operations, and includes: two bases, wherein the upper surface of the base is provided with a socket, and the opposite side of the two bases is provided with a notch; a reel motor, which is arranged at the bottom of the socket, and the motor shaft of the reel motor is coaxially arranged with the socket; a reel, one end of which is inserted into the socket, and the reel is connected to the motor shaft of the reel motor; an isolation net, the two ends of which are respectively connected to the reels on the two bases; an acoustic wave generator, which is fixed on the base; a propeller, a propeller is respectively provided at the four corners of the base, for adjusting the posture and position of the base. The seabed biological induction isolation and protection device of the present invention can isolate the target work area and the isolation protection area according to the set area. By setting the acoustic wave generator to guide migration, it is ensured that the organisms will not be affected by mining and die, and it can also prevent the expelled organisms from being easily affected by natural enemies or the environment and die.

The utility model discloses a collecting device suitable for river benthos, and relates to the technical field of biological monitoring. Comprising a shell, the bottom of the shell is an opening, and a hanging plate is arranged at the top; the two grab buckets are symmetrically arranged on the side walls of the two ends of the shell, and the grab buckets are used for opening or closing the opening; one end of the rotating shaft is fixedly connected with the grab bucket, and the other end is rotationally connected with the shell; the torsional spring sleeves the rotating shaft and is used for driving the bottoms of the two grab buckets to get close to each other; the first electromagnet is arranged on the side wall, away from the other grab bucket, of the grab bucket; the second electromagnet is arranged at the upper end of the side wall of the shell and corresponds to the first electromagnet; and the rope is arranged on the lifting plate. The device has a simple structure, is convenient to operate and carry, and is convenient for working personnel to use under various conditions.

The invention discloses a modular ecological protection and scour prevention integrated device of an offshore wind power foundation and a construction and maintenance method. The device is arranged around an offshore wind power foundation and comprises an annular support, an anti-scour assembly and an ecological protection module assembly. The annular supporting and anti-impact assembly comprises a rigid inner ring plate, a rigid outer ring frame, a rigid supporting plate and an interlocking block soft mattress. The rigid inner ring plate is fixedly connected with the circumferential outer wall of the offshore wind power foundation; the rigid outer ring frame is arranged on the circumferential outer side of the rigid inner ring plate in a surrounding mode. The opposite sides of the rigid inner ring plate and the rigid outer ring frame are connected through a rigid supporting plate. An interlocking block soft body row is arranged on the upper surface of the rigid outer ring frame in a surrounding manner; the ecological protection module assembly comprises a bionic aquatic plant frame module; and the bionic aquatic plant lattice module is arranged between the rigid inner ring plate and the rigid outer ring frame. According to the offshore wind power foundation, scouring of the seabed around the offshore wind power foundation can be slowed down, meanwhile, an inhabiting space is provided for benthos, and unitized replacement and long-term monitoring are facilitated.

The present application belongs to the field of environmental pollution and environmental chemistry, and particularly relates to a model and method for simulating the marine environmental behavior and food web migration of typical persistent organic pollutants. The model comprises atmosphere, seawater (including suspended particles), sediment (including pore water), and a food web unit composed of multi-trophic level organisms (including plankton, benthos and fish of different diets). The present application integrates biological energy science and input parameters of aquaculture activities (such as feed delivery, residual feed loss, and fecal excretion), and constructs a multi-medium non-steady-state fate model of the aquaculture marine ecosystem. The model can accurately quantify the cross-medium flux of organic pollutants (such as PFAS, PCBs, DDTs, pesticides, etc.) between the "aquaculture source-environment-organism", realizes the quantitative tracing of the environmental contribution rate of aquaculture activities, and can predict the temporal and spatial distribution and enrichment level of pollutants in wild fish, thereby providing a scientific basis for the ecological risk management of the marine aquaculture area and the overall marine environment.

The utility model relates to the technical field of underwater sampling robots, and particularly discloses a squid bionics-based shallow sea organism sampling robot which comprises a bionic main body, prefabricated connecting pieces are arranged on the front side and the rear side of the bionic main body, and a detachable picking storage bin and a detachable floating collection bin are respectively arranged on the prefabricated connecting pieces on the two sides. A mechanical arm connected with the bionic body is further arranged on the side adjacent to the picking storage bin, a bin door plate capable of being automatically opened and closed is arranged on the picking storage bin and connected with the mechanical arm in a matched mode, and the floating collecting bin and the picking storage bin are symmetrical in weight. According to the shallow sea organism sampling robot based on cuttlefish bionics, the motion mode of bionic benthos is adopted, so that disturbance of the robot to surrounding water bodies in the seabed operation process is remarkably reduced. Through the synergistic effect of the bionic appearance and the flexible driving mechanism, the robot can carry out fine operation under the condition that the stable posture is kept.

The present invention relates to a device for suppressing the plume of deep - sea mining tail - water discharge, which comprises a mineral tail - water drain pipe, a plume - suppressing pipeline cover, and damping sheets. The tail - water inlet of the plume - suppressing pipeline cover is fixedly connected to the mineral tail - water drain pipe. A plurality of the damping sheets are fixedly connected inside the plume - suppressing pipeline cover, and the size of the damping sheets gradually increases from the tail - water inlet to the outlet direction. The device for suppressing the plume of deep - sea mining tail - water discharge of the present invention effectively reduces the flow velocity of the discharged tail - water through its unique damping structure (since the effective discharge cross - sectional area of the suppression device exceeds the cross - sectional area of the tail - water drainage pipeline by at least 30%, the efficiency of tail - water discharge will not be affected), significantly reduces the diffusion velocity of the plume sediment, and thus controls the impact of the tail - water discharge particulate matter on the seabed organisms and the marine environment to the minimum extent.

The utility model relates to a kind of intelligent portable deep layer collection benthos device, the top of hemispherical cover is fixed with arc handle, and the bottom plate of hemispherical cover is opened with hanging hole;Maintenance door is arranged on the opening of the side wall of hemispherical cover;Visual digital display module is embedded in the side wall of hemispherical cover;DC motor is fixed in the inside of hemispherical cover, and its output shaft is connected with winding wheel disc, winding wheel disc is wound with wire rope, and the lower end of wire rope is movably arranged in hanging hole;Internal control component is fixed in the inside of hemispherical cover;The DC motor is electrically connected with internal control component, and internal control component is connected with visual digital display module.Utilize arc handle and corresponding frame pole, through buckle interaction, increase the overall stability when the device works and facilitate staff to carry.

This invention belongs to the field of ecological risk assessment, specifically relating to a method for assessing the ecological risk of heavy metals in sediments based on the heavy metal tissue residues of benthic organisms. The method includes the following steps: (1) extracting and testing the total amount and occurrence forms of heavy metals in surface sediment samples from the target watershed, and effectively indicative of their presence in organisms. i Heavy metal content C i (2) Construct the first sediment in the target watershed. i Biosafety thresholds for heavy metals as effective indicator organisms P i (3) Based on the improved single-factor tissue residue index (TI) and the comprehensive tissue residue index N TI This invention classifies the ecological risks of a single heavy metal element in sediments to a target watershed, as well as the combined ecological risks of multiple heavy metal elements to the target watershed. This invention will provide important reference and basis for assessing the effectiveness of existing remediation measures and for developing scientific plans for the remediation of heavy metal pollution in sediments.

The invention discloses a deep learning integrated RVFL-based benthic organism image classification method and system, and belongs to the field of computer vision and deep learning. The method comprises the following steps: uniformly scaling an input benthic organism image, and extracting initial features through a preprocessing module; the tail end of a middle-level feature extraction module outputs a feature map L1 through space attention weighting, the tail end of a deep-layer feature extraction module outputs a feature map L3 through channel attention weighting, and a high-level module outputs a feature map L2; performing global pooling and normalization on the L1, the L2 and the L3, and splicing the L1, the L2 and the L3 to form a fusion feature vector; and finally, predicting biological category probability distribution by a TRVFL classifier according to the fusion features. The method is suitable for marine organism recognition, underwater robot navigation, underwater monitoring and other tasks, and the classification accuracy is remarkably improved under the conditions of limited samples and complex backgrounds.

This invention provides a hollow reef assembly with a seaweed-like structure. The assembly comprises multiple stacked reef blocks, each a rectangular concrete block. A rectangular channel is located in the center of each reef block, with its opening on the upper surface. Four grooves connect the opening to the four sides of the rectangular channel and are communicatively connected to it. A metal mesh is installed inside the rectangular channel, and a seaweed-like structure is inserted through it, with its top extending above the upper surface of the reef block. The grooves divide the upper surface of the reef block into four areas, each covered with seaweed. This invention, by incorporating the seaweed-like structure on the reef surface, allows marine life to gather and cling, compensating for the lack of shelter during certain seasons when seaweed is absent. The metal mesh not only secures the seaweed-like structure but also provides additional support.

The invention discloses a multi-depth benthos collection device, and belongs to the technical field of benthos collection, the multi-depth benthos collection device comprises a collection frame body, one side of the bottom of the collection frame body is provided with a sliding groove, and a vibration assembly is arranged in the sliding groove. According to the benthic organism collecting device, by arranging the adjusting and mounting assembly, sewage and soil in the collecting frame body can flow out from the gaps after being elutriated by water flow, and sample organisms are reserved in the collecting frame body, so that the benthic organism samples are cleaned and separated, the situation that the collecting quality is influenced by unnecessary impurity sewage is avoided, and the collecting efficiency is improved after elutriation is completed. A worker pulls a limiting rod, so that the limiting rod is separated from a limiting hole, limiting on a mounting frame is relieved, then the worker pulls the mounting frame out of a mounting groove seat through a handle, one side of a collecting frame body is completely opened, the worker can conveniently take out benthos samples from the collecting frame body, benthos are prevented from remaining in the collecting frame body, and the working efficiency is improved. The sampling efficiency is improved.

A modularized bottom mud cleaning and benthic organism extracting device comprises a barrel and a cover, a stirring cleaning unit is arranged on the cover, a multi-layer screening unit, a rotating unit and a drainage unit are arranged in the barrel, the stirring cleaning unit comprises a stirrer arranged in a stirring barrel and a water spraying assembly fixed to the inner side surface of the cover, and the stirring cleaning unit is arranged in the barrel. A center shaft of the stirrer is fixed to the center position of the inner side of the cover body, the multi-layer screening unit is arranged in the containing space in the barrel body, the multi-layer screening unit comprises three stages of filtering barrels which are sequentially nested from top to bottom, and a uranium line of the multi-layer screening unit coincides with a rotating shaft of the rotating unit. By combining a multifunctional mechanical structure and an intelligent adjusting system, mud pies, benthos and impurities can be efficiently separated, the cleaning mode can be flexibly adjusted according to different scene requirements, and the applicability and the cleaning efficiency of the equipment are remarkably improved.

The utility model relates to the technical field of benthonic animal sampling, and discloses a portable benthonic animal sampler which comprises a sampling box, sieve plates are fixedly connected in the left side and the right side of the sampling box, a closing mechanism is arranged in the top of the sampling box, and entering mechanisms are arranged at the bottoms of the front side and the rear side of the sampling box. The closing mechanism comprises a closing assembly, a telescopic assembly, an adjusting assembly and a limiting assembly. According to the portable benthonic animal sampler, benthonic organisms located in a river channel enter the sampling box through the trapping opening through the arranged entering mechanism, a limiting block limits the overturning direction of an overturning net plate through a protruding piece, and the overturning net plate can only be overturned inwards and cannot be overturned outwards; the benthos entering the sampling box cannot leave through the trapping opening, so that the device can automatically complete sampling of the benthos after being placed in a river channel, auxiliary operation of workers is not needed, and the efficiency of trapping and sampling the benthos when the device is used is improved.

The invention discloses a multi-scale marine benthos detection method based on global-local cross-attention fusion, and the method comprises the steps: collecting and preprocessing submarine data information through an underwater robot, and building a multi-scale marine benthos data set; preprocessing the multi-scale marine benthic organism data set and dividing the multi-scale marine benthic organism data set into a training set and a test set; building a multi-scale benthic organism detection framework based on global-local cross-attention fusion; adopting the training set to train the multi-scale benthos detection framework to obtain a multi-scale marine benthos detection model; and performing deployment and hyper-parameter adjustment on the trained multi-scale marine benthos detection model, and using the obtained model to detect marine benthos. According to the method, histogram equalization is introduced to pre-process the image, so that the image contrast is effectively enhanced, and the problem of low contrast of the underwater image caused by uneven illumination, color degradation or blurring is solved.