335results about How to "Improve detection stability" patented technology

The invention discloses a rainy day target detection method based on paired image data generation and knowledge distillation, and belongs to the technical field of automatic driving target detection. The method comprises the following steps: acquiring a real image of a road scene and a semantic segmentation mask; inputting the semantic segmentation mask and the weather condition text information into a trained improved diffusion model to obtain a road scene generation image under a rainy day condition; introducing a CoT module and an ODConv module into the target detection model for optimization, and then improving the target detection model by adopting a dynamic weighted aggregation loss function to obtain an optimized target detection model; performing knowledge distillation to obtain a rainy day target detection enhancement model; and performing target detection by using the rainy day target detection enhancement model. The method provided by the invention is excellent in performance in a rainy day scene, has good generalization ability, can provide higher-precision target detection in various complex scenes, and effectively improves the detection accuracy and stability.

The invention discloses a colorimetric and fluorescent dual-mode nano-enzyme sensor. The colorimetric and fluorescent dual-mode nano-enzyme sensor comprises nano-enzyme, hydrogen peroxide and a colorimetric probe, wherein the nano-enzyme is prepared from ZrCl4, 2-aminobiphenyl dicarboxylic acid and Cu (NO3) 2 by a one-pot method; the colorimetric probe is TMB, can oxidize colorless TMB into blue oxTMB, generates an ultraviolet absorption peak at 652 nm, and has a strong fluorescence emission peak at 405 + / -5 nm. The zirconium-based metal organic framework UiO-67-NH2 with high stability is used as a carrier, Zr < 4 + > sites are replaced by part of Cu < 2 + >, the fluorescent nano-enzyme with peroxidase activity is formed, and the problems that free enzyme is difficult to recover, the stability is low and the like are solved.

The invention provides a power station photovoltaic module defect detection method and device based on unsupervised learning. The method comprises the steps that firstly, a thermal imaging image to be detected and a visible light image in a target area provided with a photovoltaic module are collected; cutting a thermal imaging photovoltaic module image from the thermal imaging image, and cutting a visible light photovoltaic module image from the visible light image; afterwards, a teacher network model and a visible light photovoltaic student network model in the photovoltaic defect detection model are used for processing the visible light photovoltaic module image, and position information of the visible light abnormal photovoltaic module is determined and returned; and the teacher network model and the thermal imaging photovoltaic student network model in the photovoltaic defect detection model are used to process the thermal imaging photovoltaic module image, and the position information of the thermal imaging abnormal photovoltaic module is determined and returned. Wherein the photovoltaic defect detection model is obtained by training in a label-free supervision mode.

The invention provides a key point prediction model training method and device, electronic equipment and a computer readable storage medium. The method comprises the following steps: acquiring a sample key point in a sample image and first sample position information of the sample key point; extracting a feature map of the sample image through a to-be-trained key point prediction model; determining first prediction position information of the sample key point and first prediction offset information of a target pixel area where the sample key point is located in the feature map; determining a model loss value based on the first sample position information, the first prediction position information, and first sample offset information and first prediction offset information of the sample key point in the target pixel region; and performing model parameter updating on the to-be-trained key point prediction model based on the model loss value to obtain a trained key point prediction model. According to the invention, the precision and stability of the key point prediction model can be improved.

The invention discloses a photovoltaic cell hot spot intelligent detection method fusing multi-modal data. The method comprises the following steps: S1, collecting multi-modal data of a photovoltaic cell; s2, carrying out preprocessing on the multi-modal data; s3, constructing a photovoltaic cell sample data set; s4, extracting features of the multi-modal data; s5, fusing the multi-modal features to generate a comprehensive feature vector; s6, training and optimizing a dynamic hot spot detection model based on the space-time convolutional network; s7, identifying a hot spot area in the photovoltaic cell by using the model; s8, positioning and outputting a hot spot detection result; and S9, optimizing the detection model in real time in combination with a dynamic adaptive algorithm. According to the method, the multi-modal data fusion, the space-time convolutional network and the dynamic adaptive optimization algorithm are introduced, so that the hot spots in the photovoltaic cell can be detected and positioned more accurately and quickly, the false detection rate in a complex environment is effectively reduced, and the operation efficiency and reliability of a photovoltaic cell system are remarkably improved.

The invention relates to an Internet of Vehicles multi-agent cooperative sensing method and system. The method comprises the following steps: constructing a collaborative sensing network, wherein the collaborative sensing network comprises a self-agent and a plurality of collaborative agents; acquiring and processing sensing data through the collaborative sensing network; performing feature extraction to obtain intermediate features; self-adaptive sparsification is carried out to obtain sparse features, and the sparse features are compressed and transmitted to a self-agent; performing time sequence feature enhancement on the features of all the agents at the self-agent end; fusing the features to obtain fused features; and constructing an Internet of Vehicles perception model, and realizing perception by the detection model according to the fused features. According to the method, the calculation complexity of traditional global attention is reduced from the square level to the linear level through an adaptive sparsification mechanism, the calculation overhead is remarkably reduced while the multi-agent feature interaction precision is kept, and the method is more suitable for real-time operation on the vehicle-mounted edge equipment with limited resources.

The invention provides a self-diagnosis method and device for a dual-redundancy pressure sensor and electronic equipment, and the method comprises the steps: determining a stability criterion of the pressure sensor according to a data curve collected by the pressure sensor and a corresponding standard input data set; determining a data confidence criterion and a data weight of the pressure sensor according to the data curve; according to the sensitivity parameter of the pressure sensor, the environment temperature curve and the offset error parameter, the sensitivity weight, the temperature weight and the offset weight of the pressure sensor are correspondingly determined; integrating the data weight, the sensitivity weight, the temperature weight and the offset weight into an evidence weight of the pressure sensor; and determining a normal or abnormal state of the pressure sensor according to the stability criterion, the data confidence criterion and the evidence weight corresponding to the two paths of pressure sensors. The weight distribution can be dynamically adjusted by fusing various data evaluation indexes, so that the detection precision and stability of a pressure sensor system are improved.

The invention relates to the technical field of gas purity detection, in particular to a gas source purity detection device and a gallium nitride production system. The gas source purity detection device comprises a main pipe body, a supporting piece, a control ring, a moving arm and a concentration sensor. The main pipe body is used for being connected with an air source pipeline. The control ring is rotationally matched with the supporting piece and driven by the driver, and the control ring and the outer wall face are rotationally sealed. A gas collecting groove is formed in the inner ring wall of the control ring. The supporting piece is provided with a mounting blind hole, and the moving arm is slidably matched with the mounting blind hole and is slidably sealed. A first elastic piece is connected between the moving arm and the hole bottom wall of the mounting blind hole in an abutting mode. The concentration sensor is arranged on the outer side wall of the moving arm. When the control ring rotates, the gas collecting groove can periodically communicate with the main pipe body and can also periodically move to the moving arm. The gas source purity can be continuously and accurately detected for a longer time, and automatic calibration is achieved while purity detection is conducted.

The invention belongs to the technical field of quality detection, and particularly relates to an intelligent quality inspection and process optimization system based on machine vision, which comprises an image acquisition unit, an illumination compensation unit, an image preprocessing unit, a feature extraction unit, a feature matching unit, a difference analysis unit, a quality evaluation unit, a process adjustment suggestion unit and a feedback execution unit. By introducing the illumination compensation unit, the illumination intensity can be automatically adjusted under various illumination conditions, and it is ensured that the quality of the obtained image is always in the optimal state. The improvement not only improves the effect of image preprocessing, but also enhances the accuracy of feature extraction and matching, and finally improves the detection precision and stability of the whole system. In addition, the illumination compensation unit is added, so that the requirement of manual intervention is reduced, and the system is more intelligent and automatic and adapts to a wider production environment.

The invention relates to the technical field of rubber tire detection, in particular to an air tightness detection device for rubber tire production and a detection method thereof. Comprising a detection platform, and a telescopic groove is formed in the top of the detection platform; and a tire supporting mechanism movably penetrates through the telescopic groove. The tire limiting mechanism is controlled to limit and fix a rubber tire, then the tire supporting mechanism is controlled to be separated from the rubber tire, the detection marking mechanism is controlled to wrap a part of the rubber tire, and then the tire limiting mechanism is controlled to drive the rubber tire to rotate; through the arrangement of the detection marking mechanism, the detection marking mechanism can mark the air leakage position while detecting the surface in the rotation process of the rubber tire, the problems that manual detection wastes time and omission is likely to happen are solved, the detection work efficiency is improved, the air leakage point can be accurately marked, and the detection efficiency is improved. The problem that follow-up repair is inconvenient due to the fact that no mark exists after existing liquid detection is dried is solved, and the using effect of the detection device is improved.

The invention belongs to the technical field of optical fiber gas sensing, and discloses an optical fiber photoacoustic gas sensing system and method suitable for a low-pressure environment. The system comprises a programmable logic gate array circuit, a distributed feedback laser, a laser transmission optical fiber, an optical fiber photoacoustic sensing probe, a super-radiation light-emitting diode light source, an optical fiber circulator and a spectrograph. The air pressure is monitored according to the linear relation between the resonance frequency of the cantilever beam and the ambient air pressure, the characteristic that the acoustic sensitivity at the resonance position of the cantilever beam is enhanced under the low air pressure is utilized to make up for signal attenuation in the photoacoustic excitation process, and the sensitivity is corrected in combination with the air pressure measured by the cantilever beam. Therefore, high-precision and high-sensitivity detection of trace gas in a low-pressure environment is realized. The system only adopts a single cantilever beam and a single demodulator, has the functions of gas pressure and gas concentration detection, and can realize accurate detection of trace gas in a gas pressure change environment. And an all-fiber scheme is adopted, so that intrinsic safety is realized, and remote measurement and multi-point network group detection are easy.

The invention provides a method and a system for mixing substances in magnetic control liquid drops. The method for mixing the substances in the magnetic control liquid drops comprises the following steps: dropwise adding the liquid drops containing substances to be mixed into a hydrophobic solid base material filled with magnetic fluid, and enabling the magnetic fluid to quickly wrap the contacted liquid drops to form a magnetic response liquid shell; the magnetic response liquid shell is controlled to move to drive the liquid drops wrapped by the magnetic response liquid shell to move, and mixing of substances in the liquid drops is accelerated. The magnetic control liquid drop internal substance mixing system comprises a hydrophobic solid base material and magnetic fluid spread on the surface of the hydrophobic solid base material. The magnetofluid response shell wrapping the liquid drops rotates around the liquid drops, shear flow at a two-phase interface is generated, kinetics in the liquid drops is increased, and the process of mixing substances in the liquid drops is obviously accelerated.

The invention belongs to the field of rockfill material grading detection, and relates to a rockfill material grading detection method and device based on an image segmentation algorithm, and the method comprises the steps: 1) obtaining an image after paving of stone particles; 2) preprocessing the stone image; 3) inputting the preprocessed image into an instance segmentation model to obtain mask information of stone particles; 4) extracting mask quantitative indexes and shape characteristic parameters in the mask information of the stone particles, and performing particle size and stone total volume conversion in each particle size range on the mask quantitative indexes and the shape characteristic parameters; (5) converting the obtained stone particle size and volume into three-dimensional stone particle gradation, and drawing a stone particle apparent gradation curve; 6, the obtained stone particle apparent grading curve is corrected through the data fusion model.The final stone particle three-dimensional overall grading curve is obtained.The method has the advantages of being high in adaptability, free of a special detection environment and high in detection precision.

The invention discloses an automatic rotation observation and analysis device for a bearing part, and the device comprises a pedestal which is provided with a rotating disc and a driving part which is used for driving the rotating disc to rotate axially. An inner clamp used for fixing an outer bearing inner ring to be detected and an outer clamp used for being matched with the inner clamp to fix an outer bearing outer ring to be detected are arranged on the surface of the rotating disc, and an adjusting structure used for adjusting the distance between the inner clamp and the outer clamp to adapt to the outer bearing inner rings or the outer rings of the outer bearings to be detected with different diameters is arranged on the rotating disc. The problems that in the prior art, when a failure bearing part is detected, the bearing part needs to be manually driven to operate, and the detection precision is low are solved.

The invention discloses a full-automatic spring plunger height and torsion detection device and method, and the device comprises an automatic feeding positioning mechanism which is provided with a spring plunger positioning groove; the automatic detection mechanism comprises a lower die assembly, a pressure sensor and a servo lifting driving unit; a detection positioning groove with an opening in the upper end is formed in the lower die assembly; the servo lifting driving unit drives the pressure sensor to move downwards to a first height, the pressure sensor makes contact with the floating head of the spring plunger, the height size is calculated, then the servo lifting driving unit drives the pressure sensor to move downwards continuously to a second height, and the height size is calculated. The torsion of the spring is tested through the counter-acting force of the spring in the spring plunger to the pressure sensor, the torsion data is automatically recorded, and qualified products and defective products are judged; and the automatic carrying mechanism is used for automatically transferring the spring plungers into the detection positioning grooves, and discharging the detected spring plungers to a corresponding qualified product area and a defective product area according to a qualified product area and a defective product area.

The invention relates to the technical field of visual inspection, in particular to a stamping part defect detection method based on vision, which comprises the following steps: S1, shooting setting; s2, synchronizing video streams; s3, acquiring a part bounding box and a hole position bounding box; s4, associating part bounding boxes belonging to the same stamping part among the frame part images, and associating hole position bounding boxes belonging to the same screw hole position; s5, correspondingly associating the bounding boxes of the target parts belonging to the same stamping part in the part image; s6, when a hole position boundary frame matched with the view angle template image of the corresponding view angle exists in any target detection image under any view angle, determining that a screw hole position exists; and when the matched hole position boundary frame is not found in the target detection image under the corresponding visual angle in the visual angle template images of all the visual angles, the screw hole position is missing. The stamping part defect detection device can be used for defect detection of stamping parts in the conveying process, the consistency of detection results is improved, and the detection efficiency is improved.

The invention relates to the technical field of telegraph pole tower state detection, and discloses a telegraph pole tower inclination condition real-time monitoring and positioning device and a corresponding method.The telegraph pole tower inclination condition real-time monitoring and positioning device comprises a monitoring shell and a monitoring and positioning assembly arranged in the monitoring shell, the opening of the monitoring shell is covered with a shell cover, and a mounting groove is formed in the upper side of the outer wall of the monitoring shell; a solar panel is embedded in the mounting groove, an arc-shaped clamping plate is fixedly arranged on the side, away from the solar panel, of the monitoring shell, rotating assemblies are arranged on the two sides of the arc-shaped clamping plate correspondingly, and a first bandage and a second bandage are arranged through the rotating assemblies correspondingly; the sides, away from each other, of the first binding band and the second binding band are provided with a plurality of first inserting grooves and second inserting grooves which are arranged at intervals respectively. According to the real-time monitoring and positioning equipment for the inclination condition of the telegraph pole tower and the corresponding method, the inclination condition of the telegraph pole tower can be continuously and accurately monitored for a long time by arranging a plurality of monitoring modules and stably connecting the monitoring modules with the outer wall of the telegraph pole tower.

The invention relates to a water quality detection method for detecting heavy metal elements in water, which comprises the following steps: S1, pretreating a water sample, and filtering to remove suspended particles and impurities; s2, acid digestion of the sample: digesting the sample by using a mixed solution of nitric acid and hydrochloric acid to release heavy metal elements; s3, electrochemical sensor detection: anodic stripping voltammetry (ASV) or differential pulse voltammetry (DPV) is adopted for detection, and the sensor uses a gold-plated electrode or a glassy carbon electrode modified with carbon nanotubes and graphene; s4, concentration calculation: calculating the heavy metal concentration according to a current peak value and a volume correction formula; and S5, verifying a result, and rechecking a detection result by adopting an ICP-MS or AAS method. The method is high in sensitivity, can accurately detect trace heavy metals in drinking water, industrial wastewater, surface water and underground water, is suitable for simultaneous detection of multiple elements, has the advantages of simplicity and convenience in operation, portability of equipment, high detection speed and the like, and is particularly suitable for on-site water quality monitoring.

The invention discloses a fruit and vegetable quality nondestructive testing imaging control system which comprises the following components: S1, a multispectral light source module which captures spectral information inside and on the surface of the multispectral light source module; s2, a precision imaging module generates multispectral image data; s3, extracting key image features related to fruit and vegetable quality by an image data preprocessing module; s4, the quality analysis module adjusts algorithm parameters in real time and generates an analysis data set corresponding to the fruit and vegetable quality indexes; s5, a spectrum enhancement and adaptation module, which optimizes the sensitivity and resolution of the multispectral signal; s6, a distributed detection control module is used for controlling system parameters to be adjusted in real time, and the detection efficiency and detection are optimized; and S7, a data management and output module which generates a fruit and vegetable quality evaluation report through a computer system. The method has the advantages of being high in detection precision, high in adaptability, good in nondestructive property and high in detection efficiency.

The invention discloses a sealant gluing quality detection method, which comprises the following steps: S1, acquiring images of a vehicle body sealant gluing area, and matching the images; s2, carrying out binarization processing on the corrected image, setting a target point and a width threshold value in the binarized image according to the vehicle model, measuring the gluing widths of the two sides of the target point, and comparing and judging the gluing centering degree; s3, performing gridding processing on the corrected image, equally dividing a gluing area in the gridding image along the length direction, setting the area of a target gluing area and a gluing coverage rate threshold value, calculating the area of an actual gluing area, and calculating the area of the actual gluing area; according to the area of the actual gluing area, the area of the target gluing area and the gluing coverage rate, whether missed gluing and glue breaking occur or not is judged; and S4, collecting image information in the step S2 and the step S3, importing the image information into a visual deep learning system, and establishing a deep learning model. The detection method disclosed by the invention can be used for comprehensively judging the quality according to the centering degree of gluing and whether missed gluing and broken gluing occur or not.

The invention relates to a precision detection device for gear processing, which is suitable for detecting the dynamic performance and single tooth precision of a gear, and comprises a base, a rotating bracket, a fixed block, a motor, a reduction gearbox, a test gear, a to-be-tested gear, a main measuring rod, a plurality of measuring rods and the like. High-precision detection of gear dynamic balance, vibration parameters and the tooth outer end and the contour line of a single gear is achieved, the reduction gearbox reduces the rotating speed of the gear and strengthens torque control, and the gear thickness difference design is matched, so that the multiple measuring rods can slide along the tooth outer end of the gear, tooth surface deformation and the contact state are accurately detected, and the device is made of high-carbon low-alloy tool steel. The gear detection device has high strength and wear resistance, adapts to various gear specifications, can output detection data and analysis results in real time, is simple to operate, high in detection precision and high in applicability, and can be widely applied to the fields of gear manufacturing, quality detection and maintenance.

The invention relates to the technical field of crushing machine detection, in particular to a detection device for a health-care food powder crushing machine, comprising: a crushing module for crushing health-care food into powdery health-care food through the crushing machine; the detection module comprises a current sensor which is connected with the crushing machine and is used for detecting the working current of the crushing machine, and a vibration sensor which is connected with the crushing machine and is used for detecting the vibration intensity of the crushing machine; the control module is connected with the crushing module and the detection module and used for judging whether the detection stability of the crushing machine meets the requirement or not according to the variance of the fluctuation amplitude of the working current of the crushing machine. When the requirements are not met, the rotation angle of the crushing machine is adjusted or the detection frequency of the pressure sensor is adjusted according to the average vibration strength of the crushing machine, and the feeding speed of the health food is adjusted according to the average vibration strength of the crushing machine and the temperature rising speed in the crushing machine. According to the invention, the detection stability of the crushing machine is improved.

The invention discloses an immunodetection kit based on a bimetal nano-bomb biological probe, and relates to the technical field of gene detection, the immunodetection kit comprises a test strip and an immunoprobe, the immunoprobe is a suspension of a bimetal nano-bomb and a monoclonal antibody, and the mixing ratio of the bimetal nano-bomb to the monoclonal antibody is 0.1-0.3 mg: 5 [mu] g. The bimetal nano-bomb has a symmetrical hollow nano-bomb structure, so that the bimetal nano-bomb has excellent colorimetric and photo-thermal properties, and the detection sensitivity of salmonella typhimurium is greatly improved.

The invention provides a reinforcing steel bar bending resistance detection device for civil engineering. The reinforcing steel bar bending resistance detection device structurally comprises a positioner, a mainframe box, a control panel, a supporting column, a parallel plate, a detection structure and a detection cavity, on the basis of newly-added components in the detection cavity, three steel bars can be inserted in parallel at the same time through the restraining plates on the left side and the right side of the bearing body and the three inserting grooves, the middle sections of the steel bars can be suspended in the cavity area, and therefore when the detection structure is pressed downwards for detection, the two ends of the steel bars are limited in the restraining plates and cannot tilt up; therefore, the parallel fixity of the steel bar is improved, the middle section area does not deviate during downward pressing detection, the detection accuracy and the safety coefficient of the steel bar can be improved, meanwhile, the limiting ring at the edge of the inserting groove can be combined with the reinforcing column to improve the overall hardness, and the deformation condition caused by the continuous influence of the pressure at the two ends of the steel bar is avoided; therefore, the steel bar can be continuously maintained in a circular shape, and the steel bar is prevented from being stuck inside after deformation.

The utility model discloses a sorting line for cigarette blast beads, which relates to the field of cigarette production, and comprises a storage bin, a roundness screening module, a diameter size screening module, an image acquisition module, a rejecting module, a qualified product bin and an unqualified product bin, the roundness screening module comprises a rack, a feeding piece, at least one screening piece, a first collecting bin and a second collecting bin, the screening part comprises an accelerating part and a rolling inclined surface; the accelerating part is used for providing an initial speed for materials to enter the rolling inclined surface; the image acquisition module is used for acquiring an image of the material and transmitting the image to the computer for image analysis and processing; the rejecting module is used for rejecting unqualified cigarette blast beads according to an analysis result of the computer; a roundness screening module and a diameter screening module are adopted, so that the detection stability is improved; a plurality of detection links of the image acquisition module and the rejection module are combined, and the whole finished product detection link of the cigarette blast beads is designed into one detection line, so that the automation degree of detection equipment is improved, and the detection equipment is advanced towards the direction of automatic and intelligent detection equipment.

The invention relates to the technical field of gas detection, in particular to a formaldehyde detection device based on multiple probes, which comprises a probe main body and a gas detection assembly, the gas detection assembly comprises a butt joint part arranged at the bottom of the probe main body, and a semiconductor gas detection element is arranged in the middle of the butt joint part. A mounting cylinder is arranged on the outer side of the semiconductor gas detection element, and a filtering piece is arranged at the bottom of the mounting cylinder; a heat conducting plate is arranged in the mounting cylinder, is in contact with the semiconductor gas detection element, and is provided with hollow grooves which are symmetrically distributed; and the drainage assembly comprises a heat insulation cylinder arranged at the empty groove, the heat insulation cylinder is fixedly connected with the mounting cylinder, a reciprocating assembly is arranged in the heat insulation cylinder, and the reciprocating assembly is heated through a heat conduction plate, so that reciprocating movement of the reciprocating assembly in the heat insulation cylinder is achieved. By arranging the drainage assembly, the effect of sucking gas at the bottom of the probe main body is achieved, and the formaldehyde detection accuracy of the semiconductor gas detection element is improved.

The invention discloses a steel structural member local corrosion intelligent identification and evaluation method based on deep learning, and the method comprises the steps: obtaining a steel structure surface image through an imaging device, carrying out the feature extraction and corrosion region identification of an input image through a YOLOv12 semantic segmentation model which introduces a CBAM self-attention mechanism, and generating a pixel-level corrosion mask; the actual physical area of the corrosion area is automatically calculated based on an area calculation algorithm of conversion of mask integration and ground sampling distance; the residual bearing performance of the corroded component is evaluated in combination with the corrosion area ratio and the morphological characteristic prediction model; and finally realizing visual display and report output of detection results, area data and performance indexes through a graphical interface. According to the invention, full-process automatic processing of local corrosion of the steel structural member from detection, segmentation, quantitative measurement and calculation to member performance evaluation is realized, and closed-loop intelligent evaluation from visual detection to mechanical evaluation is realized.

The invention discloses an AI substitute inspection method and system based on multi-modal perception, and relates to the technical field of intelligent detection. According to the method, images, audios, environmental parameters and equipment state data of an inspection area are synchronously collected through a multi-modal sensor array, and after space-time alignment preprocessing, multi-modal features are extracted and input into a pre-trained fusion model to generate unified feature representation. The fusion model adopts a Transform architecture to realize cross-modal feature interaction, an anomaly detection algorithm is used to identify and classify anomalies, a confidence coefficient is calculated in combination with an evidence theory, and finally an inspection report containing anomaly positions, types and processing suggestions is generated. According to the invention, the information limitation of single-mode inspection is solved, the detection accuracy and robustness in a complex environment are improved, unmanned inspection in a high-risk scene is realized, the labor cost is reduced, the safety is guaranteed, and the method is suitable for automatic inspection in the fields of industrial equipment, public facilities and the like.

The invention discloses a remote sensing small target detection method and device, and relates to the field of remote sensing image target detection, and the method comprises the steps: replacing a CIoU loss function in a baseline algorithm with a FocalWIoU loss function, and determining an improved baseline algorithm; the FocalWIoU loss function comprises a FocalWIoU loss function and a WIoU loss function; a modulation factor is introduced into the Focal loss function, and the weight of the sample is adjusted; the WIoU loss function measures the similarity between the target frames; training an improved YOLOv8 model by using the improved baseline algorithm; a GAM module is added behind a second context sensing cross-layer fusion module of a neck network in the improved YOLOv8 model; the GAM module comprises a channel attention sub-module and a space attention sub-module; the channel attention sub-module is used for retaining cross-latitude information; the space attention sub-module is used for fusing space information; and the remote sensing small target in the remote sensing image is detected by using the improved YOLOv8 model. The target detection precision and the detection performance of the remote sensing small target are improved.

The invention relates to the field of radar target detection, in particular to a millimeter-wave radar target detection method, and solves the problem of low detection precision of miniaturized and low-power millimeter-wave radar waves in the prior art. The method comprises the following steps: S1, calibrating a to-be-tested environment; s2, the radar works in a first working mode in a low-frame-rate state, and whether the radar is the same target or not is determined; whether mode switching is carried out or not is determined; s3, switching the working mode of the radar to a second working mode of a high frame frequency state, and if the target can be detected within T3 time, entering step S4; and S4, the radar acts after detecting the target in the second working mode. According to the invention, the average power consumption of radar detection is greatly reduced by using multi-mode switching; the environment to be detected is partitioned, interference in the environment is recognized, the anti-interference performance is improved under the condition of low power consumption, and the detection precision and stability are improved.