87 results about "Detector head" patented technology

The invention discloses a lightweight human body posture estimation method based on PLES-YOLO, and belongs to the field of computer vision, and the method comprises the following steps: carrying out the feature extraction of an input image through employing an improved down-sampling module, and obtaining an initial feature map; based on the initial feature map, a multi-scale feature fusion module is adopted to carry out progressive feature enhancement, and multi-scale fusion features are obtained; based on the multi-scale fusion features, performing feature optimization by adopting a spatial pyramid pooling module combined with an attention mechanism to obtain weighted multi-scale features; and on the basis of the weighted multi-scale features, a lightweight shared convolution detection head is adopted to predict human body key point coordinates, and a final attitude estimation result is output.

The invention discloses a fine-grained representation remote sensing image interpretation method and system based on Shearlet priori knowledge. The method comprises the following steps: performing high-frequency sub-band information extraction on preprocessed remote sensing image data by adopting Shearlet transformation; a Laplacian pyramid filter and a shearlet filter are used for obtaining a high-frequency assembly and multidirectional high-frequency sub-bands obtained after the high-frequency assembly is split in the direction; sub-bands of the same level are adaptively fused through a space attention mechanism, and a detail-enhanced feature map is generated; carrying out multi-scale and multi-level weighted fusion on high-frequency component prior extracted through Shearlet transformation and a deep residual block obtained by a deep residual network; outputting an accurate remote sensing image analysis result by using the trained detail enhancement backbone and different detection heads of a plurality of tasks; through combination of Shearlet transformation and an advanced deep learning architecture, the accuracy and efficiency of remote sensing image interpretation are significantly improved, good model generalization ability and practicability are maintained, and it is predicted that remote sensing image processing will be promoted to develop towards a more intelligent and automatic direction.

An example method of training a detector head for object detection of a training object category based on a frozen vision and language model (VLM) is provided. The method includes receiving the frozen VLM pre-trained on a plurality of image-text pairs. The method includes determining, for an image embedding generated by a pre-trained image encoder of the frozen VLM and by the detector head, a detection region embedding indicative of one or more regions of interest in an image. The method includes generating, by a pre-trained text encoder of the frozen VLM, a text embedding of the training object category. The method includes predicting, by the detector head and based on the detection region embedding and the text embedding of the training object category, an object from a target object vocabulary associated with the training object category. The method includes providing the pre-trained frozen VLM and the trained detector head.

An example method is provided for training a detector head to perform object detection on a training object category based on a frozen vision and language model (VLM). The method includes receiving the frozen VLM that is pre-trained over a plurality of image-text pairs. The method includes determining, for an image embedding generated by a pre-trained image encoder of the frozen VLM and by the detector head, a detection region embedding indicative of one or more regions of interest in an image. The method includes generating, by a pre-trained text encoder of the frozen VLM, a text embedding of the training object category. The method includes predicting, by the detector head, an object from a target object vocabulary associated with the training object category based on the detection area embedding and the text embedding of the training object category. The method includes providing the pre-trained frozen VLM and a trained detector head.

The present invention discloses a wind turbine blade detection device and detection method provided with a buckling assembly, which relates to the technical field of wind turbine blade detection, and comprises a fixed bottom plate, the top surface of which is symmetrically mounted with a slide rail, the slide rail is slidably connected with a buckling mechanism, the buckling mechanism is provided with two groups, and the inside of the two groups of the buckling mechanism is connected with a wind turbine blade body. The present invention drives the base, the buckling mechanism and the wind turbine blade body to slowly move from the bottom of the ultrasonic detection head by using a servo motor to drive and rotate the first screw rod, which is conducive to the ultrasonic detection head to effectively detect the wind turbine blade body, and in the process of the buckling mechanism moving, it can also drive the second connecting seat and the transmission rack to move synchronously, realize the driving of the flushing power mechanism, and enable the flushing power mechanism to link the sliding flushing mechanism and the dust suction mechanism to realize the cleaning of the surface of the wind turbine blade body.

The invention discloses a multi-modal adaptive fine-grained feature fusion method based on an unmanned aerial vehicle visual angle, and the method comprises the steps: carrying out the feature extraction of a visible light image and an infrared image collected by an unmanned aerial vehicle through a single-modal feature extraction branch, and obtaining multi-scale feature maps of two modals; based on the multi-scale feature maps of the two modalities, the weight is adjusted by calculating the local region structural similarity of the modal features, and an adaptive fusion weight matrix is constructed; performing fine-grained feature weighted fusion on the multi-scale feature maps of the two modalities by using an adaptive fusion weight matrix, matching the weight matrix with the spatial size of the feature maps through an interpolation method in the fusion process, and then performing weighting operation to obtain a fused feature map; and based on the fused feature map, firstly aggregating multi-scale features through the neck of the detector, then executing target classification and bounding box regression through the head of the detector, and outputting a target classification result and a bounding box regression result. The method can adapt to a complex detection scene of the visual angle of the unmanned aerial vehicle.

The invention relates to the technical field of infrared optics, in particular to a large-area-array long-rear-working-distance infrared optical system which comprises a head cover, a first lens, a second lens, a third lens, a fourth lens and a detector, the head cover serves as an optical window, the specific surface curvature radius and thickness design of the head cover optimizes the light incidence and preliminary convergence effect, and the light intensity is improved. The structure of the four lenses is reasonably distributed according to positive and negative focal powers to form a stable optical structure, and the diffraction surface design of the first lens can effectively correct chromatic aberration and improve the imaging quality. Meanwhile, the second emergent surface, the third incident surface, the fourth emergent surface and the fifth incident surface all adopt aspheric surface design, aberration is further reduced, and high-resolution imaging of the large-area-array detector is guaranteed. In addition, through reasonable spacing arrangement between the elements, the long rear working distance of the system is realized, and the requirements of adapting to detectors of different models are met. In conclusion, the system successfully realizes an imaging target with a large area array and a long rear working distance.

The utility model relates to a throw-in type oil-water mixture transmitting controller which comprises a transmitting meter head and a detection head and further comprises a detection head shell, an installation counter bore is formed in the front end of the detection head shell, an insulation gasket is fixedly connected to the bottom of the installation counter bore, the rear end of the detection head is embedded into the insulation gasket, a containing groove is formed in the bottom of the installation counter bore, and the containing groove is fixedly connected with the detection head shell. The signal amplification chip is arranged in the containing groove, a signal line at the rear end of the detection head penetrates through the insulation gasket in a sealed mode and then is electrically connected with the signal amplification chip, and the signal amplification chip is electrically connected with the transmitting meter head through a cable penetrating out of the rear end of the detection head shell in a sealed mode. According to the utility model, the signal amplification chip and the detection head are placed in the detection head shell together, and then are connected to the transmitting meter head through the cable, so that even if the measuring range is very long and the length of the cable exceeds 5 meters, the measurement is still very stable.

The present invention discloses a target detection method and system based on angle adaptive fusion. The method includes: extracting features from an input image through a backbone network, and outputting corresponding first feature maps from the output results of convolutional layers at different stages; performing lateral connection dimensionality reduction and top-down fusion processing on the original feature maps through a feature pyramid network to obtain second feature maps; performing target multi-angle feature extraction and fusion processing on the second feature maps through an angle adaptive fusion pyramid network to obtain third feature maps; classifying and regressing the third feature maps through a detector head network to output target detection results. The present invention has a wide range of applications and high computational efficiency, and can be widely applied to the field of artificial intelligence technology.

The utility model relates to the technical field of measuring devices, and discloses a mining geology crack measuring device which comprises a crack measuring instrument, a first sound wave detection head and a second sound wave detection head are inserted in the left side of the crack measuring instrument, and a first installation base is arranged outside the second sound wave detection head in a sleeved mode. A locking knob is inserted into the rear end of the first mounting seat, a sliding tooth block is connected to the bottom of the first mounting seat, and a measuring support is connected to the bottom of the sliding tooth block. According to the utility model, through the arrangement of the multifunctional detection structure, the first sound wave detection head and the second sound wave detection head are installed in the first installation seat and the second installation seat at the two sides of the measurement support in use, and the first installation seat and the second installation seat are stretched at the two sides of the measurement support through the sliding tooth blocks; the first sound wave detection head and the second sound wave detection head are arranged on the two sides of the geological fracture, the scales on the measuring support can measure the widths of the geological fracture in different stages, and therefore the multifunctional detection function is achieved.

1. The name of the design product: explosive detector detection head protection shell. 2. The use of the design product: the design product is used for explosive detector detection head protection shell. 3. The design points of the design product: in shape. 4. The picture or photo that best indicates the design points: perspective view.

The utility model relates to the technical field of spectrometers, and discloses a spectrograph with a protection structure, which comprises a spectrograph, a self-resetting protection assembly is arranged at the left end of the spectrograph and is used for realizing automatic closing for protection, and a supporting and stabilizing assembly is arranged at the bottom of the spectrograph and is used for realizing automatic closing for protection. And the supporting and stabilizing assembly is used for ensuring the use stability. According to the spectrograph with the protection structure, the self-resetting protection assembly is installed, a user needs to grasp a movable rod or limit the movable rod on the right side of a limiting block in the using process, however, limitation of the limiting block exists, and in order to be vibrated, the self-resetting protection assembly can fall off and be automatically closed to achieve the protection effect; the problem that the detection head is damaged due to the fact that the detection head falls off from the hand and directly collides with the detection head in the using process is avoided, automatic reset can be effectively achieved through the self-reset protection assembly to achieve the protection effect, and the protection cover is automatically reset after being used to avoid forgetting to close the protection cover.

A modular laser encoder system is disclosed which comprises one or more laser units (2) and one or more detector heads (4), with any one of the laser units (2) being connectable in use to any one of the detector heads (4) to provide laser light to the connected detector head (4) and with the connected detector head (4) using the laser light to generate a detection signal, wherein a signal level (S) associated with the detection signal varies depending on which combination of laser unit (2) and detector head (4) is used, and wherein each of the laser units (2) and detector heads (2) is associated with one or more coefficients (c1, c2) which aim to normalise a response of that laser unit (2) or detector head (4) such that a corrected signal (S) is derivable from the detection signal by a signal processing unit (50) using the coefficients (c1, c2) associated with the laser unit (2) and detector head (4) of the combination concerned, with the corrected signal (S) having a signal level that is substantially consistent regardless of what combination of laser unit (2) and detector head (4) is used.

The invention provides a blade creep test device, and relates to the technical field of structural design. The test device comprises a furnace cylinder, a temperature measuring ring, a temperature sensor and a first control assembly, the furnace cylinder is provided with an accommodating space, and the accommodating space is used for accommodating a to-be-detected blade; a resistance wire is arranged on the inner wall of the furnace barrel; the temperature measuring ring is arranged in the furnace barrel and sleeves the periphery of the to-be-detected blade, and the temperature measuring ring can reciprocate along the axial direction of the furnace barrel so as to align the temperature measuring ring with a to-be-detected area of the to-be-detected blade; the temperature sensor is arranged on the temperature measuring ring and can reciprocate along the radial direction of the temperature measuring ring, so that a preset distance is kept between a detection head of the temperature sensor and the to-be-detected blade; the first control assembly is electrically connected with the resistance wire and the temperature sensor, and the first control assembly is used for receiving an actual temperature value detected by the temperature sensor, comparing the actual temperature value with a preset temperature value, and controlling the resistance wire to be heated or cooled when a difference value between the actual temperature value and the preset temperature value exceeds a preset range.

The utility model discloses a wearable eye myopia prevention physical therapy device which comprises an adjusting assembly and a mounting disc arranged on the front side of a mounting plate, a mounting groove is formed in the inner side of the mounting disc, a fixing block is arranged on the inner side of the mounting groove, a telescopic rod is arranged on one side of the fixing block, a movable block is arranged at one end of the telescopic rod, and the movable block is arranged on the other end of the telescopic rod. During eye massage, the detection head detects the eye contour of a user, the control box controls the telescopic rods to stretch out and draw back, the telescopic rods arranged in a plurality of mounting grooves in front of the mounting disc drive the moving blocks to adjust the positions in the mounting grooves, and the moving blocks are driven by the control box to move back and forth. And after the moving block finishes moving according to the eye contours, a massage rocker on a fixed disc on the front side of the moving block starts to rotate to massage the eyes of the user, so that the device can be adjusted according to the eye contours of different users, the situation that the massage rocker cannot effectively massage the users is avoided, and the applicability of the device is improved.

The invention discloses a screw type air compressor rotor flaw detection device, and relates to the technical field of air compressor rotor detection, the screw type air compressor rotor flaw detection device comprises a workbench, a rotating shaft and a plurality of spiral tooth surfaces arranged on the side wall of the rotating shaft, and further comprises a detection assembly arranged on the workbench and used for carrying out flaw detection on each spiral tooth surface, the flaw detection assembly comprises a detection ring connected to the two workbenches in a sliding mode, a plurality of detection pipes are connected to the side, close to the spiral tooth surfaces, of the detection ring in a sliding mode, and an ultrasonic flaw detection head is arranged at the end, close to the spiral tooth surfaces, of each mounting frame. According to the flaw detection device for the rotor of the screw-type air compressor, through the arrangement of the detection assembly, synchronous flaw detection testing on a plurality of spiral tooth surfaces is realized, and meanwhile, under the action of the control assembly, the ultrasonic flaw detection head moves on the surfaces of the spiral tooth surfaces in a path with the same distance with tooth roots; therefore, the conditions of the tooth surface and the interior can be reflected more accurately.

The utility model relates to the technical field of constructional engineering, and discloses a structural crack detection device for a high-rise building, which comprises a base and a detector body, universal wheels are fixedly connected with the bottom of the base in a front-back and left-right symmetrical manner, a height adjusting mechanism is arranged at the top of the base, and the height adjusting mechanism comprises a fixed seat. The fixing base is fixedly connected to the top of the base, and the top of the fixing base is fixedly connected with a fixing cylinder. The servo motor drives the worm to rotate, the screw rod is driven to rotate through meshing transmission of the worm and the worm wheel, and the lifting plate can drive the lifting column to stably move up and down under the constraint action of the limiting rod, so that the height of the detection head body is accurately adjusted. The high-rise building crack detection device can meet the detection requirements of structural cracks of high-rise buildings at different heights, is stable and reliable in operation, avoids errors possibly caused by manual adjustment, ensures that the detection head can be accurately aligned with a detection area, and improves the reliability of detection data.

The beam parameter measurement device for boron neutron capture therapy provided in this application includes: a detector; a beam inlet window provided on the side of the test chamber facing the beam outlet, with the detector head extending into the interior of the test chamber; a moving mechanism on which the detector is mounted for adjusting the position of the detector head inside the test chamber; an adjustment mechanism on which the test chamber is mounted for adjusting the position of the test chamber to make the geometric center of the test chamber coincide with the center of the beam; and a control system including a data acquisition module and a motion control module. The data acquisition module is electrically connected to the detector for acquiring beam data output by the detector, and the motion control module is electrically connected to the moving mechanism and the adjustment mechanism for controlling the moving mechanism and the adjustment mechanism. This application can reliably adjust the position of the detector inside the test chamber, and can also adjust the position of the test chamber to ensure that the geometric center of the test chamber coincides with the center of the beam, thereby improving the accuracy of the measurement data.

The invention relates to the technical field of optical cable fault positioning, in particular to a communication optical cable fault positioning device and method based on the computer technology, and the device comprises a housing, a pressing rolling device disposed in the housing, and a marking device disposed at the front end of the housing. The abutting rolling device comprises a rotating disc, an abutting wheel used for dynamically clamping the optical cable, a double-head motor and a red light detection head. According to the communication optical cable fault positioning device based on the computer technology and the method thereof, after a fault optical cable is connected with a red light emitter, the device slides along an overhead suspended optical cable, and an abutting wheel abuts against the optical cable to enable the optical cable to be bent and fully expose the fracture position of the optical cable, so that a red light detection head receives red light at the fault fracture position; and then the rubber band is released through the marking device to sleeve the fault position, so that the fault fracture position is accurately positioned under the condition that the optical cable is not dismounted, the condition that a worker mistakenly dismount a normal optical cable due to inaccurate positioning is reduced, and the fault fracture positioning efficiency and accuracy are improved.

The invention relates to the technical field of steel rail flaw detection, in particular to a welded rail intelligent flaw detection device and method.The welded rail intelligent flaw detection device comprises a mechanical arm, and a flaw detection assembly is arranged at the front end of the mechanical arm and comprises an ultrasonic detection head and a reciprocating deflection assembly; the reciprocating deflection assembly drives the ultrasonic detection head to deflect back and forth between a first deflection position and a second deflection position so as to switch the detection direction; and the coupling assembly is used for conveying a coupling agent to the detection surface. The structure of the flaw detection device is optimized, and the detection area can be enlarged through the reciprocating deflection structure of the ultrasonic detection head, so that the detection coverage is improved; under driving of the mechanical arm, the flaw detection assembly can be driven to advance along the scanning path more automatically and intelligently, flaw detection of the steel rail is completed accurately, and therefore flaw detection efficiency and result accuracy are guaranteed.

Methods and systems are provided for improving the quality of an image generated by a single photon emission computed tomography (SPECT) imaging system (100, 200). Photons penetrating a protective shield (580) around a detector head (308, 408, 210) of the SPECT system (100, 200) may not be correctly processed by a reconstruction algorithm used to reconstruct an image from a plurality of projection views (701) acquired by the SPECT imaging system (100, 200), thereby generating artifacts (1121, 1112, 730) and reducing the quality of the image. In one embodiment, the artifacts (1121, 1112, 730) may be reduced or eliminated by selectively applying the reconstruction algorithm to a first partial column of pixels (1007, 1012) of the projected views (701) that does not include the artifacts (1121, 1112, 730), and not applying the reconstruction algorithm to a second partial column of pixels (1007, 1012) of the projected views that includes the artifacts (1121, 1112, 730). The number of the second partial pixel columns (1007, 1012) may be based on a sweep angle (550) of a detector (471) that acquires the projected view (701).

Systems and methods for controlling motion of detectors having moving detector heads are provided. One system includes a gantry and a plurality of detector units mounted to the gantry, wherein the plurality of detector units are individually movable including translational movement and rotational movement. The system further includes a controller configured to control movement of the plurality of detector units to acquire Single Photon Emission Computed Tomography (SPECT) data, wherein the movement includes both the translational movement and the rotational movement coordinated to position the plurality of detector units adjacent to a subject.

1. Name of the product in this design: Detector head of a Doppler underwater measuring device. 2. Application of this design: A device for emitting sound waves into water and receiving scattered signals to calculate the velocity, direction, and profile distribution of water flow. The probe head of the Doppler underwater measuring device is used to assemble the Doppler underwater measuring device. 3. The key design feature of this product is its shape. 4. The picture or photo that best illustrates the key design points: Design 1 front view. 5. Design 1 is designated as the basic design. 6. Other situations requiring explanation: In Design 1, the part depicted by the dashed line is not protected, while the part drawn by the solid line is protected; in Design 2, the part covered by red is not protected.

Systems and methods for controlling motion of detectors having moving detector heads are provided. One system includes a gantry and a plurality of detector units mounted to the gantry, wherein the plurality of detector units are individually movable including translational movement and rotational movement. The system further includes a controller configured to control movement of the plurality of detector units to acquire Single Photon Emission Computed Tomography (SPECT) data, wherein the movement includes both the translational movement and the rotational movement coordinated to position the plurality of detector units adjacent to a subject.

The utility model discloses a multi-beam depth sounding device, which comprises a sleeve seat, a sound velocity profiler main body arranged on the inner side of the sleeve seat, a protective cover arranged at the bottom end of the sound velocity profiler main body, the protective cover is of a cylindrical structure, the protective cover is provided with a hollowed-out opening around the outer side of the circumference, and the protective cover is connected with a protective net through the hollowed-out opening. The inner wall of the top end of the sleeve seat is fixedly provided with an electric telescopic rod, one side, adjacent to the electric telescopic rod, of the inner wall of the top end of the sleeve seat is fixedly connected with a guide rod in a surrounding manner, and the inner side of an opening in the bottom end of the sleeve seat is provided with a cutting knife. It is effectively guaranteed that the detection head cannot be collided and damaged in water; weeds on the outer side of the withdrawn protective cover can be cut off and removed through the cutting knife arranged on the inner side of the bayonet of the sleeve seat, so that the structure is simple, mounting and dismounting are convenient, and later maintenance is convenient.

The invention relates to an integrated composite detector based on infrasound audio vibration. A collection detector integrating three sensors is composed of an infrasound sensor, a vibration sensor, a microphone and a collection control calculation circuit board. A Beidou antenna, a 4G antenna and an LR-WIFI antenna are installed in the detector head, head solar energy serving as an auxiliary power supply is installed on the surface of the head of the detector head, and external solar energy serving as a main power supply is installed outside the detector head. The detector head is installed on the ground nail barrel, the battery pack is installed in the ground nail barrel, and the ground nail head is connected below the ground nail barrel. A vibration sensor is installed in a ground nail head, a sound guide pipe is installed outside a detector head, a microphone is connected with the sound guide pipe, and sound waves enter the microphone from the sound guide pipe. According to the invention, good detection of homologous infrasonic waves, audio sound waves and vibration is realized.

The invention discloses an interval supporting movable position type semiconductor detection device, and relates to the technical field of semiconductor detection. The detection device comprises a chamber top cover; the top of the monitoring window is an integrated quartz window supported by the monitoring frame; the supporting frame body is supported on the top of the monitoring window at intervals through a supporting piece and is used for movably supporting more than one probe seat; the supporting piece comprises an end supporting piece, a first central supporting piece and a second central supporting piece; the first central lug is fixedly connected with the first central supporting piece, and the second central lug is fixedly connected with the second central supporting piece; the probe seat comprises a cylinder body connecting body and a connecting lug block. A detector head, a probe seat and the like of the monitoring equipment are prevented from being influenced by different deposition temperature fluctuations of the cavity; the probe seat can be installed and positioned at any position, the test position can be positioned at any position, and various sizes of substrates and tray changes are compatible.

This invention discloses a detection end stress detection device based on an infrared sensor, belonging to the technical field of infrared stress detection devices. It includes a detection device body, a detector head fixedly connected to one side of the detection device body, and mounting plates symmetrically fixedly connected to the side of the detection device body near the detector head. A frame is fixedly connected between the mounting plates and the detection device body. An obstruction removal mechanism is provided on the frame, including a vertical plate fixedly connected to the upper surface of the upper mounting plate. This mechanism can blow away floating objects and impurities from the infrared illumination path and the surface of the object to be detected. This not only eliminates propagation obstacles of infrared radiation, ensuring that radiation can smoothly reach the detector end and avoiding problems such as signal strength weakening, signal distortion, and background interference, thereby improving measurement accuracy, but also reduces the influence of external factors on the measurement, keeping the illumination path clean and unobstructed, reducing error sources, and improving measurement stability.

The utility model relates to the technical field of building crack detection, and discloses a building engineering crack detection device which comprises a sliding seat, the lower surface of the sliding seat is fixedly connected with an electric telescopic rod, the output end of the electric telescopic rod is fixedly connected with a supporting plate, and the lower surface of the sliding seat is provided with a supporting assembly. The supporting assembly is used for connecting the sliding seat and the supporting plate, four clamping plates are rotatably connected to the interior of the supporting plate, the outer walls of two of the clamping plates are fixedly connected with first adapters, the outer walls of the other two clamping plates are fixedly connected with second adapters, and a rotating shaft is fixedly connected to the interior of the supporting plate. According to the utility model, by pulling the clamping plates on the two sides, the first adapter and the second adapter rotate in a matched manner, so that the insertion block slides out from the interior of the mounting plate, and the sound wave detection head can be dismounted, thereby solving the problem that most of existing detection devices are inconvenient to dismount and mount the sound wave detection head.

Methods and systems are provided for increasing a quality of images generated by a single photon emission computed tomography (SPECT) imaging system. Photons penetrating a protective shielding around detector heads of the SPECT system may not be handled correctly by a reconstruction algorithm used to reconstruct an image from a plurality of projection views acquired by the SPECT imaging system, generating artifacts and reducing a quality of the image. In one embodiment, the artifacts may be reduced or eliminated by selectively applying the reconstruction algorithm to a first portion of pixel columns of the projection views not including the artifacts, and not applying the reconstruction algorithm to a second portion of pixel columns of the projection views including the artifacts. A number of the second portion of pixel columns may be based on a sweep angle of a detector acquiring a projection view.