55results about How to "Solve drift" patented technology

The invention discloses a real-time relative pose estimation method in an unmanned aerial vehicle-unmanned vehicle cooperative system, and the method comprises the steps: 1, achieving the real-time pose estimation of each platform through visual feature point tracking and IMU pre-integration through a monocular camera and IMU data, and constructing a sparse feature point cloud map; step 2, combining laser radar point cloud data with an IMU, and applying a Fast-LIO algorithm and an IESKF to complete dense three-dimensional point cloud map construction of each platform; and step 3, fusing the data in the step 1 and the step 2, constructing a factor graph optimization problem through point-surface matching residual error, visual reprojection residual error and IMU pre-integration residual error factors, and solving the relative pose in real time. According to the method, vision and laser radar information are integrated, the pose estimation precision and robustness are improved, the method is suitable for an unstructured environment, and the collaborative operation efficiency is enhanced.

The invention provides a diabetic retinopathy image recognition method based on federated learning and differential privacy, and belongs to the field of federated learning. The method comprises the following steps: carrying out preprocessing and Poisson sampling on a diabetic retinopathy image data set to initialize a global model located in a central server; the local client obtains parameters of the global model from the central server and constructs a local model; the local client performs training and differential privacy processing on a local model by using local data; the central server receives the parameters uploaded by the clients for secure aggregation, and updates the global model by using the parameters after secure aggregation; if the number of training iterations does not reach the threshold value specified by the user, a result is transmitted to the local model to continue the training process; and when the number of training iterations reaches a threshold value specified by the user, terminating a further iteration process to obtain a trained model. And the problems of gradient leakage and data loss possibly existing in privacy protection by introducing differential privacy in federated learning are balanced.

The invention discloses a Non-IID federated learning method based on dynamic knowledge distillation and dynamic regularization. The method comprises the steps that after a local model of a client is trained, distribution Ps of local data tags is output; uploading local model parameters to a server; the server gathers a group of client model parameters obtained by random sampling, and outputs global model parameters for the next round of use; the client downloads global model parameters, loads the global model parameters to a local model and outputs global data label distribution Pt; performing dynamic knowledge distillation at the client, integrating Ps, Pt and a dynamic function d (r), and calculating a local loss function of the client; and updating the local model of the client through back propagation. According to the method, the convergence speed and accuracy of the model can be remarkably improved under the condition of non-independent and identically-distributed Non-IID client data.

The invention is suitable for the technical field of metallurgical processes, and provides a continuous casting, continuous rolling and continuous straightening method for a steel rail. Comprising the following steps that a continuous casting billet is produced through a continuous casting machine; performing scanning, typical defect judgment and alarm on the surface of the continuous casting blank through a 2D surface defect detection system so as to obtain the position of a defect on the surface of the casting blank; carrying out follow-up targeted cleaning according to the surface defect position of the continuous casting billet; the continuous casting billet is subjected to induction supplementary heat treatment; the continuous casting billet is subjected to descaling treatment through high-pressure water; rolling the continuous casting billet by using a rolling mill to obtain an initial steel rail; marking the initial steel rail; waste heat quenching treatment is conducted on the initial steel rail, and deformation limiting is conducted on the rail bottom / rail head of the steel rail; straightening the initial steel rail by using a straightening machine; the initial steel rail is subjected to rough fixed-length cutting treatment of any length; the cut initial steel rail is subjected to quality detection; and performing fixed-length cutting treatment on the initial steel rail subjected to quality detection to obtain a finished steel rail.

The invention relates to the technical field of laser cutting, and discloses a polymer film full-automatic laser cutting machine which comprises a machine frame, horizontal rails for a movement adjusting module to move are horizontally installed on the two sides of the upper end of the machine frame, the movement adjusting module is located above the machine frame, and the movement adjusting module comprises a cantilever. An X-axis moving assembly is fixedly arranged on the cantilever, the X-axis moving assembly acts on the laser cutting head module, and the cantilever is provided with an inner rail for the laser cutting head module to move horizontally. According to the full-automatic laser cutting machine for the polymer film, the multiple sets of adsorption boxes are gathered together and spliced to form the adsorption area corresponding to the pattern range, the matrix structure layout is adopted, splicing is conducted according to the pattern range or the size of the film, negative pressure in each adsorption box is very uniform, each area of the film can be evenly adsorbed, and the cutting efficiency is improved. The laminating effect of the film is improved, and the laser cutting effect is also improved.

The invention discloses a multi-source data-based mining equipment positioning system and method. The system is characterized in that a plurality of ultra-wideband base stations are sequentially mounted in a crossheading roadway at intervals; the coding identifiers are mounted on supporting columns of the hydraulic supports in a one-to-one correspondence manner, and are positioned on one side close to the coal mining machine; the inertial sensing unit, the camera, the ultra wide band positioning module and the controller are all mounted on the body of the coal mining machine; the visual angle of the camera faces a supporting column of the hydraulic support. The ultra-wideband positioning module is positioned on one side, close to a crossheading roadway, of a coal mining machine body and is connected with the ultra-wideband base station; the controller is respectively connected with the inertial sensing unit, the camera and the ultra-wideband base station. The method comprises the following steps: carrying out initial position calibration by using the ultra-wideband positioning module; predicting the position of the coal mining machine based on inertial navigation; updating the prediction state of the inertial sensing unit based on the identification of the coding standard by the camera; and the positioning continuity is ensured by fusing the data of the inertial sensing unit and the camera. According to the system and the method, real-time and accurate positioning operation of the coal mining machine in the working process can be realized.

The invention provides a CT data correction method, electronic equipment and a CT machine. The CT data correction method is used for pet CT machine scanning data processing and comprises the steps that actual scanning data at each measurement angle in a CT machine are collected in real time; acquiring endpoint channel data from the actual scanning data; judging whether the channel data is shielded or not; actual measurement current data and actual measurement channel data of the shielding channel are obtained; determining a function relationship between the actually measured current data and the actually measured channel data; obtaining correction current data of an unshielded channel; and substituting the correction current data into the function relation to obtain standard channel data of the shielding channel. The problem that the CT value of a pet CT image drifts due to the fact that current feedback of a pet CT machine is not timely is solved, and the accuracy of CT scanning is improved. In the image generation process, a CT image which is more practical can be formed, and misdiagnosis caused by CT value drift is prevented.

A visual odometer initialization method based on point cloud multilayer sampling is suitable for an unmanned system of unmanned aerial vehicles (UAV) and unmanned ground vehicles (UGV) cooperative tasks, and comprises the following main steps: firstly, performing feature point matching in multiple frames of images by using a KLT algorithm, and constructing a sliding window; secondly, through 2D-2D feature association and a PnP algorithm, constructing a scale-free SFM point cloud based on the feature point data; then, IMU data and the SFM point cloud are fused, and the pose and the scale are calculated through pre-integration; thirdly, performing vision-inertia alignment by applying a multi-level sampling method to obtain a coarse registration pose; and finally, through a BA optimization structure, carrying out fine adjustment to obtain an accurate pose. The method can effectively improve the pose estimation precision and efficiency in an unknown environment, and has an important application value.

The utility model discloses a high-precision servo electric cylinder which comprises an electric cylinder body, a piston rod is arranged at the front end of the electric cylinder body, a guide enhancing assembly is connected to the front end of the piston rod, a transmission assembly is fixedly installed on the outer surface of the rear end of the electric cylinder body, and the guide enhancing assembly comprises a fixing plate, a guide rod, a fixing block and a guide ring. The transmission assembly comprises a worm gear, a rotating shaft, a transmission box, a servo motor, a driving shaft and a worm. According to the high-precision servo electric cylinder, the worm in the transmission assembly is a single-head worm, and the lead angle of the worm is smaller than the equivalent friction angle, so that a worm gear cannot reversely drive the worm, mechanical self-locking is achieved, stability is improved, radial loads can be shared through the arranged guide enhancing assembly and an external guide rail, the lateral force of a ball screw is reduced, and the service life of the ball screw is prolonged. And the bending moment resistance and the vibration resistance are obviously improved (particularly suitable for long-stroke or high-unbalance-load scenes), so that the precision is improved.

The invention relates to the technical field of network security and information security, and particularly discloses an internal threat detection method based on entity behavior analysis, which comprises the following steps of: firstly, converting a multi-source heterogeneous log into a unified event semantic vector, and constructing a double-flow neural network architecture: extracting a long-term portrait vector from a long-period history by utilizing LSTM (Long Short Term Memory), and in the other path, a short-term intention vector is extracted from the neighbor session by utilizing Transform. And the short-term intention is projected to the semantic space of the long-term portrait to carry out consistency residual calculation, so that the hidden slow poison attack is effectively identified. In order to solve the problem of misinformation caused by feature dimension stability difference, a variance estimation and anisotropy measurement mechanism is introduced, portraits are modeled as probability distribution, variance vectors are utilized to carry out reliability weighting on residual errors, and global entropy regularization is combined to dynamically adjust penalty sensitivity, so that the probability distribution of the portraits is determined. Therefore, legal concept drift and malicious behaviors can be accurately distinguished in a complex dynamic scene.

The present invention relates to the technical field of game controller calibration, and in particular to a method and system for automatically calibrating a game controller drift fault, which can effectively improve the efficiency and accuracy of game controller drift calibration and enhance the gaming experience of players. The method comprises: obtaining joystick operation information and joystick cursor detection image information; performing cluster analysis on the joystick cursor detection image information according to the joystick operation information to obtain a cursor detection image set of the joystick in a non-action state and a cursor detection image set of the joystick in an operation state; calculating the distance between the cursor and the image center on each image according to the cursor detection image set of the joystick in a non-action state to obtain a static drift distance of the joystick; dividing the game controller into pre-set drift intervals, summarizing the number of static drift distances falling within each drift interval, and obtaining a joystick drift feature vector; and performing weighted calculation on the joystick drift feature vector.

This invention provides an intelligent analysis system for photoelectric detection spectral data based on deep learning, belonging to the fields of photoelectric detection, spectral data analysis, and deep learning application technology. The system includes core units such as a spectral data acquisition unit, a preprocessing unit, a feature extraction unit, and a deep learning analysis unit, as well as auxiliary units for anomaly detection and data storage, forming a closed-loop architecture of acquisition-processing-analysis-output-optimization. It improves data quality through step-by-step preprocessing algorithms, reduces computational load by adaptively selecting key features, deeply mines spectral patterns using a model that integrates convolutional neural networks and attention mechanisms, achieves rapid scenario adaptation through transfer learning, and dynamically optimizes model parameters based on a feedback mechanism. The system solves the problems of low accuracy, poor adaptability, and long training cycles in traditional spectral analysis, achieving accurate, efficient, and intelligent spectral data analysis, and is suitable for various photoelectric detection scenarios such as qualitative and quantitative analysis of substances and environmental monitoring.

The present invention discloses a method for predicting silicon content in molten iron based on TD-LightGBM, and belongs to the technical field of industrial process monitoring. The method comprises: first, using the decision tree model structure of a lightweight gradient booster to judge the importance of the feature according to the number of times the feature is used to segment training data; then, combining a time difference algorithm to establish a dynamic feature extraction model; finally, using a local weighted algorithm to establish a local model based on the Euclidean distance between samples. Predicting the silicon content in molten iron during the ironmaking process based on this model can effectively remove redundant items and drift phases in the blast furnace environmental parameters. Compared with the existing prediction scheme for silicon content in molten iron, the present invention further improves the accuracy of the prediction.

The present invention provides a single-boiler segmented fire tube indirect heating boiler, comprising a boiler shell, an economizer is arranged at one end of the boiler shell, and a water inlet pipe is connected to one side of the economizer; the economizer and the boiler shell are connected by a connecting pipe; the outer wall and / or inner tube of the boiler shell and the economizer are arranged with a fire tube bundle as a heating surface. The present invention has the advantages of simple structure, convenient manufacturing, maintenance and overhaul, and can ensure sufficient heat exchange area, reduce exhaust gas temperature, improve thermal efficiency, and realize integrated superheated steam production without superheating tubes.

This invention relates to the field of grain dryer technology, specifically to a grain dryer discharge mechanism with an anti-breakage structure. The mechanism includes a main body with a discharge chamber at its bottom, connected to a grain receiving hopper at the bottom; a frame unit located within the discharge chamber, comprising a first frame slidably connected to the inner wall of the discharge chamber, and a second frame slidably nested in the middle of the first frame; both the first and second frames have connecting grooves on their outer walls, with the two sets of grooves facing opposite directions; a partition plate fixed inside the discharge chamber above the frame unit, with several discharge ports extending through its top; and a drive unit located within the discharge chamber. This invention enables low-breakage discharge throughout the entire grain dryer process, single-drive multi-mode controllable discharge, and also helps to break up knots and prevent damage. The purely mechanical structure ensures stable operation and is suitable for drying various grains.

The application discloses a kind of gas-liquid two-phase leakage ultrasonic safety detection method and device, it is related to mechanical engineering and nondestructive testing technical field, the method includes: deployment multichannel ultrasonic sensor array;Dynamic noise baseline based on adaptive Gaussian mixture model is constructed;Extract amplitude attenuation rate, center frequency offset and other multidimensional acoustic characteristics and calculate difference vector;Combining dynamic threshold and multi-probe time delay-amplitude information carries out leakage preliminary judgment and positioning;Through convolution-long short-term memory hybrid neural network, identify leakage mode and implement three-level risk early warning according to leakage rate.The system is composed of multichannel ultrasonic sensor array, signal processing module, feature extraction unit, positioning solution module and intelligent early warning engine.The application can realize high sensitivity detection, strong working condition adaptability, accurate positioning and quantification and closed-loop intelligent response, significantly improve the intrinsic safety level of gas-liquid two-phase flow system.

The invention discloses a molecular logic gate based on a DNA tile self-assembly result. The molecular logic gate comprises an input signal, a signal conversion unit and an output signal, wherein the input signal comprises at least two input chains; the signal conversion unit comprises a multi-chain cross-type element formed by complementary pairing of bases, and an output signal is whether the DNA tile is successfully assembled into the nanotube or not. According to the method, a complex strand displacement reaction network is simplified into intuitive structure formation / failure judgment, so that the reliability and readability of molecular logic operation are greatly improved, and a foundation is laid for realizing higher-level molecular calculation system integration.

The application discloses a power magnetic core common mode inductor and a preparation method thereof, and belongs to the technical field of electronic component manufacturing, and comprises a winding execution unit, a magnetic circuit assembly unit and a material flow transfer module, wherein the winding execution unit comprises a main shaft rotating motor and a wire arranging servo module, the magnetic circuit assembly unit comprises a servo press and a glue injection nozzle, and the material flow transfer module is connected with the winding execution unit and the magnetic circuit assembly unit. The scheme physically increases the average distance of interlayer conductors through staggered span actions, changes the distribution direction of electric field vectors, effectively reduces the parasitic capacitance between windings, directly improves the high-frequency characteristics of the common mode inductor, expands the high-frequency impedance bandwidth, and enables the common mode inductor to more effectively inhibit high-frequency noise transmission.

This invention relates to the field of agricultural aviation intelligent equipment and precision spraying technology, and particularly to a method for the formulation and spraying control of chemical adjuvants in agricultural drones, comprising the following steps: Step S1, acquiring flight status parameters of the agricultural drone and three-dimensional topographic data of the crop canopy below, and constructing a real-time aerodynamic flow field model superimposed with the downwash airflow of the agricultural drone rotor and the environmental wind field; Step S2, predicting the sedimentation trajectory of the pesticide droplets based on the real-time aerodynamic flow field model, calculating the surface tension and dynamic viscosity required for the droplets to reach the target canopy based on the sedimentation trajectory, and generating an instantaneous mixing ratio including the main pesticide solution and various adjuvants; In this invention, by constructing a real-time aerodynamic flow field model and directly sending the instantaneous mixing ratio of the main pesticide solution and various adjuvants to a microfluidic mixing device installed at the end of the nozzle for in-situ mixing, the physical space limitations and response lag of traditional agricultural drone pipeline mixing are broken.

The application discloses a preparation method of a red mud composite phase change heat storage material and relates to the technical field of material preparation process optimization. The method constructs a historical database containing temperature, power and product latent heat, calculates basic heating power based on data quality, collects current heating data in real time, calculates an inertia factor representing heat diffusion capacity and a temperature rise factor representing specific heat capacity characteristics, constructs a thermal response feature vector to quantify the composition similarity between the current batch and the historical record, dynamically determines the real-time target temperature for precise control by screening high-similarity high-quality historical records, and updates the database with new data. The application indirectly reflects the composition difference of red mud raw materials through thermal response characteristics, solves the problem of process window drift caused by composition fluctuation, realizes self-adaptive temperature control without complex composition detection, and significantly improves the phase change latent heat and batch consistency of the product.

The application discloses a kind of fusion modal and time sequence attention multi-modal physiological feature joint identification method and system, it is related to biomedical detection and artificial intelligence identification cross technical field, the method includes: control multispectral light source module emits stimulating light signal and probe light signal to the eye of measured object, real-time acquisition original image signal of the eye of measured object, calculate pupil shrinkage peak speed, local blood oxygen change rate and response time difference;Double parallel processing is carried out, and weighted physiological features and weighted time sequence features are respectively generated;Weighted physiological features and the weighted time sequence features are input into multi-modal weighted fusion layer, and after fusion, comprehensive feature vector is obtained, comprehensive feature vector is input into classification prediction layer, the confidence probability that measured object belongs to non-normal physiological response category is calculated, according to the comparison result of confidence probability and preset judging threshold, the identification result of non-normal physiological response category or normal physiological response category is output.

The invention discloses an automatic cleaning and maintenance control method and system for water quality monitoring equipment and a medium, and belongs to the field of water quality monitoring. The method comprises the steps that pure water generated by a built-in pure water production assembly of equipment serves as a measurement anchor point, and whether sensors are normal or not is judged on the basis of comparison results of detection numerical values of the sensors in a pure water environment and a preset pure water threshold value; monitoring operation state parameters of the pure water production assembly, and dynamically adjusting and updating the pure water threshold value corresponding to each sensor; when the value of the sensor deviates from the threshold value, first-stage pure water flushing is executed; if not, second-stage maintenance is started according to the sensor type, the optical sensor performs optical path self-cleaning or backwashing, and the electrochemical sensor performs electrode cleaning; and detecting again after maintenance, and locking the sensor and generating a fault alarm if the sensor is still not recovered. Through dynamic threshold configuration, progressive maintenance and sensor adaptive cleaning, the problems of reference drift and deep pollution are solved, the data credibility is improved, and the operation and maintenance cost is reduced.

This invention provides a method for measuring and locating marine station structures suitable for silty coastal zones. The method includes acquiring multi-source monitoring data for the silty coastal zone; extracting deformation feature points of the marine station structure from the multi-source monitoring data and generating planar displacement maps and elevation displacement maps of the feature points relative to a reference period for each monitoring cycle; calculating spatial offsets based on the planar displacement maps and elevation displacement maps to generate geological deformation distribution maps for each monitoring cycle and mapping them to generate a foundation stability coefficient matrix; and fusing the foundation stability coefficient matrix for each monitoring cycle with historical positioning benchmarks to generate structural safety positioning coordinates. This invention can improve the accuracy and anti-interference capability of structural displacement monitoring in complex geological environments.

This utility model discloses a high-precision touch testing system for curved screens, comprising a worktable. An adaptive curved screen leveling and clamping device is mounted on the top of the worktable. The worktable also includes an X-axis linear drive device, a Y-axis linear drive device, a Z-axis vertical drive mechanism, and a multi-functional stylus switching device. A copper head device that contacts the screen is mounted on the Z-axis vertical drive mechanism. A seven-degree-of-freedom collaborative robotic arm is mounted on one side of the adaptive curved screen leveling and clamping device, and an intelligent end effector is mounted at the end of the seven-degree-of-freedom collaborative robotic arm. Above the worktable are a machine vision positioning device for coarse positioning of the product under test and a visual calibration device for spatial mapping calibration of touch point positions. This utility model achieves automated leveling, curvature modeling, and quantitative testing of edge touch performance for curved screens, effectively solving the problems of poor compatibility and insufficient edge touch accuracy.

This invention discloses an active thermal management system and control method for silicon carbide devices based on aloe vera-inspired stomatal cell microvalves, belonging to the field of active thermal management technology for silicon carbide devices. The active thermal management system includes a hygroscopic phase change working fluid module and a stomatal cell-inspired intelligent control module. The hygroscopic phase change working fluid module includes a heat sink body, a hygroscopic salt solution contained within the heat sink body, and a porous permeable membrane covering the surface of the hygroscopic salt solution for basic passive heat dissipation. The stomatal cell-inspired intelligent control module includes a stomatal cell-inspired microvalves array and a detection and control unit, simulating the diurnal rhythm and stress response mechanism of biological stomata. By adjusting the number of microvalves opening and closing, the water vapor discharge rate is precisely controlled, thereby controlling the heat exchange rate. This invention significantly reduces the average junction temperature of silicon carbide devices and greatly suppresses temperature cycling stress, effectively delaying thermomechanical fatigue failure.

This invention relates to the field of chemical reaction equipment technology, specifically to a formaldehyde oxidation reactor with an easy-to-clean airflow distribution plate. The reactor includes a reactor body, an inlet chamber, an airflow distribution plate, and a cleaning system. The cleaning system includes a gas collecting hood disposed within the inlet chamber, a gas delivery mechanism that drives the gas collecting hood to rotate around the axis of the reactor body while adhering to the bottom surface of the airflow distribution plate, and a high-pressure pulse source located outside the reactor body. By setting a sliding gas collecting hood within the inlet chamber and cooperating with the external high-pressure pulse source, online cleaning is achieved without shutting down the reactor or lowering the temperature. This transforms traditional top mechanical scraping into a bottom-directed jet, eliminating the interference and jamming risks caused by mechanical scrapers moving between high-temperature, dense air caps, and solving the problems of flow deviation and localized overheating caused by pore blockage.

The invention relates to the technical field of cable manufacturing, in particular to an intelligent tension regulation and control method and system in a cable twisting process, and aims to solve the problems of low regulation and control precision, response lag and weak interference resistance in the prior art. The method comprises the following steps: firstly, introducing a multi-wire tension coupling coefficient according to cable specifications, material flexural rigidity and the number of stranded layers, calibrating a tension reference value by combining real-time environment temperature and humidity, and synchronously extracting roughness and correcting friction parameters through optical characteristics of wire surfaces so as to adapt to equipment speed; training historical data by using an LSTM (Long Short Term Memory) time sequence model, predicting the tension trend in the future 2-5 seconds in real time, performing classified regulation and control according to deviation statistical characteristics, and performing micro-step rotation speed regulation by a servo motor and torque compensation execution by a magnetic powder brake; updating the model when the error exceeds the threshold value or the specification is changed, and periodically finely adjusting the reference value by combining the quality data. The system comprises seven functional modules, the tension control precision and the anti-interference capability can be improved, the adaptability is high, and the cable quality and the production efficiency are effectively improved.

The application provides a fuse determination method, device, equipment and medium for an energy storage power station. The application relates to the technical field of lithium battery safety of an energy storage power station. The fuse determination method comprises the following steps: determining, based on calibration parameters of an energy storage battery, an expected discharge duration of the energy storage battery under the condition that cumulative discharge charge of the energy storage battery reaches a preset state of charge variation; discharging the energy storage battery within the expected discharge duration under the condition that the energy storage battery is in multiple states of charge, so as to collect current and voltage output by the energy storage battery and obtain voltage sequences and current sequences of the energy storage battery in the multiple states of charge; predicting, based on the voltage sequences and the current sequences, current when the energy storage battery is subjected to external short circuit, so as to obtain short circuit current of the energy storage battery in the state of charge; and determining the model of a fuse of the energy storage power station according to the structure of the energy storage power station and the short circuit current of the energy storage battery in the multiple states of charge.

The invention provides a four-piece glass-plastic hybrid lens which is composed of a diaphragm, a first lens, a second lens, a third lens, a fourth lens and a plane mirror which are sequentially arranged from the object side to the image side in the optical axis direction, the object side face of the optical axis area of the first lens is a convex face, the image side face of the optical axis area of the first lens is a convex face, and the object side face of the optical axis area of the second lens is a convex face. One of the first lens and the fourth lens is a glass lens, and the four-piece glass-plastic hybrid lens meets the following relational expression: f1'gt; 0, f3'lt; 0, f4'gt; 0, f5'gt; 0; 31t; 31t; f1 ' / f3'lt; 0, 1lt; f1 ' / f4'lt; 5, 0.5 lt; f ' / f12'lt; 2,-1.5 lt; f ' / f34'lt; 0.8,-1lt; f12 ' / f34'lt; 1; 0.15 lt; iH / TTLlt; 0.18, 0.18; according to the four-piece glass-plastic hybrid lens, through the arrangement of the structures and parameters of the first lens, the second lens, the third lens and the fourth lens, the temperature of the lens of the four-piece glass-plastic hybrid lens is out of tilt; therefore, the focus of the lens is stable.

The invention belongs to the technical field of photoelectric detectors, and particularly discloses a fixing clamp of a four-quadrant APD assembly automatic test system, which comprises a test seat and a test cover, a tested piece is arranged in a central cavity of the test seat, the test cover is arranged on the test seat, an outer rotating ring and an inner rotating ring are arranged on the test cover, and the outer rotating ring is connected with the inner rotating ring. Rotating handles are arranged on the outer rotating ring and the inner rotating ring, a mounting seat is arranged in the testing cover, an optical fiber patch cord extends into the mounting seat during testing and is fixed on the mounting seat through an FC flange plate, the outer rotating ring pushes the mounting seat to press downwards to tightly press a tested piece when rotating, an eccentric hole is formed in the inner rotating ring, and the eccentric hole is communicated with the FC flange plate. And when the inner rotating ring rotates, the optical fiber patch cord is driven to eccentrically rotate to complete quadrant switching. According to the fixing clamp, a precise mechanical alignment and fixing unit is adopted, the problem of light path drifting caused by temperature circulation is solved through a rigid structure, the stability of the light spot position in the whole testing process is guaranteed, therefore, high-repeatability testing data is obtained, and the accuracy and reliability of a testing system are improved.