124results about How to "Reduce the impact of detection accuracy" patented technology

The invention discloses a plant pest and disease detection method based on multiple features and a support vector machine, including the specific steps of preprocessing a plant image collected by an unmanned aerial vehicle, extracting the color feature, HSV feature, texture feature and shape feature of the image, and inputting the features to a support vector machine to detect possible pests and diseases in the process of growth. The plant pest and disease image is detected by combining multi-feature fusion and the support vector machine. For different plants and different pests and diseases, plant pest and disease detection can be carried out using the method. Moreover, the method is real-time, accurate and practical.

The invention discloses a perimeter intrusion detection method, device and system. The Doppler microwave radar sensor is set on the enclosure in advance, the method includes: collecting the I / Q channel signal generated by the Doppler microwave radar sensor according to the received reflection echo, the I channel signal is an in-phase component, and the Q channel signal It is an orthogonal component; the collected I / Q channel signal is analyzed and processed for intrusion; according to the processing result, a warning message is sent to the terminal server, and the warning message is used to instruct the terminal server to alarm or eliminate the alarm. In this scheme, the electromagnetic waves emitted by the Doppler microwave radar sensor are used for intrusion detection. Since the electromagnetic waves are less disturbed by climate and environmental factors, the impact of climate and environmental factors on the detection accuracy of enclosure intrusions can be effectively reduced.

The invention discloses a battery management system and a balanced state online monitoring method thereof. The battery management system comprises a micro-control unit, a battery voltage detection circuit controlled by the micro-control unit, a battery voltage balancing circuit controlled by the micro-control unit and a battery. The battery management system has the remarkable advantages of being capable being supported to be a large current balancing system; reducing the effect of balancing current on the voltage detection accuracy; being capable of completely separating voltage detection and battery balancing by aid of the framework design mode that single harness and dual channels are adopted for balancing harnesses and voltage detection harnesses which do not affect each other functionally, and meanwhile, not affecting complexity of system harness distribution due to the fact that the single harness and dual channel design is adopted for the harnesses and the quantity of the harnesses is as same as that of the original harnesses performing balancing by aid of the voltage detection harnesses; and being capable of monitoring balancing circuit impedance in real time and accordingly finding whether the balancing harnesses are opened or damaged timely.

The invention discloses a mobile robot high-precision processing method based on visual servo. The method comprises the following steps: relative pose relations T1TL and BTL of a laser tracker coordinate system to a workpiece local target coordinate system and a workpiece overall coordinate system are calibrated, and a relative pose relation T1TB of the workpiece local target coordinate system tothe workpiece overall coordinate system is calculated; local target points are shot to determine a relative pose relation T1TC of the workpiece local target coordinate system to a visual coordinate system; a relative pose relation R1TB of a robot base coordinate system to the workpiece overall coordinate system is calculated through robot kinematics; errors of an actual position of a milling terminal and a theoretic position are updated in real time; and each joint corner of a robot is compensated through robot inverse kinematics. The method adopts a visual system for large component processing reference conversion, mobile robot precise positioning and robot processing precision compensation to realize high-efficiency and high-precision processing of large components in such fields as aerospace, rail traffic and energy.

The invention discloses a calibration pattern of a camera and a calibration and detection method based on the same. In the invention, the center of a standard circle is taken as a characteristic point of a calibration pattern; and an equivalent center at a specific region is calculated, thus the algorithm is simple, the noise resistant property is strong, and even though an image is distorted (e.g. a circle is distorted into an ellipse), the centroid position of the image can also be correctly extracted. In addition, in the calibrating process, the object capable of being detected is arranged in a rectangular region delineated by calibrating circles and simultaneously the calibration and detection are completed, thereby reducing the influence of environment variation on the accuracy of detection. The surface of the calibration pattern is provided with eight calibrating circles, and is provided with a rectangular region, and the surface degree of the calibration pattern is different from the circular degree of the calibrating circles.

The invention discloses an aspheric optical element surface shape detection device, which is used for carrying detection on surface shape information of an aspheric optical element, and comprises an interferometer and a reflector, a phase shifter, a standard mirror, a compensating mirror and an aspheric optical element, which are arranged in sequence from the bottom up in the vertical direction. The interferometer is used for emitting a measuring beam in the horizontal direction; the reflector is used for converting the measuring beam in the horizontal direction into a measuring beam in the vertical direction; the phase shifter is used for adjusting the phase of the measuring beam in the vertical direction; the standard mirror is used for providing a reference plane to reflect the measuring beam in the vertical direction; the compensating mirror is used for compensating the surface shape of the aspheric optical element; and the measuring beam reflected by the aspheric optical element and the measuring beam reflected by the reference plane of the standard mirror form interference fringes in the interferometer. The aspheric optical element surface shape detection device improves surface shape detection accuracy of the aspheric optical element.

The invention discloses a vehicle classification method for double-row vertical radar waves. A microwave vehicle detector is installed on the side of the road. The microwave vehicle detector adopts an array type double microwave radar. The microwave vehicle detector is composed of a vector decomposition module and a frequency analysis module. and a microprocessor. The method of the present invention is a comprehensive identification and analysis of the echo characteristics, analyzing the characteristic value of the echo to obtain a group of typing results, analyzing the frequency of the echo, and then obtaining a group of typing results; finally, the The two classification results are weighted to obtain the final vehicle classification result. Because this product adopts the two-level small network classification method in the neural network, it can greatly improve the classification speed, solve the real-time problem of vehicle classification, and improve the accuracy and stability of classification.

Provided is a method for realizing synchronous characterization of the mechanical properties of graphene based on the fold formation principle. Two identical PMMA / graphene composite films are made. The graphene in the composite films is placed on a test substrate and on a target substrate respectively through technologies such as a dry graphene transfer technology. After cooling treatment under the same conditions, folds are generated due to stress mismatch because the thermal expansion confidents of the test substrate and the target substrate are different from that of graphene. Finally, themechanical properties of graphene are evaluated by measuring the wavelength and amplitude of graphene folds. The method can be carried out synchronously with a graphene process, and thus, the elasticmodulus, surface stress, number of layers and other mechanical properties of graphite can be characterized simply and quickly.

The invention relates to a method and a system for realizing remote sensing image target detection based on a deep neural network, and a storage medium thereof, so as to realize detection of horizontal and rotary arrangement targets of a remote sensing image. According to the method, an anchor point box generation module is designed, an anchor point box is generated in a self-adaptive mode throughfeature information of different positions, and the influence of the difference of preset anchor point boxes on detection precision is reduced; aiming at the characteristic that more small targets exist in a remote sensing image, an improved feature pyramid structure is provided, and deep and shallow layer feature information is fused by adopting a transposed convolution method; aiming at difficulties such as complex background of a remote sensing image, a receptive field expanding module is adopted to extract more characteristic information, and the detection precision of a small target under a complex background is improved; a SmoothLn function is adopted as regression loss, so that the algorithm performance is further improved; for a rotation arrangement target, regression of an anglefactor is introduced to realize rotation frame detection. In addition, in order to facilitate the use of a user, the remote sensing image target detection system designed by the invention has the functions of horizontal frame and rotary frame detection and result statistics.

The invention provides a road scene other vehicle detection system and method based on a space-time domain multi-dimensional fusion technology, and the method comprises the steps: obtaining a projection matrix through employing calibration parameters of a camera and a radar, fusing radar point cloud data and camera image data through the projection matrix according to the distances and different tones, and forming multi-dimensional fusion image data; respectively detecting the camera image data and the multi-dimensional fusion image data by adopting two YOLOv3 models, calibrating a detection frame of the target vehicle, forming a first annotation picture and a second annotation picture, and performing training to obtain a first target detection model and a second target detection model; carrying out tracking and data association on the detection frame of the target vehicle, and outputting a detection result after screening. Target tracking is achieved through data association in combination with the time domain information of the image sequence frames to perform other vehicle detection, so that the method is closer to an actual road scene, the influence on the other vehicle detection precision caused by the mobility of vehicles in the road scene is reduced, and the detection precision and robustness are improved.

The invention discloses a vehicle driving state detection method. The method comprises the steps: obtaining a monitoring video stream, carrying out vehicle detection on the monitoring video stream, and obtaining a target detection frame of a current vehicle in a plurality of monitoring images in the monitoring video stream; determining a vehicle movement track vector according to the target detection frame positions of the Nth frame of monitoring image and the Mth frame of monitoring image in the monitoring video stream, wherein N is smaller than M, and the difference between M and N is largerthan a preset value; calculating a vector included angle between the vehicle movement track vector and a preset traffic flow direction vector; and if the vector included angle is greater than a preset included angle, judging that the current driving state of the vehicle is reverse driving. According to the invention, the detection accuracy of the vehicle driving state can be improved. The invention further discloses a vehicle driving state detection device, electronic equipment and a storage medium, which have the above beneficial effects.

The invention discloses a mobile equipment human body pose estimation method based on three-dimensional skeleton extraction. The method comprises the steps: putting a human body video into a light-weight human body skeleton recognition model, and obtaining 2D articulation points; mapping the obtained 2D articulation points through a regression device constructed by a neural network to obtain 3D articulation points in a space; calculating an interlimb joint angle of the photographed person according to the obtained coordinate information of the 3D joint points; judging the pose of the photographed person and the photographing position of the photographing equipment according to the inter-limb joint angle of the photographed person and the position information of the 3D joint points; and mapping 2D human body articulation points obtained by the lightweight human body skeleton recognition model to 3D human body articulation points in a three-dimensional space through a deep neural networkregression device, thereby increasing the dimension of geometric information of the 2D human body articulation points. The limb included angle is calculated based on the geometrical relationship of the 3D human body articulation points, and the pose of the human body and the shooting position of the mobile equipment are effectively predicted according to the limb included angle and the position relationship of the articulation points.

The invention belongs to the technical field of chemical ingredient detection, and particularly relates to a method for detecting multiple kinds of vitamin simultaneously. According to the detecting method, multiple lipid-soluble vitamin samples contained in a sample can be fast and accurately measured simultaneously through a two-dimensional liquid chromatographic analysis method. According to the method, effluent in a specific period of a one-dimensional chromatographic column is injected to a two-dimensional chromatographic column, effluent of the two-dimensional chromatographic column is subjected to vitamin D measurement, the other effluent of the one-dimensional chromatographic column is subjected to measurement of non-vitamin D lipid-soluble vitamin, peakoverlap of vitamin D and other lipid-soluble vitamin can be avoided, and the influence of lipid matter in the sample on the detection accuracy of vitamin D is effectively reduced. According to the detecting method, the normal-phase column purifying process on vitamin D in a traditional method can be omitted, tedious detecting steps are simplified, detection time is saved, and high accuracy can be guaranteed.

The invention relates to a pesticide detection device with an extraction function for agricultural product detection. The pesticide detection device with the extraction function for the agricultural product detection comprises a main body, an extraction mechanism and a disinfection mechanism, wherein the extraction mechanism comprises a rotating sleeve, a supporting sleeve, a bottom plate, a piston block, a first driving assembly, a connecting hole, a filter screen, a sealing assembly and two air cylinders; the disinfection mechanism comprises a second driving assembly, an ultraviolet lamp tube, a connecting ring, a cleaning brush and two supporting plates; and the sealing assembly comprises a cover plate, two sliding rods and two springs. According to the pesticide detection device with the extraction function for the agricultural product detection, a mixture of agricultural products and an extracting solution can be grinded through the extraction mechanism, and then residues of the agricultural products and the extracting solution are separated, the extraction of pesticides in the agricultural products is realized, the detection efficiency of a detector is improved, bacterium onthe agricultural products can be killed through the disinfection mechanism, the impact of the bacterium on the detector is reduced, and the accuracy of the detector is improved.

The invention relates to a river levee side slope landslide detection device, and belongs to the field of geological disaster detection equipment. The river levee side slope landslide detection devicecomprises a ground pile implanted into slope ground along a direction perpendicular to a horizontal plane and a detection box arranged at one end of the ground pile, which faces away from the ground;a datum rope is suspended inside the detection box; a through-beam type photoelectric switch and a power supply are also arranged inside the detection box; a transmitter and a receiver of the through-beam type photoelectric switch are respectively arranged at both sides of the datum rope; when the detection box is not shifted, the datum rope is positioned between the transmitter and the receiverfor blocking off light; the through-beam type photoelectric switch is coupled with an alarm mechanism; the power supply provides required electric energy for the through-beam type photoelectric switchand the alarm mechanism; and when the detection box is inclined, the alarm mechanism sends out an alarm signal. The river levee side slope landslide detection device disclosed by the invention replaces artificial cyclic detection to carry out real-time inspection, so that when labor cost is saved, detection accuracy is improved.

The invention discloses a rapid testing device for pavement deflection. The rapid testing device comprises a dragging type test vehicle and a shock absorption system; a steel beam bracket is arranged in a compartment of the test vehicle; a testing steel beam is longitudinally arranged on the steel beam bracket; a plurality of Doppler laser sensors are arranged on the testing steel beam; and the shock absorption system comprises three shock absorption systems, wherein a first shock absorption system is a shock absorption system for testing the chassis of the vehicle, a second shock absorption system is a servo control system which is used for adjusting the gesture of the testing steel beam in real time, and a third shock absorption system comprises a shock-isolating cushion arranged between the steel beam bracket and the bottom of the compartment and a damping cylinder arranged on the steel beam bracket. With the adoption of the rapid testing device, the testing efficiency is increased, the time and the human resources of road testing departments are saved, and the testing is carried out on an express way at a normal running speed, so that the personnel safety of workers is effectively ensured; meanwhile, no lanes are occupied or the traffic is not closed, so that the road resources are saved, and the effective utilization ratio of a road when the deflection testing is carried out is improved.

The invention discloses a control device for a gas water heater, the gas water heater and a calculating method for the water flow of the gas water heater. The control device comprises a front ultrasonic simulation end, a time-to-digital converter, a microprocessor as well as a first ultrasonic sensor and a second ultrasonic sensor which are oppositely arranged on two sides of a water inlet pipe of the water heater; the front ultrasonic simulation end is in controlled connection with the first ultrasonic sensor and the second ultrasonic sensor and is connected with the time-to-digital converter; and the microprocessor is connected with the front ultrasonic simulation end and the time-to-digital converter. With the control device, the obtained accuracy of the water flow can be improved, and the service life of the water heater can be prolonged.

An ultrasonic flow rate detection system comprises a system control part, a circuit part and an assistant device, wherein the system control part comprises a single chip, a digital signal processor, and an on-site programmable gate array; the circuit part comprises a switching circuit, an automatic gaining control amplifying circuit, an analog / digital converting circuit, and a digital / analog converting circuit; the assistant device comprises a storage part, an input device, a display device a power supply and an ultrasonic transducer; the on-site programmable gate array is connected with the input device and the ultrasonic transducer respectively, and used for receiving the input instruction of the input device and conducting time control on the ultrasonic transducer; the digital signal processor is connected with the storage part, and used for data information exchange; the on-site programmable gate array is in bi-directional connection with the single chip and the digital signal processor respectively; the single chip is connected with the display device, and used for displaying data information through the display device.

The invention relates to a hot forming cylindrical workpiece geometrical parameter detection device. The device comprises a tail rod for a bottom thickness test, an LVDT sensor for the bottom thickness test, a bottom thickness test contact and a hole depth test device. The detection method is as follows: a workpiece to be detected is mounted on a main shaft, a positioning cylinder loosens to enable a tail supporting cylinder to act, a rotary pressing head moves forward and presses the workpiece to be detected, the LVDT sensor for the bottom thickness test tests bottom thickness, the main shaft is rotated, an LVDT sensor for a wall thickness difference test tests wall thickness difference, an LVDT sensor for a hole depth test measures the bore diameter of an inner bore of the workpiece, if a measured value is greater than a specified value, the hole depth is judged to be smaller than the specified value, the workpiece is a defective product, and if any value of the bottom thickness, the wall thickness difference and the hole depth is not up to standard, the value is displayed on a display touch screen, and an alarm is given. The detection method adopts clamping for one time, wall thickness difference and bottom thickness numerical values of two or more points are automatically tested, whether the hole depth is up to standard is judged, the automation degree is high, labor intensity is lowered, and detection efficiency is improved.

The invention relates to a loop heat pipe ammonia working medium purity analysis device. The loop heat pipe ammonia working medium purity analysis device is composed of a high-purity carrier gas steel bottle, a first high-purity ammonia steel bottle, a second high-purity ammonia steel bottle, a gasification chamber, a constant-temperature control system, ball valves, a pressure valve, an electromagnetic valve, a re-pressing valve, a check valve, pressure gauges, a flow meter, a vacuumizing set, a waste gas collecting and processing device, a gas chromatograph and a computer. The high-purity carrier gas steel bottle is connected with the gas chromatograph through a carrier gas pipeline. The first high-purity ammonia steel bottle is connected with the gas chromatograph through a purging pipeline, the first ball valve, the electromagnetic valve, the first pressure gauge and the flow meter. The second high-purity ammonia steel bottle is connected with the gasification chamber through a sample inlet pipeline. The vacuumizing set is connected with the gasification chamber through the check valve. The gasification chamber, the vacuumizing set and the gas chromatograph are connected with the waste gas collecting and processing device through discharge pipelines. By means of the integrated design of the device, the detection rate of impurity gas in loop heat pipe liquid ammonia is increased, air inside the pipelines is reduced in the gas sample inlet process, and the measurement accuracy is improved.

The invention discloses a gas analysis and detection device in the environment, which comprises a casing, an air inlet is arranged on the top of the casing, a valve matching the air inlet is arranged on the top of the air inlet, and a filter is arranged above the inside of the casing. The filter screen divides the interior of the shell into a filter chamber and a detection chamber. The top of the filter chamber is located on the side of the air inlet with a suction fan, and the bottom of the filter chamber is provided with a filter mesh. The filter chamber and the detection chamber are detachable. Connection, the bottom of the filter chamber is symmetrically provided with inserts on both sides, and the top of the insert is provided with a stopper. The stopper is located at the bottom of the filter chamber, and the diameter of the stopper is larger than the diameter of the insert. A slot matching the plug is provided at the corresponding position, and an insertion port matching the stop is provided on the top of the slot. Beneficial effects: the filter screen can be cleaned in time conveniently, which is more conducive to the cleanness and tidiness of the gas environment inside the housing, and the detection head can fully contact the gas in the housing, thereby greatly improving the detection accuracy.

The invention discloses an electric multi-leaf grating encoder closed-loop control system, which comprises an electric multi-leaf grating and is characterized by further comprising a code bar, an integrated sensor, an integrated circuit board and a code bar guide slot, wherein a connector arranged on the code bar is fixed with a blade of the electric multi-leaf grating, the code bar is integrally arranged in a guide slot of the code bar guide slot, and the code bar carries out linear motion along with the blade; the integrated circuit board is fixed at one surface of the code bar guide slot; and the integrated sensor is composed a circuit board and a precision sensor installed on the circuit board, and the circuit board of the integrated sensor is connected with the integrated circuit board installed at one surface of the code bar guide slot. By adopting the structure disclosed by the invention, a closed-loop detection device applying a miniature high-precision encoder and the code bar is combined with the blade of the electric multi-leaf grating, accurate control for the position of the blade is realized through data processing of the encoder, the failure rate is reduced, and a purpose of accurate position control with high stability and good durability is achieved.

The invention relates to a method for measuring a titanium element or titanium dioxide content in a diacetate cellulose tow. In the method, a dynamic colorimetric method is adopted, a continuous flow analyzer is used for detecting, and an external standard method is used for quantifying. The method has the advantages of being simple, fast, high in detection safety, accurate in detection result, small influence of personnel and environment, suitable for multi-batch detection demands of factories, etc.

The invention discloses a human body action recognition method. According to the method, firstly, minimum neighborhood construction and filtering preprocessing are carried out on an image, then imagechannel transformation, target contour enhancement and differential image extraction are carried out, threshold processing and foreground image processing are carried out on a foreground image, and finally model training or action recognition and action positioning are carried out on the basis of a three-dimensional convolutional network. According to the method, the problem that in an existing action recognition method, the detection precision of a model is reduced in a large scene, a small target and a complex background is solved, and meanwhile action detection and action positioning in anycontinuous borderless video stream are realized, the precision of human body motion recognition and the robustness in different application scenes are improved, and the normative application capability of the model is improved.

The invention relates to a monitoring system and method for acquiring coal mine underground dust concentration based on images, and belongs to the technical field of coal mines. The monitoring system comprises a walking track, a detection unit and a control unit, wherein the detection unit and the control unit move along the walking track; a camera, an illuminometer, a hygrometer, a light source set, a data processing unit and a control unit are arranged on the walking track; the data processing unit is connected with the control unit, and the control unit is connected with the camera, the illuminometer, the hygrometer and the light source set; the walking track is composed of two parallel tracks which are arranged along the two sides of the side wall of the roadway respectively; and the camera is installed at the height of a breathing zone of an operator so as to obtain the dust concentration near the breathing zone of the operator. Compared with traditional coal mine underground dust concentration fixed-point monitoring, the monitoring system and the method can obtain coal mine underground particulate matter concentration space distribution, and a foundation is laid for coal mine underground dust concentration global monitoring.

The invention discloses a high-precision temperature monitoring circuit which comprises a temperature sensor, a resistor R1, a resistor R2, a triode, a resistor R3 and an optical coupler U, wherein after the resistor R1 and the resistor R2 are connected in series, a non-public end of the resistor R2 is connected onto the temperature sensor; a non-public end of the resistor R1 is connected onto a power supply; public ends of the resistor R1 and the resistor R2 are connected with a base electrode of the triode; an emitting electrode of the resistor R1 is connected onto the power supply; a collector electrode is connected onto the resistor R3, and the other end of the resistor R3 is connected onto an input end of the optical coupler U; a pull-up resistor R4 is connected onto the output end of the optical coupler U; and the temperature sensor is connected with the power supply at the same time. The high-precision temperature monitoring circuit has the advantages as follows: the output of the optical coupler is isolated, common-mode voltage on a line is restrained, and the monitoring precision is improved.

The invention belongs to the technical field of fluorescent detection, and particularly relates to a pulse width modulation-based fluorescent detection and optical-to-digital conversion system. The pulse width modulation-based fluorescent detection and optical-to-digital conversion system comprises an optical current detection circuit, a dark current detection circuit, a reference voltage source, a wireless transmitting module and a wireless receiving end, wherein the wireless receiving end consists of a wireless receiver and an MCU (Micro Controller Unit). Detection of output current of a photodiode is divided into a resetting phase, a heavy current detection phase and a low current detection phase. At the resetting phase, an integrator and a ramp signal generator are reset to initial states. At the heavy current detection phase, reference voltage of a comparator is kept at a fixed level by a reference voltage control signal Sel. At the low current detection phase, the reference voltage of the comparator is controlled to be a ramp signal by the Sel signal. By using the pulse width modulation-based fluorescent detection and optical-to-digital conversion system, conversion from a fluorescent signal to a digital signal at low power consumption and transmission and reception of fluorescent detection data can be realized, and the detection accuracy and the dynamic range of the fluorescent signal by the circuit are improved.

The invention discloses an automobile sideslip inspection bench. A side sliding plate comprises a first side sliding plate and a second side sliding plate which are symmetrically arranged. The first side sliding plate and the second side sliding plate are connected through a return lever, a return mechanism is arranged at one side of the return lever, a reset mechanism and a rapid positioning device are arranged inside the return mechanism, and an induction switch is arranged in the front of the rack. The inspection bench has the advantages that by arranging a spring return device, an electromagnet positioning mechanism and the induction switch, the side sliding plates can be rapidly returned and stabilized at the initial zero point positions in the dual-steering bridge detection process,the left-right oscillation phenomenon in the side sliding plate return process is eliminated, the influence on the rear wheel detection precision is lowered, the inspection is used for eliminating loose sliding of wheel deformation, tire deformation is eliminated before axle tire sideslip detection, and therefore the sideslip quantity detection precision is improved.

The invention discloses a gain testing method and device for a receiver, and relates to the technical field of wireless communication. The method comprises the following steps: acquiring the bandwidth, current thermal noise power and noise coefficient of a receiver; and computing the gain of the receiver according to the current thermal noise power, the noise coefficient and the bandwidth. The gain of the receiver is computed through the thermal noise power, so that errors generated due to differences of simulation devices are reduced, and the gain testing accuracy is increased; and meanwhile, the performance of the receiver is enhanced. Moreover, a noise system statistical value at a normal temperature and a temperature change compensation value are taken as the noise coefficient, so that the gain testing accuracy can be further increased.

The invention provides a liquid level sensor of a compressor. The liquid level sensor comprises a compressor housing. An oil storing cavity of freezing lubricating oil is arranged in the compressor housing. An opening is arranged in the compressor housing. The liquid level sensor comprises a housing body with an opening which is in sealed connection with the opening of the compressor housing. The liquid level sensor also comprises a filtering screen which is arranged on the inner side of the opening of the housing body. The filtering screen surrounds at least parts of inner wall of the housing body, so a detection cavity communicated with the oil storing cavity via the filtering screen is formed. The liquid level sensor also comprises a floating ball which is arranged in the detection cavity and rises and descends along with change of liquid levels of the freezing lubricating oil. At least one permanent magnet is arranged on the floating ball. The liquid level sensor further comprises an inductor which is installed on the outer wall surface of the housing body. The inductor can be easily assembled later and pre-assembled for mass production, individually changed and more suitable for small semi-enclosed and fully-enclosed compressors.