30results about How to "Solving Detection Errors" patented technology

The invention provides a method and a system for detecting the parking space states of a parking lot. The detection method comprises the following steps of acquiring a first grayscale image of a parking lot in a non-vehicle state; converting the first grayscale image into a contour image; acquiring the regional information of each parking space according to the contour image, wherein the regionalinformation comprises the regional range and the position information of each parking space; acquiring a second grayscale image of the parking lot in the current state; performing contour detection oneach parking space in the second grayscale image according to the regional information, obtaining the sum of the contour area corresponding to each of all parking spaces through calculation, whereinthe sum of the contour area is the sum of the areas of all contours in the region of a certain parking space in the second grayscale image; respectively comparing the sum of all contours with a presetjudgment threshold value to obtain the idle state of each parking space in the parking lot in the current state. According to the invention, an idle parking space can be accurately fed back to the owner of a vehicle. Therefore, the owner can conveniently find the idle parking space, and the parking time is saved. The detection accuracy and the detection timeliness are improved.

The invention relates to an outer building wall crack detection system and method based on machine vision. The system comprises a microcomputer processor, a CMOS sensor and a laser illumination system; the laser illumination system comprises four point-like laser transmitters arranged in the form of a rectangle; a transmission light beam is perpendicular to the imaging plane of a CMOS sensor; thecenter of the rectangular is positioned at the imaging center of the CMOS sensor; the microcomputer processor collects the image of the CMOS sensor in real time. According to the method, crack data isextracted via an image gray threshold processing and framework refining algorithm, the actual length corresponding to a unit pixel is acquired according to the pixel length of the rectangular formedby the extracted point-like laser, the absolute position of the current CMOS sensor under an outer wall reference system is acquired according to accumulation of relative movements between two frames,and thus the actual length of a crack and the coordinate of the crack under the outer wall reference system are acquired. According to the outer building wall crack detection system and method basedon machine vision provided by the invention, the automatic detection, size computation and coordinate positioning of the crack can be achieved, and the detection and positioning efficiencies of the crack are greatly improved.

The invention provides a surface defect detection method and device for an injection molding part, equipment and a storage medium, and relates to the technical field of computer vision. The method comprises the following steps: acquiring a product picture of an injection molding part; carrying out first preprocessing on the product picture to obtain a to-be-detected picture; and detecting the obtained to-be-detected picture by adopting a pre-trained detection model to obtain a detection result about whether the surface of the injection molding part has defects or not. According to the technical scheme, the computer vision and the neural network are combined in advance to train the detection model, and the detection model is adopted to detect the to-be-detected picture, so that whether defects exist on the surface of the injection molding part or not is quickly and accurately detected. The problem of detection errors caused by external factors such as emotion changes of detection personnel in current manual visual inspection solved, and the problems that a traditional solution is low in compatibility and cannot quickly adapt to new requirements are solved. The enterprise training and management cost is reduced, and the problem that defective goods flow into the market to cause loss to enterprises is effectively avoided.

The invention discloses a battery capacity measuring circuit, and a battery capacity accurate calculating method and system. The battery capacity accurate calculating method includes: S10, calculating resistance of a resistor; S20, calculating a current value passing across a device load; and S30, calculating current capacitance of a battery. Instead of direct use of a resistor nominal value, actual resistance of the resistor is measured and calculated at first, thereby solving the problem of terminal device battery capacity detection errors caused by inconsistency between the actual value and the nominal value on two ends of the resistor. In this way, the battery capacity calculation precision is increased, problems of wrong guidance, electric quantity jump and shutdown in advance caused by resistance errors are solved, the user experience is improved, and users are helped to correctly arrange usage of devices.

The invention discloses a switching tube control circuit. The switching tube control circuit comprises a switching tube, a differential circuit, a comparator and a control circuit, wherein the input end of the differential circuit is connected with a drain of the switching tube, the output end of the differential circuit is connected with a first input end of the comparator and used for sampling differential current generated when the drain voltage of the switching tube changes, a second input end of the comparator inputs a voltage predetermined threshold value, the comparator is used for comparing the output voltage of the differential circuit with the voltage predetermined threshold value, the output end of the comparator outputs a switching tube control signal, the input end of the control circuit is connected with the output end of the comparator, the output end of the control circuit is connected with a grid of the switching tube, and the control circuit is used for judging whether the drain voltage of the switching tube reaches a minimum value or not according to the switching tube control signal output from the comparator, and controlling the switch-on of the switching tube when the drain voltage of the switching tube reaches the minimum value.

The invention relates to a branch vehicle afflux early warning system for vehicles coming and going on a main road. The system comprises a branch vehicle detection system used for detecting branch vehicle appearance information through two-dimensional laser radar detection equipment, and a main road intelligent early warning sign used for prompting a main road driving vehicle through a detection result. The branch road vehicle detection system is arranged at the edge of a branch road and is 30-50 meters away from a branch road intersection, the two-dimensional laser radar detection equipment is vertically downward, and is installed facing a road, the main road intelligent early warning sign is arranged on a main road, is about 5-10 meters away from the branch road intersection and is arranged facing the direction of coming vehicles of the main road, and the branch road vehicle detection system is in real-time communication with the main road intelligent early warning sign in wired andwireless transmission modes. Compared with the prior art, the system has the advantages of stable, accurate, efficient and all-weather detection and the like.

The invention discloses an intelligent monitoring system and a method for mildew status of the storage tobacco leaves based on the internet of things, which includes a monitoring center data server. The environmental information parameters of tobacco leaves storage space collected by multiple data acquisition terminals are received by the monitoring center data server through a relay device, and the mildew state of tobacco leaves in the tobacco leaves storage space is intelligently identified through a tobacco leaves status recognition model based on the BP neural network algorithm, and finally the mildew state of tobacco leaves is output. The relay device is a hand-held relay device, one end of the relay device wirelessly communicates with multiple data acquisition terminals, and the other end wirelessly communicates with the monitoring center data server. At least three data acquisition terminals are placed inside each standard tobacco leaves stack, and the data acquisition terminalsare evenly placed in the plane one third of the height from the bottom of the tobacco leaves stack. The invention solves the problem that the mildew state of tobacco leaves can not be accurately determined in time due to the heavy manual detection workload, untimely and small coverage.

The invention discloses a fuel oil system of an automobile and a leakage detection method of the fuel oil system. The fuel oil system comprises a fuel oil steam recovery tank and a leakage detection device; the leakage detection device comprises a microcontroller, a detector and a pressure regulator; the microcontroller is in signal connection with a pressure sensor; an execution mechanism of the pressure regulator is in signal connection with the microcontroller, and a gas outlet of the pressure regulator is connected with the fuel oil steam recovery tank; the detector is installed between the fuel oil steam recovery tank and the atmosphere and comprises a first cavity, a second cavity and a third cavity, an air inlet of the first cavity is connected with an air outlet of the fuel oil steam recovery tank, and an air outlet of the second cavity is communicated with the atmosphere; a first valve element is arranged between the first cavity and the second cavity and closed through a driving mechanism of the first valve element; a second valve core is arranged between the third cavity and the first cavity; and the third cavity is communicated with the second cavity. According to the fuel oil system provided by the invention, the problems of detection errors and false alarms caused by reference holes are solved.

The invention relates to a vibration device and a control method thereof, electronic equipment and a storage medium. The vibration device comprises a motor, a vibration part, an acceleration detectionmodule and a control circuit, and the motor is connected to the vibration part and used for driving the vibration part to vibrate; the acceleration detection module is connected to the vibration partand is used for detecting the acceleration of the vibration part; the control circuit is connected with the acceleration detection module and the motor and used for controlling the motion state of the motor according to the acceleration detected by the acceleration detection module. The vibration detection and control precision is improved.

The invention discloses a power tester for an electric device, comprising a voltage transformer, a current transformer, an A/D (Analog-to-Digital) conversion module, a parameter input unit, a theoretical working current conversion module, a theoretical working voltage conversion module, a current average processing module, a voltage average processing module, a timing circuit, a statistic time clearing preset key, a display unit and a controller, wherein the A/D conversion module is respectively connected with the current transformer and the voltage transformer; the timing circuit is used for counting the actual working time of a tested electric device; and the controller is used for calculating the electric power and electric consumption of the tested electric device according to average processing results of the current average processing module and the voltage average processing module by combining working time mounted by the timing circuit. The power tester is reasonable in design, convenient to install and arrange, simple and convenient in use and operation, high in intelligent degree, good in using effect and capable of effectively solving various practical problems such as lower detection precision, not high intelligent degree, incapability of directly inquiring electric consumption and the like existing in the traditional electric power testing device.

The invention provides a method for detecting the quality of the parking technology of a driver. The method comprises the steps that based on a number of sensors, motion data such as a steering wheel and a brake during the driver's parking process are acquired, and a multi-modal time sequence is acquired in a time sequence manner; a neural network architecture based on a convolution variation self-encoder is used to extract the features of the multi-modal time sequence and generate a simulation sample; reconstruction probability detection is carried out based on the simulation sample; and the quality of the parking technology of the driver are judged based on the detection result. According to an intelligent recommendation method provided by the invention, based on the judgment of the quality of the parking technology of the driver, an automatic parking function or other executable functions are recommended to the driver when needed. The invention further provides a detection system for the quality of the parking technology of the driver and an electronic device, and the system and the electronic device have the same beneficial effects as the above method.

The invention discloses a fuel system of an automobile and a leakage detection method of the fuel system. The fuel system comprises a fuel steam recovery tank and a leakage detection device; the leakage detection device comprises a microcontroller, a detector and a pressure regulator; an execution mechanism of the pressure regulator is in signal connection with the microcontroller, and a gas outlet of the pressure regulator is connected with the fuel steam recovery tank; the detector comprises a first cavity and a second cavity, an air outlet of the first cavity is provided with a first communication port, the first communication port is provided with a first valve element, and the first valve element is opened and closed through a locking mechanism; in the inflating and pressure maintaining stage, the locking mechanism is in the first state, and the locking mechanism controls the first valve element to be closed; and in the normal exhaust stage, the locking mechanism is in the second state, and the locking mechanism controls the first valve element to be opened. According to the fuel system of the automobile and the leakage detection method of the fuel system, the problems of detection errors and false alarms caused by a reference hole are fundamentally solved by adopting a mode of measuring pressure and judging pressure drop.

The invention discloses a fuel oil system of an automobile and a leakage detection method of the fuel oil system. The fuel oil system is provided with a pressure sensor, and the fuel oil system comprises a fuel oil steam recovery tank and a leakage detection device; the leakage detection device comprises a microcontroller and a detection part, the detection part is installed between the fuel steam recovery tank and the atmosphere, and the detection part comprises a pressure regulator and a normally open solenoid valve; an air inlet of the normally open solenoid valve is connected with an exhaust port of the fuel steam recovery tank, and an air outlet of the normally open solenoid valve communicates with the atmosphere; and the normally open solenoid valve is in signal connection with the microcontroller, and the microcontroller is in signal connection with a pressure sensor and the pressure regulator. The problems of detection errors and false alarms caused by a reference hole are fundamentally solved by adopting a mode of measuring pressure and judging pressure drop.

The invention discloses a multi-scale feature extraction and fusion method and device, belongs to feature extraction and fusion, and aims to solve the technical problem of how to realize multi-scale feature extraction and fusion on the premise of reducing noise and redundant data. According to the method, multi-scale feature extraction is carried out through a plurality of Faster R-CNN deep learning networks to obtain a plurality of groups of different feature sets, DCA feature fusion is carried out on the plurality of groups of different feature sets, and each feature set comprises a plurality of feature vectors with the same dimension. The device is used for carrying out multi-scale feature extraction and fusion through the multi-scale feature extraction and fusion method. The device comprises a feature extraction module and a feature fusion module. Compared with a conventional feature extraction and feature fusion method, the method is better in robustness, improves the target detection accuracy, and solves a problem that a single-scale feature may cause a detection error.

The invention provides a method for improving detection near-space precision. The method comprises the following steps: under a calibration mode, recording a first calibration propelling quantity NE1 when a light blocking sheet on a moving component of an injector blocks a photoelectric switch on an injection pump, and recording an empty propelling quantity NE2 when a piston of the injector movesto the end; calculating a calibration relative quantity difference delta NE between the first calibration propelling quantity NE1 and a near-space relative quantity NE0 according to the preset near-space relative quantity NE0 of the injector, the first calibration propelling quantity NE1 and the empty propelling quantity NE2; and under a normal injection mode, recording a first propelling quantity E1 when the light blocking sheet blocks the photoelectric switch, using the difference of a real-time second propelling quantity E2 recorded when the piston of the injector moves and the first propelling quantity E1 as a working relative quantity difference E3 to be compared with the calibration relative quantity difference delta NE, and judging that the injector is propelled to a near-space state when the working relative quantity difference E3 is greater than the calibration relative quantity difference delta NE.

The invention belongs to the field of electronic equipment, and provides a label detection method and device based on double sensors and a handheld thermal printing head. The method comprises the steps of detecting a label through a first sensor and a second sensor, and outputting first current intensity and second current intensity; and obtaining a detection result of the label based on the first current intensity and the second current intensity. According to the label detection method and device based on the double sensors and the handheld thermal printing head provided by the invention, the problems of paper detection errors and label learning failures of a current handheld thermosensitive bill printing head are solved, and the paper detection precision is improved through hardware and an algorithm scheme.

The invention discloses a machine vision technology based leather detection device. The device includes a support, an industrial camera, an LED lamp and a control device; the industrial camera can take pictures of leather on a conveyor belt; the LED lamp can illuminate the leather on the conveyor belt; and the control device can obtain the area and position information of the leather by taking pictures of the leather on the conveyor belt through the industrial camera, and control a nozzle to perform spraying on the leather on a corresponding position or with a corresponding area on the conveyor belt when the leather is conveyed to a spraying chamber by the conveyor belt. A leather detection method is also disclosed. The method includes the following steps: 1, performing camera calibrationon the industrial camera, and determining the position of a spray gun and the speed of the conveyor belt being V; 2, obtaining the area of the leather or the position information of the leather on theconveyor belt; and 3, controlling, by the control device, the spray gun to perform spraying on the leather on the corresponding position or with the corresponding area. Thus, the positions of the leather can be accurately detected and measured; the scope of detecting and measuring the positions of the leather is large; and the device and method are stable and reliable in operation.

The invention discloses a multi-lithium-battery protection system, comprising a plurality of level conversion circuits, a battery module, an RC module and a detection module, wherein the battery module comprises a first battery to an x battery, the negative electrode of the first battery is connected with a vss end, and the negative electrode of the k battery is connected with the positiveelectrode of the (k-1)battery; and the detection module comprises first to x detection circuits and a control logic unit, each of the first to x detection circuits comprises a fifth resistor, a sixth resistor and a detection operational amplifier, one end of the fifth resistor is connected with one end of the sixth resistor and the in-phase input end of the detection operational amplifier, the other end of the sixth resistor is connected with the vss end, and the output end of the detection operational amplifier is connected with the control logic unit. According to the invention,level conversion is realized in virtue of the resistors and the operational amplifiers, so the voltages of all the ports of the operational amplifiers are controlled, the grid electrode of a low-voltage CMOS is conveniently used as the input of the operational amplifiers, and errors caused by current flowing into the operational amplifiers are avoided; and thus, the problem of errors of the lithium battery protection system is solved.