264results about How to "Achieve culling" patented technology

The invention discloses an optimization operation method based on energy consumption analysis for power plant and a system thereof. The method includes the following steps: a data acquisition subsystem acquires real-time operational data and carries out pre-processing on the specific real-time data; a data processing subsystem analyzes and calculates the real-time data, so as to obtain the real-time energy consumption index under the current working condition, the real-time energy consumption index under the current working condition is compared with the energy consumption index under the corresponding working condition stored in an optimal working condition database, comparison result is output to a data display subsystem, and meanwhile if the real-time energy consumption index obtained by calculation is superior to the optimal energy consumption index under the corresponding working condition, the current real-time working condition is stored to be corresponding optimal working condition; and the data display subsystem displays the real-time energy consumption index under the current working condition and the comparison result of optimal energy consumption indexes under the corresponding working conditions. The method of the invention can more accurately analyze various energy consumption differences of power plant and the reason of exceeding of energy consumption difference, and a reliable optimization scheme is provided.

The invention discloses a timber optimized truncate method and relative optimized truncate saw. First, inputting the saw data list which comprises target saw length and quality parameters into the computer; when the sawed timber that longitudinally supplied passes the position relative to the detecting device, the detecting device reads the information of sawed timber into the computer, and the computer according to the input saw data list and said information to select the optimized sawing method; the saw machine according to said method processes truncating on the timber. Two sides of said truncate machine individually have a feeding platform and a finished product classifying platform; wherein, said feeding platform has a detecting device; said finished product classifying platform has a finished product classifying device; and a waste removing device is arranged between them.

The invention discloses an SLAM-based visual perception mapping algorithm and a mobile robot. The algorithm comprises the steps of obtaining image data of a to-be-mapped area environment and rotatingspeed data of a driving wheel; performing feature extraction on a color image in the image data; after the rotating speed data is calculated and processed according to a kinematics equation, fusing and optimizing the rotating speed data and a depth image in the image data by using a method of minimizing an encoder and a re-projection error to obtain a current frame pose of the robot; screening current frame image data acquired when a robot walks in a to-be-mapped area to obtain a key frame, identifying and classifying objects in an environment by utilizing a color image of the key frame, deleting the classified dynamic objects from the key frame, carrying out sparse processing on static objects, and carrying out local loopback for correction; taking pixel points of the color image and thecorresponding depth image in the key frame as map points, storing the map points through an octree, and carrying out mapping.

The invention discloses an oil filler target recognition and positioning method based on machine vision. The method comprises the steps that license plate information of a to-be-refueled vehicle is acquired through a refueling level camera, and then stored vehicle refueling port height and fuel tank cap template information are obtained; then the tail end of the robot moves to the height of an oilfiller, and a tail-end camera collects images; graying, gray-scale equalization and histogram equalization stipulation preprocessing are carried out on the image; secondly, edge extraction is performed on the preprocessed image, false noise contours are screened by setting a connected domain area threshold value, vehicle body contour interference is screened by setting a circumscribed circle radius threshold value, and accurate extraction of real edge features of the side face of the vehicle under ambient light interference is achieved; and finally, the shape template matching method based onthe pyramid hierarchical search strategy is adopted to quickly and accurately identify the edge features of the fuel tank cap. According to the method, the vehicle oil filling port target can be quickly and accurately identified under the influence of various light rays and shadows, and the working requirement of an oil filling robot is met.

The invention discloses a classification and identification method for small light unmanned aerial vehicles and birds based on radar data. The classification and identification method for small light unmanned aerial vehicles and birds put forward by the invention is based on target measurement information acquired by radar. Through the four steps of multi-model target tracking, multi-model probability extracting, target motion pattern judging and target motion feature extracting, the target features of a target motion model such as the transformation frequency are extracted finally to distinguish between a small light unmanned aerial vehicle target and a bird target. The defect that a radar surveillance system cannot distinguish between a small light unmanned aerial vehicle and a bird in a low-altitude airspace is overcome. Target motion characteristics are extracted based on radar data to distinguish between a small light unmanned aerial vehicle target and a bird target. The identification capability of a radar surveillance system in a low-altitude airspace is improved.

The invention discloses a small bearing quality detection method, device and system based on machine vision, wherein the method comprises the steps of S1, judging whether a to-be-detected small bearing exists in a to-be-detected region, and if yes, skipping to the step S2; S2, collecting an image of a small bearing; S3, matching the template images in a database according to the collected images; S4, obtaining an image difference between the collected image and the matched template image through an image frame difference method, and analyzing the image difference to determine whether a defect region exists or not; and S5, controlling the small bearing to be conveyed to a corresponding discharging rail. According to the present invention, the small bearing quality detection, the automatic feeding and the automatic removing based on the machine vision are achieved, so that the accuracy is improved, and the purpose of rapid and automatic detection is achieved.

The invention provides a pneumatic drying method and equipment thereof for realizing large-flow tobacco material puffing drying treatment; the method comprises the following steps: puffed tobacco material is conveyed to a drying pipe for being dried by hot air flows, and then the tobacco material is separated from hot sir; the drying pipe is composed of a horizontal drying pipe and a vertical drying pipe of which the cross section is more than the cross section of the horizontal drying pipe; hot air with the temperature at 100-250 DEG C and air flow speed of10-50m / s of is sent in the horizontal drying pipe, wherein the air flow feeding direction is vertical to the material feeding direction; the material driven by hot air flows enters in the vertical drying pipe and is ascended in a suspension way, and the ascending speed of the material is reduced to 0.5-5m / s; meanwhile, sundries such as slivers with large density in the material can not be suspended and are separated pneumatically under the wind speed. By arranging the vertical drying pipe with large caliber, the speed of the material, moving upwards along the air flow direction, is reduced, on the premise of not increasing the drying route, the drying time of the material is prolonged, heat-transferring and mass transferring coefficient is improved, and the drying time of the material is improved.

The invention discloses an automatic packer. The automatic packer comprises a frame I, a hopper I fixed on the frame I, a qualified product conveying belt, a rejecting device, an unqualified product conveying belt, a frame II and a hopper II fixed on the frame II, wherein an overturning plate is arranged in the hopper I; a weight sensor is arranged on the overturning plate; the lower part of the hopper I is connected with a bagging machine; the rejecting device comprises a cylinder I and a pushing plate connected to a pushing rod; an automatic plastic packaging machine, a labeling machine and a label making machine are arranged on the qualified product conveying belt; an infrared sensor is arranged on the automatic plastic packaging machine; the unqualified product conveying belt penetrates through the frame II. According to the automatic packer, the overturning plate with the weight sensor and the pushing plate are arranged in the hopper I; the materials are fed stably; the uneven stirring and the secondary pollution caused by artistic stirring are avoided; the rejecting device is capable of rejecting unqualified materials and supplementing; the qualified rate of products is ensured; the production efficiency and the automation degree of the equipment are improved.

The present invention relates to a cotton isomerism fiber detecting and measuring device, and a detecting and measuring method thereof. The device is characterized by comprising a cotton sample network preparing device, a cotton network collecting device, a detection box and a computer; the forefront of the cotton sample network preparing device is provided with a pair of cotton transmitting rollers and a pair of drawing rollers; one of the cotton transmitting rollers and the drawing rollers is provided with an encoder which is used for recording the transmission velocity and the position of the cotton network; the detection box is positioned between the two cotton transmitting rollers and the two drawing rollers; a detection channel is positioned in the detection box and is a transparent rectangular tube; one side or two sides of the detection channel are provided with a set of image acquisition systems; the computer is respectively connected with the signal acquisition card of the encoder and the image acquisition cards of the image acquisition systems; a set of isomerism fiber image identifying and measuring software is arranged in the computer; and the computer is connected with a displaying and printing device. The device and the method can be widely used in cotton quality inspection departments, cotton processing enterprise and cotton spinning enterprises.

The present invention provides an airborne LIDAR air strips overlay region redundancy identification and elimination method, belonging to the remote sensing data processing technology field. The method comprises: determining the air strips to which each laser point in the airborne LIDAR point cloud belongs, and extracting an overlapping region; extracting the homonymous points in the overlapping region; calculating the conversion parameters according to the least square method, converting coordinates of the laser points of the overlapping region, and realizing the overlapping of the homonymous features; calculating a distance threshold value, and performing redundancy removal according to the laser point distance; employing the information entropy to calculate the information content of the LIDAR point cloud data in the overlapping region prior to the redundancy removal and after the redundancy removal, and performing quantitative evaluation of the redundancy removal result. The airborne LIDAR air strips overlay region redundancy identification and elimination method can effectively remove the redundant information of the overlapping region, reduce the data processing pressure, improve the efficiency, can be directly applied to the redundancy elimination of the ground and between the data of the vehicle-mounted LIDAR adjacent observation stations; adaptively obtaining the distance threshold values, determining whether the laser point has redundancy or not, making the selection rule of the data points, and realizing the rejecting of the redundancy laser point; and providing the information entropy measurement of evaluating the redundancy validity.

The invention discloses a method for detecting a moving target of a Video SAR (Synthetic Aperture Radar) by utilizing space-time correlation. The method comprises the following specific steps of: step1, carrying out shadow detection on a single frame image of the Video SAR by adopting a basic static shadow detection model; step 2, obtaining a target motion parameter by adopting a matching searchmode, and obtaining a motion target estimation speed; step 3, carrying out track tracking on the moving target through a Kalman filtering tracking algorithm, and preliminarily screening a real movingtarget by using a track association method; and step 4, tracking the target in the step 3 by adopting a space-time correlation method, comparing the distance between the centroid of the shadow area ofthe first frame of image and the shadow area of the last frame of image and a given threshold after a period of time, and screening a real moving target shadow. Static shadows are removed from a high-probability moving target set, the accuracy and registration error robustness of Video SAR moving target detection are improved, and real moving targets can be screened out under the conditions thatthe moving target speed is high and the signal-to-noise ratio caused by target defocusing is low.

The invention relates to an appearance defect detection apparatus for a small-modulus gear with a shaft. The appearance defect detection apparatus comprises a gear transferring disc mechanism, a gearfeeding mechanism, a visual detection mechanism and a gear removal mechanism, wherein the gear feeding mechanism, the visual detection mechanism and the gear removal mechanism are circumferentially distributed around the gear transferring disc mechanism, the gear feeding mechanism enables the gear with the shaft to enter the gear transferring disc mechanism in a state that the axial direction is in the vertical direction, the gear with the shaft is transferred by the gear transferring disc mechanism to successively pass through the visual detection mechanism and the gear removal mechanism, andthe gear removal mechanism sorts and blanks the gear with the shaft according to a gear shape visual detection result of the visual detection mechanism. Compared with the prior art, by adopting an automatic feeding way, a visual detection way and an automatic removal way, the appearance defect detection apparatus is high in accuracy, low in power consumption, capable of solving the problems thatthe gear detection efficiency is low, and one set of devices is suitable for detecting a product and the like, capable of realizing the rapid and automatic detection, and widely applicable to the appearance defect detection of various small-modulus gears with the shaft.

The invention discloses a method for processing ground synthetic electric field data measured practically. The method comprises: according to a missing value writing-in rule of ground synthetic electric field data measured practically, a missing value is determined and rejected; on the basis of a Pauta criterion, determination and rejection are carried out on abnormal value segmentation of ground synthetic electric field data measured practically; the rejected missing value and abnormal value in the ground synthetic electric field data measured practically are made up by using a linear regression method, thereby obtaining ground synthetic electric field data after making up; if the ground synthetic electric field data after making up do not meet expectation, a segmentation way is adjusted and the abnormal value is made up again by using the linear regression method; and if the ground synthetic electric field data after making up meets expectation, a random interference value in the ground synthetic electric field data after making up is filtered by singular value decomposition. Therefore, the impact on the ground synthetic electric field data by measurement equipment abnormality or external environment interference can be reduced; and thus a mechanism analysis of a ground synthetic electric field can be promoted.

The invention provides an online monitor method of a transmission zero value of a digital mulitbeam system, aiming to provide a RF closed loop online transmission zero value monitor method that monitors accurately and has no limitation on array size of a multibeam antenna and a quantity of beams. The method comprises the steps of coupling a transmission working signal containing m beams from n FRports of n channels on each sub-array, directly superposing, synthesizing and then delivering to an online monitor unit in the sub-array, measuring distance of m beam signals in one direction in realtime, calculating real-time phase-shifted values of m beams in n channels in the sub-array, and deducting a corresponding real-time phase-shifted value from a real-time distance measurement value of each beam in the sub-array, so as to obtain a real-time zero value of each beam signal; and fitting data according to signal quality so as to obtain an online monitor zero value of the sub-array, delivering online monitor zero values of all sub-arrays to an all-array online monitor unit for fitting, so as to obtain an online monitor value R0 of a transmission zero value of an all-array digital multibeam system.

The invention relates to the technical field of medicine production and in particular relates to medicine bottle detection equipment and a medicine bottle detection system. The medicine bottle detection equipment comprises a conveying device, a detection device, an elimination device and a control device, wherein the detection device and the elimination device are arranged in sequence along a conveying device of the conveying device; the detection device comprises an image acquisition unit and first laser sensors; the image acquisition unit is arranged above the conveying device, forms a detection station which horizontally surrounds a medicine bottle to be detected and is used for acquiring image information of the medicine bottle to be detected in all directions; the first laser sensors are arranged at two sides of the conveying device and are used for feeding back a signal for indicating the medicine bottle appears at the detection position back to the control device; the control device is in signal connection with the image acquisition unit, the first laser sensors and the elimination device respectively. According to the medicine bottle detection equipment and the medicine bottle detection system, a whole medicine bottle detection flow can be used for automatically and intelligently finishing detection work of the medicine bottle; a detection result is accurate and has high reliability and the detection efficiency is high; the labor cost of enterprises is saved; the medicine bottle detection equipment and the medicine bottle detection system have wide application and popularization prospect.

The invention discloses an adhesive tape foreign matter synchronous removal device which comprises a material transfer system, a foreign matter positioning mechanism, a synchronizing mechanism and a removal mechanism, wherein the material transfer system can transfer an adhesive tape; the foreign matter positioning mechanism can move the removal mechanism to the foreign matter position; the synchronizing mechanism can realize synchronous movement of the removal mechanism and the adhesive tape; the removal mechanism can remove the adhesive tape containing the foreign matter. In cooperation with x-ray detection equipment, the device disclosed by the invention can realize synchronous removal of foreign matters in an adhesive tape conveying process, and can be used for solving the problem of stopping material transfer for removing foreign matters, and improving the production efficiency.

The invention relates to an integrated simulation platform and method for an electromagnetic rail gun. The integrated simulation platform and method for the electromagnetic rail gun aim at solving the problems that in the prior art, whole trajectory simulation, model verification, simulation optimization, simulation test design and firing accuracy analysis and assessment cannot be performed on the electromagnetic rail gun. According to the technical scheme, the integrated simulation platform for the electromagnetic rail gun comprises an electromagnetic rail gun simulation subsystem, a model verification subsystem, a simulation optimization subsystem, a test design subsystem and a firing accuracy analysis and assessment subsystem; the electromagnetic rail gun simulation subsystem comprises a bullet module, a rail module, a pulse power supply module, an armature module, an object module, a parameter configuration module, a simulation process management module and a data recording module. The integrated simulation platform and method for the electromagnetic rail gun are applied to the technical field of computer simulation.

The invention relates to a method for removing an abnormal spectrum of a full-spectrum direct-reading spectrometer for analyzing a single-matrix material, comprising: performing spectrometer calibration operation; setting the monitoring pixels of respective detectors, forming a monitoring light intensity by using the light intensity values of respective monitoring pixels; setting a position correction reference spectrum; and removing the abnormal spectrum. The method can eliminate the data corresponding to the abnormal spectrum and improves the repeatability and accuracy of the analysis data.

The invention relates to a battery detection machine. The battery detection machine comprises a conveying device. A code scanning device, a detection device and a sorting device are arranged in the conveying direction. The code scanning device comprises a code scanner and an auxiliary code scanning device body. The detection device comprises detection probes which are installed on a pressing device. An eliminating device comprises an eliminating bracket. A transporting device is installed on the eliminating bracket. Material taking heads are installed on the transporting device. According to the battery detection machine, code scanning, detecting and eliminating of battery cores are synchronously achieved, and meanwhile the efficiency is high, so that the requirements for production and processing of an automated assembly line are met.

The invention relates to a cylindrical surface image matching method combining with SURF feature extraction and curve fitting, comprising steps of arranging two images (A and B) on an upper position and a lower position in a left-right alignment manner, adopting a SURF characteristic detection algorithm to perform characteristic detection on the two images, finding out a matched pair set, calculating the angle between the straight line defined by each matched pair in the characteristic point matched pair set and the horizontal direction and the angle and an Euclidean distance between the two characteristic points in each matched pair, establishing an angle set K of the image matched pair and an Euclidean distance set D of the image, performing curve fitting on the angle set K of the matched pair, wherein the independent variable of the curve fitting is the abscissa X1i of the matched pair, the ordinate of the curve fitting is angle Theta I, removing the mismatching, performing curve fitting on the Euclidean distance set D of the matched pair, wherein the independent variable of the curve fitting is the abscissa X1i of the characteristic point in the image A, and removing the mismatching. The invention can more accurately remove the mismatching of the cylindrical object.

The invention relates to a plastic film printing character defect detection device and method. The device comprises a static electricity generating module, a conveying module, an illumination module and a collection module, wherein the static electricity generating module is arranged on the conveying module; a detection region is arranged on the conveying module; an illumination module and a collection module are arranged above the detection region. The plastic film product can be flatly conveyed according to the preset speed; the illumination brightness and angle are regulated; a CCD area-array camera is used for regularly and completely acquiring image data; a progressive compensation algorithm and a bidirectional difference image method are used for accurately and effectively determining the existence and the type of the printing character defects; meanwhile, the defect information is subjected to automatic warning and deletion operation. The whole system does not need the participation of personnel; the intelligence of the whole system is realized; the character defect automatic detection, classification and deletion can be realized; the problems of high manual participation cost, low speed, low recognition rate and the like can be avoided; the detection on the printing character defects can be accurately and efficiently realized.

A high-energy particle interference suppression method of a star sensor is divided into an all-celestial acquiring mode and a window tracking mode according to working modes of the star sensor. The high-energy particle interference suppression method in the all-celestial acquiring mode and the high-energy particle interference suppression method in the window tracking mode comprise the steps of primary star point selection and primarily selected star point verification. The method effectively suppresses charged particle interference, and the tracking performance and the identification success rate of the star sensor are improved and increased in various space radiation environments.

The invention discloses a method and device for standing-wave ratio detection. The purpose is that the standing-wave ratio can be rapidly and accurately detected only by the utilization of a downlink service signal emitted by a TD-LTE base station system, and it is avoided that the base station system is extra disturbed by a specific training sequence. The method includes the steps that output power detection (OPD) data of the service signal emitted by the base station system and reaction power detection (RPD) data of a device to be detected inside a base station are captured; within a first preset bandwidth range, according to a preset data length, OPD feedback signals and RPD feedback signals in multiple time periods are respectively extracted; the spectral characteristics of the corresponding OPD feedback signal and the RPD feedback signal in each time period are determined, and the reflection coefficient of the base station system is determined according to the spectral characteristics of the corresponding OPD feedback signal and the RPD feedback signal in each time period; according to the reflection coefficient of the base station system, the standing-wave ratio of the base station system within the first preset bandwidth range is determined.

The invention discloses a strip missing detection device for a loading sealing case machine. The strip missing detection device comprises an intelligent camera, mounting brackets, a trigger pulse photoelectric switch, a positioning sensor, an illumination light source, a touch display screen, a photoelectric coupler and a controller. The intelligent camera is mounted on the upper portion of a cigarette pushing plate of a first-push pushing rod through the first mounting bracket; the strip-shaped light source is mounted on cigarette strip protective baffle organic glass through the second mounting bracket; the trigger pulse photoelectric switch is mounted on a first-push pushing rod track through the third mounting bracket; the positioning sensor is mounted on a second-push pushing rod track through the fourth mounting bracket; and the touch display screen and the controller are mounted above an electric cabinet. The detection device can discover missing of cigarette strips in a case body of a loading sealing case in time, the production process does not needed to be interrupted, automatic detection and removing of cigarette strip missing products in the case can be achieved, and perfect matching of the detecting speed and the equipment running speed is achieved.

The invention discloses an ear band detection device for a mask and relates to the technical field of mask manufacturing equipment. The ear band detection device comprises a detection table, wherein an air cylinder is mounted on the detection table, a lower pressure block is mounted at a telescopic rod end of the air cylinder, rectangular notches are symmetrically formed in the detection table relative to the lower pressure block, the lower surface of the detection table is connected to two detection rods slidably through chutes, a wedged plate and a force cell are connected to the side surface, facing the rectangular notches, of each detection rod, a fixed column is fixed to the lower surface of the detection rod, the lower surface of the detection table is slidably connected to a cam trough plate, two cam troughs are formed in the upper surface of the cam trough plate and the two cam troughs are arranged in a splashed manner, and the cam trough and the fixed column are in clearance fit, and the cam trough moves, so that the two detection rods move close to each other or far away from each other. By arranging the detection rods, the detection rods advance to drive an ear band to stretch. Whether the ear band is welded firmly or not is detected by the force cell, so that the problems that the detection efficiency of an existing ear band of the mask is poor and the detection efficiency is low are solved.

The invention discloses a capsule waste removing and polishing all-in-one machine which comprises a machine frame, an inner roller, a polishing cylinder, a dust collector and a motor. The machine frame is fixedly provided with the polishing cylinder, the inner roller is arranged in the polishing cylinder in a sleeved mode, and a gap exists between the polishing cylinder and the inner roller; the machine frame is provided with the motor serving as a power source, the motor drives a roll shaft to rotate through a belt wheel, and the roll shaft is fixedly connected with one side of the inner roller; the other side of the inner roller is provided with a feed inlet, and a cover plate is disposed on the inner roller; the dust collector is arranged at the bottom of the polishing cylinder and connected with the inner roller; a discharge port is formed in the bottom of the inner roller, and the bottom of the polishing cylinder is provided with a negative-pressure waste removing channel. By means of the machine, polishing is achieved while hollow capsules and other defective products are removed, the structure is simple, brushes are not used, the problems that the brushes abrade capsules and bristles fall are avoided, continuous production can be achieved, production efficiency is improved, the defective products and qualified products are collected separately, and production cost is lowered.

A method for calibrating data of chemical industry process based on bilinear model includes calculating and obtaining measurement data of total flow rate and composition content of certain chemical industrial process unit according to relevant chemical balance equation, setting up calibration model of material flow in chemical industrial process unit according to said basic relation and measurement data table, decomposing bilinear problem in calibration operation to be two linear constraints for making derivation be more easier.

The invention provides a transgenic carrier with function of automatic control and elimination of selection marker and application thereof in zea mays marker-free transgenic breeding, and belongs to the technical field of breeding of zea mays molecules. The carrier uses a pEGAD carrier as a skeleton carrier, and a fusion gene of EGFP (enhanced green fluorescent protein) and IPT (isopentenyl transferase) genes is used for replacing a bar gene and is used as the selection marker; an expression framework of the selection marker is integrated into a Cre / lox site specific recombination system. Theapplication comprises the following steps of transferring the carrier into a zea mays embryonic callus, obtaining the zea mays embryonic callus with the green fluorescent expression effect; after heatshock, culturing the zea mays embryonic callus without the green fluorescent expression effect by differentiating and rooting, so as to obtain a transgenic plant. The carrier has the advantages thatthe selection efficiency is improved, the removal of the transgene selection marker is realized, and the theoretical and application basis is provided for the large-range development of marker-free zea mays transgene.

The invention discloses a non-standard flange double-sided polishing device capable of detecting a flange center column. The non-standard flange double-face polishing device specifically comprises a bottom plate, a machining base, a paint spraying mechanism, a polishing mechanism and a bearing table, a rotating base is installed in the middle of the top of the bottom plate, a rotating motor is fixed to the bottom end of the interior of the rotating base, a rotating column is mounted on the top end of the interior of the rotating base, and an output end of the rotating motor is connected to the rotating column through a rotating shaft. According to the non-standard flange double-sided polishing device capable of detecting the flange center column, two clamping bases are arranged, the two clamping bases are matched to clamp and limit a to-be-machined flange, the fact that the grinding or paint spraying precision is affected due to the displacement of the to-be-machined flange in the grinding or paint spraying process is avoided, the clamping bases can drive the flange to turn over, accordingly, the two faces of the flange are polished or painted, manual turn-over is not needed, and machining is more convenient.

The invention discloses a detecting and removing machine for a bubble cap capsule plate containing capsules. The detecting and removing machine comprises a machine frame and at least one set of capsule detecting and removing assembly. Each set of the capsule detecting and removing assembly comprises a conveying device, a detecting device, a removing device and a control device. Each conveying device sequentially comprises an upper conveying device body and a lower conveying device body according to the direction of the capsules. Each detecting device comprises an industrial camera, a light source, a first counting sensor and a second counting sensor. Each industrial camera is installed on a camera support, and each light source is installed on a light source support. The camera supports and the light source supports are both installed on the machine frame. The industrial cameras and the light sources are both located above upper conveying belts. The first counting sensors, the second counting sensors and the industrial cameras are all connected with the corresponding control devices. Each removing device comprises a removing base and an overturning power device. By means of the detecting and removing machine, the detection rate of unqualified products is guaranteed.