174 results about "ENTEROSCOPE" patented technology

The present invention relates to a cover having a plurality of moveable, external, angled projecting elements for use with flexible medical scoping devices such as endoscopesor enteroscopes. The invention includes the cover with an over cuff and use of the disposable removable covering in methods of medical scoping procedures or examinations. The invention also includes an applicator for assisting in placing the covering about or over a medical device and a kit of parts.

The invention belongs to the technical field of fecal microbiota transplantation, and particularly relates to a fecal microbiota capsule, as well as a preparation method and application of the fecal microbiota capsule. The preparation method comprises the following steps: (1) a fecal microbiota solution is prepared; (2) the fecal microbiota solution, obtained in the step (1), is added with a freeze-drying protective additive, and freeze drying is performed to obtain fecal microbiota freeze-dried powder; (3) a capsule is filled with the fecal microbiota freeze-dried powder, obtained in the step (2), to obtain the fecal microbiota capsule. Through adoption of the fecal microbiota capsule, flora is enabled to be comparatively concentrated and can be preserved for a long time at a low temperature; implementation of flora transplantation is facilitated, and invasive implementation manners, such as a gastroscope, an enteroscope and a nasointestinal feeding tube, are not needed; when the fecal microbiota capsule is applied to weak patients, the safety is increased.

The invention provides a real-time monitoring method and system for the enteroscope retreating speed based on computer vision. the enteroscope retreating speed and stability are actively prompted to adoctor, the doctor is reminded when the speed is overspeed or a picture is unstable, the doctor is supervised to control the enteroscope retreating speed within a safe range all the time, and the enteroscope examination quality is improved. The perception Hash algorithm technology is adopted, the enteroscope withdrawing speed is reflected in real time by analyzing the stability of the enteroscopeimage instantly or for a period of time, a doctor is reminded to control the enteroscope withdrawing speed within the safety range all the time during enteroscope examination, and the detection comprehensiveness and effectiveness are improved.

The invention discloses an auxiliary diagnosis system and method for ulcerative colitis under enteroscopy based on deep learning. The system comprises an enteroscopy image automatic acquisition subsystem, a client, a server and a database. The enteroscope image automatic acquisition subsystem is used for acquiring an enteroscope image; The client is used for uploading the enteroscopy image acquired by the enteroscopy image automatic acquisition subsystem to the server, and judging whether the image is qualified or not and whether the image comprises ulcerative colitis judgment or not, and an analysis result fed back by the server is received and displayed; the database is used for storing the sample set for training the convolutional neural network, the acquired enteroscopy image and the analyzed and output information. According to the method, an image recognition technology is utilized to monitor an endoscope video in real time, images containing key organ parts and suspicious focusareas are automatically collected, a neural network model is utilized to automatically screen the images, the most valuable image can be extracted from a global video, and more reliable and efficientsupport is provided for doctors to diagnose.

Provided is an application of natural faeces flora transplanting in intestinal disease treatment. Faeces of healthy people screened through multiple conditions is collected, intestinal canal mixed probiotic flora is collected through the steps of homogenating, suspending, diluting, filtering, centrifuging and the like and transplanted in the ways of a gastroscope, an enteroscope, a nose-jejunum tube, a fistulization opening, clysis and the like, flora balance in the intestinal canal is restored, and the effect of treating intestinal diseases is achieved. A method for treating bacteria with bacteria is adopted, the stability of the intestinal flora of a human body is maintained, a microecological treating method free of drug resistance and toxic and side effects is achieved, treatment is carried out in the mode of faeces bacterium capsules or bacterium liquid, and a good treatment effect is achieved on diarrhea, clostridium difficile infection, Crohn's diseases and the like.

The invention provides an intelligent enteroscope, belonging to intelligent electronic medical equipment. The enteroscope is indispensable equipment for routine examination of gastrointestinal diseases. At present, enteroscopy is operated manually and the doctors ensure the enteroscope to move in the intestinal tract with a manual device by virtue of handfeel, experience and feelings of the patients, so the enteroscope blindly enters into the intestinal tract, uncertainties exist and the clumsy enteroscope and unscientific operation bring many harms to the patients. Under the premise of respecting lives and complying with the objective laws of the human body, the intelligent enteroscope is manufactured by adopting the electronic information technology, specific material dynamic and humanized design and adapts to bend of the large intestine by bending by virtue of own bendability and freely moves in the bent large intestine. The intelligent enteroscope ensures the patients to be examined easily and quickly without any pain or risk, simultaneously frees the hospitals and doctors from the unbearable examination pressure, improves the examination quality, guarantees the safety of medical treatment, effectively elevates the electronic informatization level of treatment of the gastrointestinal diseases and brings benefits to the patients and the society.

The invention discloses an enteroscopy quality evaluation method and device based on image recognition and a storage medium. The evaluation method comprises the following steps: splitting a video acquired in an enteroscopy process into a plurality of examination images; dividing the plurality of inspection images into close-shot mode images and non-close-shot mode images based on the first classification condition and the second classification condition; and accumulating the collection duration of the close-shot mode image, and evaluating the enteroscopy quality based on the collection duration of the close-shot mode image. According to the method, the collection time of the close-shot mode image is used as an evaluation index, a scientific and reasonable enteroscopy quality evaluation standard is established, and an endoscopy doctor can be guided to improve the collection time of the close-shot mode image by changing an examination operation mode within a shorter enteroscopy time; furthermore, a longer close-range view field capable of effectively observing intestinal mucosa is obtained, so that the examination quality of the enteroscope is improved.

The invention relates to the technical field of image recognition and intelligent medical treatment, in particular to an enteroscopy quality standard and scoring system based on image recognition, which is characterized by comprising a middle enteroscopy quality standard and a scoring system, wherein the enteroscopy quality standard judgment table comprises the number of suspicious intestinal segments found in the whole enteroscopy withdrawal process, the total check time, the enteroscopy withdrawal time, the effective enteroscopy withdrawal time, the total number of screenshots, whether a blindness return part is reached or not, whether rectum inversion operation is carried out or not, the check integrity index, the check definition and the safety index. The enteroscopy quality standard and scoring system can provide an enteroscopy quality standard to quantitatively evaluate each examination of the whole life, and meanwhile, can perform individual scoring on each examination of each doctor according to the provided enteroscopy quality standard. In this way, a doctor can more carefully examine the enteroscope, and the doctor can also discover the defects of the doctor and continuously improve the operation technique, meanwhile, a patient can be more relieved, and huge economic and social benefits are achieved.

Provided is application of transplantation of artificially-cultured fecal floras in treatment of intestinal diseases. According to the application, feces of healthy persons screened out under various conditions is collected, and intestinal mixed probiotic floras are collected through the steps of homogenization, suspension, dilution, filtration, centrifugation and the like; culture is performed under the anaerobic condition and the aerobic condition by the adoption of a plate isolation technology, and culturable intestinal probiotic floras are picked out; transplantation of the artificially-cultured intestinal floras is performed by the means of a gastroscope, an enteroscope, a nose-jejunum tube, fistulation, clysis and the like, intestinal flora balance is recovered, and then the effect of treating intestinal diseases is achieved. Particularly, the application has a good treatment effect on diarrhea, clostridium diffieile infection, Crohn's diseases and the like.

The invention discloses an enteroscopy Crohn's disease auxiliary diagnosis system and method based on deep learning. The system comprises an enteroscopy image automatic acquisition subsystem, a database, a client and a server. The enteroscope image automatic acquisition subsystem is used for acquiring an enteroscope image; wherein the database comprises a sample set for training the convolutionalneural network; and the client is used for uploading the enteroscopy image acquired by the enteroscopy image automatic acquisition subsystem to the server, and receiving and displaying an analysis result fed back by the server. According to the method, an image recognition technology is utilized to monitor an endoscope video in real time, images containing key organ parts and suspicious Crohn focus areas are automatically collected, and after global preferences are made according to a weighting algorithm, whether the image is Crohn disease or not is diagnosed. According to the method, after the neural network model is utilized to automatically screen the images, the most valuable image can be extracted from the global video, auxiliary diagnosis is given, and more reliable and efficient support is provided for doctors to diagnose.

The invention discloses a pCR prediction method based on multi-type medical data. The method includes the following steps that: clinical data, CT diagnosis reports and enteroscope images are obtainedfrom a medical department; an SVM is trained by using the clinical data, a BERT model is obtained by means of transfer learning training with the CT diagnosis reports, and a Faster-RCNN model is obtained by means of transfer learning training on the basis of the enteroscope images; and the clinical data, CT diagnosis report and enteroscope image of a patient to be predicted are inputted into the three trained models, predicted pCR probabilities p1, p2 and p3 are obtained, the predicted pCR probabilities p1, p2 and p3 are fused, a final predicted pCR probability p is obtained, and if p is greater than a set threshold T, it is predicted that the patient has pCR. According to the method, the Faster-RCNN network is used; a tumor ROI can be automatically generated; a whole process does not needmanual intervention; and prediction efficiency is improved. The neural network is used for representation learning; manual setting and feature selection are not needed; and therefore, the accuracy and efficiency of prediction are improved. PCR prediction is carried out by combining the clinical data and the CT diagnosis report of the patient, so that prediction accuracy is improved.

The invention relates to an intelligent colonoscopy and belongs to intelligent electronic medical instruments. A gastrointestinal endoscope is a very indispensable and important instrument for routine examination of digestive tract diseases. However, an existing colonoscopy has no digestive tract intervening operation function and is manually operated to cause harm to a patient. Due to the facts that the physical performance of the contact face between the colonoscopy and the inner wall of the gastrointestinal tract is improved and the friction coefficients between the colonoscopy and the inner wall of the gastrointestinal tract are increased, the colonoscopy can acquire safe, stable and reliable running power in the gastrointestinal tract. The self-adaptive curve of the colonoscopy corresponds to the curve of the gastrointestinal tract, so that scientific and accurate colonoscopy examination can be smoothly and conveniently finished in the situation without any pain risk for the patient. Thus, the medical safety is guaranteed, and the medical quality is improved.

The invention discloses a method for intelligently identifying and recording alimentary canal tissues and foreign matters in endoscopic examination. Images detected by the gastrointestinal endoscopy are stored in a classified manner; a deep learning neural network model is established, image data stored in a classified mode are randomly divided into a training set and a test set for deep learningtraining, real-time enteroscope image information is obtained and subjected to vector gridding processing, then the real-time enteroscope image information is input into the deep learning neural network model for recognition, and the beneficial effects of being rapid, accurate and autonomous in learning are achieved.

The invention discloses a real-time colonoscope polyp detection device based on a deep convolutional neural network, and the real-time colonoscope polyp detection process comprises the steps: obtaining a to-be-detected current frame of colonoscope image, and inputting the to-be-detected colonoscope image into a colonoscope polyp detection model; after the image feature map of the enteroscope image to be detected is extracted by using the feature extraction unit, predicting and outputting the spatial attention feature map corresponding to the current frame of enteroscope image by using the attention prediction unit according to the image feature map and the spatial attention feature map corresponding to the previous frame of enteroscope image; and predicting and outputting a detection result by using a detection unit according to the image feature map and the spatial attention feature map corresponding to the previous enteroscope image. The real-time enteroscopy polyp detection device can improve the accuracy and efficiency of clinical enteroscopy.

The invention relates to the technical field of medical assistance, in particular to a real-time monitoring method for enteroscope withdrawal time based on a random forest algorithm, which comprises the following steps of S1, collecting endoscopic image data of intestinal non-cecum, ileocecum and appendix opening, S2, dividing and marking the image data of the non-cecum, the ileocecum and the appendix opening according to the endoscope entering sequence of enteroscopy to obtain an enteroscopy data set, S3, based on a random forest algorithm, constructing and training an enteroscope ileocecum recognition model, and S4, judging whether the current digestive endoscopy video image is an ileocecum or an appendix opening or not, judging that the current digestive endoscopy video image is the ileocecum or the appendix opening, starting endoscope reversing timing until the enteroscope is moved out of the body. According to the method, a random forest algorithm is adopted to train an enteroscope ileocecum model, and the ileocecum or appendix opening at the tail end of the intestine is recognized through the model. Automatic recording and displaying of enteroscope return operation time are achieved, the purpose of reminding an endoscopist is achieved, and enteroscope examination quality is guaranteed.

The invention discloses a proctoscope provided with a light source and a camera lens and belongs to the technical field of instruments for anorectal surgery. The proctoscope comprises a proctoscope body and a guide rod, wherein a proctoscope handle is arranged above the rear end of the cone-shaped proctoscope body, a through opening is longitudinally formed in the lower part of the proctoscope body, a front port is formed in the front end of the proctoscope body, a cable is mounted in the proctoscope body, one end of the cable is led out from a through hole in the proctoscope handle to be connected with an imaging device, the light source and a controller, a combined camera is mounted at the other end, the camera lens is mounted in the middle of the combined camera, light emitting lamps are arranged on the periphery of the camera lens, a spherical adjusting cavity is formed in the rear end of the camera lens, and a control rod is connected in the connecting cavity and connected with the controller through a wire; a guide head is arranged at the front end of the guide rod, the rear end of the guide head is connected with a connecting rod, a rear baffle is arranged at the rear end of the connecting rod, a crescent groove is formed in the lower end of the rear baffle, and the rear end of the rear baffle is connected with a guide rod handle. The proctoscope enters the rectum from the anus, an obtained image is transmitted to a display through the camera lens, the visual operative field is amplified, a focus is clearly displayed, and excision of a tumor, hemostasis and suturing can be completed.

The invention discloses a rigid microwave enteroscope system, belonging to the field of medical equipment. The rigid microwave enteroscope system comprises a rigid enteroscope and a cold light source main machine, a main camera and a monitor which are connected with the rigid enteroscope, and the rigid enteroscope is also provided with a microwave knife operating device used for cooperating with treatment and a micro knife system main machine used for cooperating with the microwave knife operating device. As a novel medical treatment means for treating various diseases by using microwave energy for local heating, the microwave technology has wide application range, obvious curative effect, safety and no side effect. In the invention, by introducing the advanced microwave technology into the enteroscope operation and taking the advantages of the microwave technology, the diagnosis of the enteroscope and the therapeutic action of microwave are superposed, and the cure rate and the accuracy of the operation can be increased to achieve the purpose of reducing recurrence rate. The invention can improve the medical quality and the medical safety and better overcome the defects of the traditional operation for treating anorectal diseases, thereby obtaining unexpected results.

A guide and method for tilting a colonoscope in the rectum leftward toward the descending colon to advance the colonoscope through the colon during colonoscopy. The guide comprises a rigid elongated body member and a handle attached to or integral with the body member. The body member has an insertion length selected so as to not be inserted past the patient's rectum. The body member can be curved or tubular to better engage the colonoscope. A soft liner can be utilized to protect the endoscope during use. Preferably, the body member has a semi-circular cross-section with a flexible inner tubular member attached thereto that receives the endoscope. A longitudinal slot facilitates loading the endoscope into the tubular member. In another embodiment, the body member comprises upper and lower rigid body members connected by an engaging mechanism. The user tilts the colonoscope with the guide to minimize or avoid sigmoid looping.

The invention discloses an enteroscope and belongs to an electronic medical apparatus. The enteroscope is an indispensable apparatus for performing routine examination on a patient who suffers from a gastrointestinal disease. Current enteroscopy is operated manually, and a doctor turns the enteroscope around in an intestinal canal through the handfeel and the experience of a doctor and the feeling of the patient with a manual operating device, so the enteroscope enters the intestinal canal blindly, a plurality of uncertainties exist, the clumsy enteroscope and the unscientific operation bring various damages to the patient and big physical and psychological pressure on the doctor is caused; on the premise of respecting life and complying with a human body objective law, a bionics principle, humanity design and advanced electronic information technology are adopted, and the enteroscope which can freely crawl into and out of a gastrointestinal tract, records a video of a correct lesion position, and takes a test object is manufactured; and the enteroscope makes the patient do not suffer from pain and receive examination easily and quickly, and relieves the examination pressure burdened by a hospital and the doctor, improves examination quality, guarantees medical safety, effectively increases the electronic informatization level of diagnosis and treatment of the gastrointestinal disease, and benefits the patient and the society.

The invention relates to the technical field of medical equipment, in particular to an enteroscope balloon. The enteroscope balloon comprises a longitudinally arranged fixed tube, a movable sleeve sleeved with the fixed tube, a first rubber piston, a dilation balloon, a sealing rubber ring and a push-pull device; the fixed tube and the movable sleeve are each of a hollow structure, the extended end of the movable sleeve stretches into the dilation balloon, the sealing rubber ring is used for making the movable sleeve and the dilation balloon connected, the upper end of the fixed tube is connected with the push-pull device, retraction and stretching of the movable sleeve are controlled by means of vertical movement of the second rubber piston in the push-pull device, the movable sleeve retracts, the sealing rubber ring and the dilation balloon achieve automatic separation from the movable sleeve, the dilation balloon is left in the colon, a dilation balloon physically blocks port side intestine peristaltic wave to be transmitted to the anorectal colon, large intestine peristalsis is weakened, and by means of the physical supporting effect of the dilation balloon, an operation spaceand clear vision are provided for endoscopic treatment.

The present invention relates to the technical field of the medical instrument, and in particular relates to a disposable enteroscope outer tube and necessary accessories thereof. The enteroscope outer tube comprises a plastic thin film tube (2), a transparent film (1), and two ducts (3, 4). The enteroscope outer tube is characterized in that the plastic thin film tube (2) is arranged in the enteroscope outer tube, the closed transparent film (1) is arranged on the front end of the plastic thin film tube (2), two ducts (3,4) are arranged on the outer side of the plastic thin film tube (2), one duct (4) is connected with a water vapor switch, and the other duct (3) is connected with a sucker. The disposable enteroscope outer tube provided by the invention has the characteristics of being novel in structure, economical, time-saving, convenient for disassembly and assembly, and convenient for use. According to the invention, the time-consuming problem of the expensive enteroscope due to cleaning and disinfection after examination is mainly solved, and the disposable enteroscope outer tube which can be conveniently disassembled and assembled and conveniently used and necessary accessories thereof are provided,.

The present invention relates to a cover having a plurality of moveable, external, angled projecting elements for use with flexible medical scoping devices such as endoscopesor enteroscopes. The invention includes the cover with an over cuff and use of the disposable removable covering in methods of medical scoping procedures or examinations. The invention also includes an applicator for assisting in placing the covering about or over a medical device and a kit of parts.

The invention relates to the technical field of medical technology assistance, in particular to an intestinal tract endoscope retreating speed smoothing method, which comprises the following steps: performing frame collection on an intestinal tract real-time endoscope retreating video to obtain an enteroscope picture; calculating a Hash fingerprint of the enteroscopy picture; calculating the Hamming distance between a current frame and the previous n frames of enteroscopy pictures according to the Hash fingerprint of the current frame and the previous n frames of enteroscopy pictures; performing weighted averaging on a Hamming distance vector; mapping the weighted average Hamming distance into a speed value as the speed value of the current enteroscope picture; taking the speed values of the latest X enteroscopy pictures as a sequence, and smoothing the sequence by using Savitzky-Golay filtering to obtain a smoothed speed value sequence; and taking out the last value of the smoothed speed sequence, and carrying out truncation processing to obtain the endoscope retreating speed value of the smoothed current image. According to the method, oscillation is eliminated through the smoothing method, the robustness of the endoscope retreating speed and the operation experience of doctors are improved, and the enteroscopy quality is guaranteed.

The invention discloses a suspicious early-stage cancer lesion examination device suitable for an endoscopy. A medical endoscope (a gastroscope, an enteroscope, a bronchoscope, a cystoscope, a colposcope, a laparoscope and the like) used in medical clinical examinations becomes an important tool for disease diagnosis. Many cancers can be found through observing the morphological characteristic of a pathological changed region, however, the pathological changes of early-stage cancers without remarkable morphological characteristics are difficultly detected, a great number of medical researches show that the early-stage pathological changes of the cancers are not remarkably changed in shape, but fluorescent substances in the cancer tissue are changed along with the formation of the canceration, then, the fluorescence of a cancer tissue is caused to be greatly different from that of a normal tissue, and thus, the suspicious lesions of the early-stage cancers are identified easily through fluorescence detection and expression of a difference by using a false color image way, the pathological biopsy can also be precisely indicated, and the suspicious early-stage cancer lesion examination device plays an important role in increasing the detection rate and the accuracy rate of early-stage cancer diagnosis.

The invention discloses a detection method for judging whether colonoscopy is complete or not based on image recognition. The method comprises the steps that in the whole enteroscopy process, a completeness detection model is used for conducting average gray value detection on four corners of each frame of image in enteroscopy video stream, and if the Boolean value output by a preset number or more of continuous frames of one corner is 0, a prompt that completeness detection is not conducted is output. According to the detection method for judging whether colonoscopy is complete or not based on image recognition, the problem that a doctor does not completely inspect the interior of the intestinal tract in the colonoscopy process can be solved, so the colonoscopy quality of an operator is improved, and the missed lesion diagnosis rate of colonoscopy is reduced.

The invention discloses a method for increasing the adenomatous polyp detection rate of a colonoscope based on deep learning. The method comprises the following steps: dividing a video stream transmitted by an enteroscope lens in an operating table into two parts, transmitting one part of the video stream to an operation platform of a doctor, preprocessing the other part of the video stream, and sending the preprocessed video stream to a polyp detection model embedded into an enteroscope operating system for identification; allowing the polyp detection model to detect whether polyps appear ineach frame of image and the appearing probability of polyps; and returning a detection result of the polyp detection model to the operation platform of the doctor for displaying, and if polyps appearin the video stream, framing the polyps for prompting. By means of an artificial intelligence deep neural network, polyps appearing in a lens in the operation process of colonoscopy can be automatically detected, a polyp recognition rate in colonoscopy examination process is increased, and therefore, the adenomatous polyp detection rate is indirectly increased.

The invention discloses a narrow-band imaging detection method in a colonoscope operation based on image recognition. The method comprises the following steps: detecting a colonoscope video through abuilt colonoscope polyp detection and recognition model, and judging whether polyps exist or not; and if the polyps are detected, judging the relationship between an R channel average value of intestinal images and a threshold value by utilizing an OpenCV image processing technology, and giving a prompt of whether NBI observation needs to be carried out or not according to the relationship betweenthe R channel average value of the intestinal images and the threshold value. The method is used for detecting colonic polyps in real time; when polyps are found, whether an operator observes the polyps through the NBI technology or not is judged through the OpenCV image processing technology; and if not, a prompt is given to standardize the operation method of the operator, so the missed diagnosis rate of polyps is reduced to the maximum extent, and therefore, the death rate of patients is reduced.

The invention discloses an endoscope tube cleaning and disinfecting device for gastroenterology in a hospital, and belongs to the technical field of medical equipment. The endoscope tube cleaning anddisinfecting device comprises a casing, operation ports, a liquid outlet and vent holes. The operation ports are formed in two sides of the casing and convenient for a user to fix an endoscope tube onan inner assembly of the casing. The liquid outlet is disposed at the bottom of the casing, and the plurality of vent holes are formed in the top of the casing. The device further comprises a cleaning assembly, a disinfecting assembly, a drying assembly and a transmission assembly. The cleaning assembly is disposed on the left side of the casing, the disinfecting assembly is disposed on the rightside of the casing, the drying assembly is disposed at the top and inside the casing, the transmission assembly is disposed inside the casing, and the transmission assembly is disposed under the cleaning assembly, the disinfecting assembly and the drying assembly. The device can be used for disinfecting and cleaning endoscopic tubes such as enteroscope tubes and gastroscope tubes for the gastroenterology, and has broad application prospects.

The invention discloses a detection method for a stationary state of an enteroscope lens based on image recognition matching. The method comprises the following steps: reading each frame of a target image in an enteroscope video, converting each frame into a grey-scale map, then normalizing the grey-scale map to a size of 512*512, and cutting the normalized the grey-scale map into small blocks, each with a size of 4*4; calculating a gray average value of each small block to obtain an average value set of 16 numerical values; calculating the difference between the gray average value sets of a previous frame and the current frame, and further solving a change proportion; and if the change proportion of continuous N frames is lower than a threshold value, determining that the lens is static.According to the invention, whether the effective retreating examination time of a doctor during enteroscopy surgery of a patient reaches a preset time or not can be effectively judged, and a prompt can be given.

The invention discloses an enteroscope image polyp error detection inhibition device based on deep learning. The device comprises the following steps: obtaining a to-be-detected enteroscope image, and inputting the to-be-detected enteroscope image into an enteroscope image polyp error detection inhibition model; after the feature extraction module is used for extracting a feature map of the enteroscope image to be detected, adopting the detection head module for detecting the feature map, calculating and outputting a detection result, and the invalid frame discrimination module is used for carrying out invalid frame discrimination on the feature map and outputting a discrimination result; performing suppression output by the suppression processing module according to the detection result and the judgment result, suppressing the output of all detection boxes in the detection result when the enteroscope image to be detected is classified as an invalid frame according to the judgment result, and otherwise, outputting the detection boxes and confidence after the detection result is normally screened according to a confidence threshold. Under the condition that the recall rate and the running speed of the detection method are hardly influenced, the number of false detection frames is reduced, so that the detection efficiency is improved, and the practicability of the detection method is improved.