47results about How to "Reduce the impact of extraction" patented technology

The invention discloses a method for quickly and accurately detecting and tracking a human face based on a video sequence, which relates to the technical field of mode identification. The method comprises the following steps of: 1, extracting a video frame image from a video stream; 2, preprocessing the video frame image, namely compensating light rays, extracting skin color areas, performing morphological processing and combining the areas; 3, detecting the human face, namely representing the human face by using Harr-like characteristics and detecting the human face by using a cascaded Adaboost algorithm with an assistant decision function; 4, establishing the characteristics of the human face, namely detecting the area characteristics of the detected human face and the shape characteristics of the edge profile of the human face; 5, tracking the human face, particularly tracking the human face by using a human face area characteristic model when an intersection does not occur in a human face area, and further matching when the intersection occurs according to the shape characteristics of the edge profile of the human face; and 6, extracting the sequence of a human face image. By the technical method, the human face can be detected and tracked quickly and accurately on the basis of the video sequence.

The invention provides a business card identification method and device, and the method comprises the steps: obtaining a to-be-identified business card image; inputting the business card image into acharacter detection model to obtain each text line area, wherein the character detection model learns and obtain a corresponding relation between image features and each text line area; and inputtingeach text line area into the character recognition model to obtain business card information corresponding to each text line area. According to the method, the text line areas in the business card image can be identified through the text detection model based on deep learning, the robustness is high, the influence of low-quality and noise data on text extraction can be reduced, and therefore the universality and the application space of the method are improved. Moreover, the text line areas are subjected to end-to-end recognition based on the deep learning character recognition model, single word segmentation is not needed, higher accuracy is achieved, higher recognition capability is achieved for various complex changes, and the universality and the recognition effect of the method are improved.

The invention discloses a method for extracting and continuously enriching polysaccharides from culture media for cordyceps militaris. The method includes steps of drying the rice residual culture media for the cordyceps militaris; carrying out superfine grinding; carrying out enzymatic hydrolysis by the aid of biological multi-enzyme systems; carrying out centrifugal filtration; controlling backflow; carrying out multifunctional membrane concentration and enrichment; carrying out spray drying; carrying out qualitative analysis and quantitative analysis. The method has the advantages that the method for quickly, efficiently and economically separating the cordyceps polysaccharides is created; the cordyceps polysaccharides are prepared by the aid of functional membrane separation technologies and green extraction technologies such as enzyme processes, accordingly, the purity and the quality of the cordyceps polysaccharides can be guaranteed, the cordyceps polysaccharides can be quickly enriched and concentrated, and the biological activity of the cordyceps polysaccharides can be guaranteed; the waste culture media for the cordyceps militaris can be effectively utilized, and accordingly resources and the costs can be saved; the method is suitable for large-scale industrial production; the content of the cordyceps polysaccharides in polysaccharide concentrated liquid obtained by the aid of the method by means of separation can reach 80% at least, and the product recycle rate can reach 85% at least.

The invention discloses a method for extracting eyebrow areas. The method comprises the following steps: 1) graying a face image to be identified; 2) segmenting the image in accordance with the rule that the face is divided into three equal parts in length and into five parts in width, so as to primarily select the left and right eyebrow areas; 3) calculating the differences of the each row of adjacent pixel points of the primarily selected left and right eyebrow areas in sequence, and generating a difference curve of the gray value; 4) taking the maximum value and the minimum value of the difference curve of each row as the upper side edge and the lower side edge of corresponding eyebrows; and 5) gathering the upper side edge and the lower side edge, and extracting the eyebrow areas from the primarily selected left and right eyebrow areas according to the upper side edge and the lower side edge of the eyebrows. According to the method, the eyebrow area can be precisely extracted while the lighting influence is effectively reduced.

The invention discloses a human face contour extracting method, wherein the difficulty of subsequent treatment is reduced by standardizing the size of an image to be treated; and the influence of illumination on human face extraction is reduced according to the human face natural quality of three parts and five holes and with dependence on the average contours of human faces. Therefore, the human face contour is fitted through the intersection point of a ray passing through the nasal tip with the human face contour and then by finding out a demarcation point; and the accuracy is higher.

The invention provides an SAR image registration method based on an MRF image segmentation algorithm, and is used for solving the technical problem that an existing feature-based SAR image registration method is low in registration efficiency and poor in stability. The SAR image registration method comprises the following steps: segmenting an SAR image reference diagram and a to-be-registered diagram by using the MRF image segmentation algorithm; carrying out regional interception on the segmented reference diagram and the segmented to-be-registered diagram; respectively corresponding the intercepted segmented reference diagram image blocks and the intercepted segmented to-be-registered diagram image blocks into an SAR image reference diagram and a to-be-registered diagram; establishing SAR-Harris scale space of the SAR image reference diagram image blocks and SAR-Harris scale space of the to-be registered diagram image blocks; extracting SARSIFT feature points of the reference diagram image blocks and the to-be-registered diagram image blocks to form a matching point pair set; removing mismatched points by using a RANSAC algorithm; and optimizing the matching point pair set by using a mutual information method.

The invention discloses a voice data extraction method and a voice data extraction device, relates to the field of voice identification and aims to solve the problem of inaccurate voice identification in the prior art. The technical scheme disclosed by the embodiment of the invention comprises the following steps of: acquiring an average noise value of the environment in which a voice device is positioned; after a user starts the voice device, segmenting a signal input by the user according to preset time to obtain at least one signal segment; and acquiring voice data from the signal input by the user according to the relation between an average audio value and the average noise value, which corresponds to at least one signal segment. The technical scheme provided by the embodiment of the invention can be used in a voice identification system.

The invention discloses an ECG identity recognition method based on an IWT and AGA-BP model. The ECG identity recognition method specifically comprises the steps of 1, collecting and denoising electrocardiosignals; 2, positioning an R wave peak point of the denoised ECG signal by adopting a wavelet positioning method; 3, determining the position of a QRS wave group through the R wave peak point obtained in the step 2, and determining the peak points, starting points and ending points of the P waves and the T waves; and step 4, combining the peak points, the starting points and the ending points of the QRS wave groups, the P waves and the T waves obtained in the step 2 and the step 3 to obtain a feature vector, and then performing ECG signal identification by using the AGA-BP model. According to the ECG identity recognition method based on the IWT and AGA-BP model, the problem that in the prior art, the electrocardiosignal recognition accuracy is poor is solved.

The invention belongs to the technical field of glutamic acid production, and discloses a production method of high-transmittance glutamic acid. The key technology comprises separating mycoprotein from a glutamic acid fermentation liquor through a disc centrifuge and carrying out a film filtration and activated carbon adsorption process to further separate mycoprotein and macromolecule impurities. Mycoprotein is subjected to spray granulation to produce mycoprotein feed, and a clear liquid is subjected to a concentration continuous iso-electric method to separate glutamic acid. The separated glutamic acid crystal is subjected to crystalloblastesis. The crystal undergoes a vacuum filter and is washed with clear water, impurities located in the crystal are removed through washing, and the wet glutamic acid is finally dried. The glutamic-acid-containing evaporated condensate water and washed crystal produced in the process are re-used for film filtration and taken as fermentation liquor washing water for cyclic utilization. The loss of glutamic acid is reduced, and the water cyclic utilization rate is increased.

The invention relates to a radar multi-target identification method based on a high-resolution 1D range profile, and the method is mainly suitable for radar multi-target identification of a early-warning detection system. The method mainly comprises that data of the 1D image profile is preprocessed; an energy aggregation area is extracted from the 1D image profile; strong scattering points are extracted from the energy aggregation area of the 1D image profile; distribution characteristic of the strong scattering points is estimated; the radial length and attitude of the energy aggregation areaof the 1D image profile are estimated; and multi-target attribute of the radar is discriminated. The method is characterized by easy engineering realization, good multi-target identification effect and sufficient theoretical bases, and the problem in radar multi-target attribute identification of an air fleet target or a group target can be solved rapidly and effectively.

The invention discloses a method for extracting colorless pectin by using Jingxi Crataegus pinnatifida Bge, wherein the method comprises: raw material pretreatment preparation, leaching, concentration, alcohol precipitation, pectin preparation and other steps. According to the present invention, the leaching liquid is maintained in the acidic environment so as to effectively reduce the interference of the impurities in the leaching liquid on the pectin extraction; by adding the complex, the blocking on the pectin by the calcium ions and the magnesium ions in the leaching liquid is reduced; byselecting the Jingxi Crataegus pinnatifida Bge to extract the colorless pectin, the characteristic of the extremely low pigment content of the Jingxi Crataegus pinnatifida Bge can eliminate the decolorization step of the preparation process; and the galacturonic acid content in the pectin powder can reach more than 87%, the gelation degree can reach more than 171%, and the utilization rate of theraw material pectin can reach more than 92%.

The invention discloses a method for improving the efficiency of ferulic acid and p-coumaric acid separation. According to the method, basic hydrolysis and acid hydrolysis are adopted in sequence to separate ferulic acid and p-coumaric acid from bagasse cell walls, and GC-MS quantitative analysis is carried out on the content of ferulic acid and p-coumaric acid. According to the method, a comprehensive method including basic hydrolysis and acid hydrolysis in sequence is adopted to separate ferulic acid and p-coumaric acid from bagasse cell walls, thereby eliminating interference of hemicellulose and lignin; ethyl acetate is taken as an extraction solvent, thereby obviously improving the yield of ferulic acid and p-coumaric acid; and GC-MS fast efficient quantitative analysis is carried outon ferulic acid and p-coumaric acid. The method establishes the foundation for efficient separation and industrial application of ferulic acid and p-coumaric acid.

The invention relates to an installing method for a collapse prevention pipe of a horizontal well section of a mining area L-shaped ground well and belongs to the field of mining area L-shaped groundwell coal bed gas extraction. At least one collapse prevention pipe unit and a variable diameter pup joint are provided, and ventilation holes are formed in the collapse prevention pipe units. The method comprises the following steps of calculation, wherein the number of the collapse prevention pipe units is calculated according to the length of the horizontal well section; pipe lowering, whereinthe collapse prevention pipe units and the variable diameter pup joint are lowered into the well in sequence through a drilling machine; separation, wherein a drill rod is rotated in the connecting thread loosening direction, and connecting threads on the variable diameter pup joint and the collapse prevention pipe units are separated; and lifting, wherein the separated variable diameter pup jointis lifted to the surface of the ground along with the drill rod. By means of the installing method, rock debris of collapse holes of the horizontal section of the mining area L-shaped ground well canbe supported to prevent hole blocking, the conical collapse prevention pipe can reduce the resistance when a collapse prevention pipe body enters the horizontal well section, and the rock debris in the horizontal well section is contained by a gap at a circular truncated cone and the tail end of the horizontal well section.

The invention discloses a diabetic retinopathy automatic screening method based on naive Bayes and a support vector machine, which comprises the following steps: preprocessing an original color fundusimage, and extracting a blood vessel image from the original color fundus image; redrawing the pre-processed image according to the blood vessel image, wherein a blood vessel pixel value in the pre-processed image is determined by non-blood vessel pixels in the surrounding area to obtain a blood vessel redrawn image for reducing the blood vessel contrast; extracting a candidate microaneurysm areafrom the vascular redrawing image, and then extracting features of the candidate microaneurysm; and outputting, by a naive Bayesian classifier, a mark matrix of the microaneurysm according to the candidate microaneurysm features so as to obtain the number of the microaneurysm; and giving, by the support vector machine, a glyconet disease detection result according to the preprocessed image and the number of the microaneurysm. The method can effectively avoid accurate segmentation of the blood vessel and the microaneurysm, the extracted features are simple and easy to operate, and whether a patient suffers from the diabetic retinopathy or not can be automatically detected.

The invention discloses a rapid detection method for an SAR image flood inundation area, and aims at the problems of low efficiency and slow response of a traditional flood detection method, the SAR image is enhanced, noise information is removed, gray features and texture features are extracted, a water area of the SAR image is segmented by using an FCM clustering algorithm, and the flood inundation area of the SAR image is obtained. A morphological filtering processing and connected domain analysis method is used to segment an adhesion area of a lake and a river, a length-width ratio and an area threshold value are set, pits and ponds and the river are removed, the lake is reserved, DEM topographic information is used to remove a shadow area generated by a mountain body, false alarm is eliminated, SAR images of two time phases before and after flood are compared, and the method is used to determine whether the flood occurs or not. By using an overlay analysis method and a space overlapping-based block parallel technology, the processing efficiency is improved, the detection speed of a flood inundation area is accelerated, the area, distribution and ground feature type information of the flood inundation area is quickly and accurately obtained, the change of the lake inundation area is detected, and a scientific basis is provided for flood control and disaster relief in time.

The invention discloses a traditional Chinese medicine quality determination system based on near infrared spectroscopy. The system comprises an extraction tank, a liquid pumping mechanism, a liquid storage cavity, a drawing cavity, a fine adjustment cavity and a detection mechanism, wherein a middle lower part of a side wall of the extraction tank is provided with a conduit in a penetrating and communicating manner, the liquid pumping mechanism comprises a catheter cover detachably connected with a catheter port and a dip pipe penetrating through the catheter cover, a drawing piston is arranged in the drawing cavity in a sealing and sliding manner, a fine adjustment piston is arranged in the fine adjustment cavity in a sealed and sliding mode, a front end of the drawing cavity, a front end of the fine adjustment cavity and the top of the other end of the liquid storage tank are communicated with one another, the detection mechanism comprises a detection container for bearing a to-be-detected sample from a sampling port, a near infrared spectrometer and a detection probe connected with the near infrared spectrometer, the detection probe is used for detecting the to-be-detected sample in the detection container, and the volume of the liquid storage cavity is larger than that of the detection container. The system has advantagest of effectively reducing influence of the liquid medicine in the flow path on detection of the next time period in the detection process of the previous time period.

The invention provides a method for extracting and purifying fustin from fruit of rhus punjabensis. The method comprises the steps of: smashing the fruit of rhus punjabensis, adding cellulase and aspergillus niger for fermenting degradation, conducting alkali treatment and extraction through reflux of alcohol and water, conducting filtering, conducting reduced pressure concentration on obtained filtrate until the filtrate is dried, dissolving the dried filtrate in ethanol, making the obtained ethanol solution to pass through a macroporous resin column, after eluting, conducting reduced pressure concentration on the obtained eluate to obtain a concentrated solution, and conducting crystallizing and recrystallizing to obtain the pure fustin product. Through an enzymatic method and alkali extraction, the extraction rate of the fustin is greatly improved, and the influence of the cellulose in the rhus punjabensis on the extraction is reduced; by combining column chromatography with crystallization for separation, the purity of the fustin is greatly improved, the operation steps of purification are simplified, and the production cost of the fustin is reduced.

The invention discloses an infrared image sea-sky-line detection method based on a linear lookup matrix. The method comprises the steps: firstly carrying out column gradient calculation of an image in a large scale through customized template operation, and increasing the gray value of an image of a sea-sky-line region; secondly carrying out the processing of the image through employing a nonlinear segmentation method, and reducing the interference from water surface ripples and noise; finally carrying out the scanning of the image, generating the linear lookup matrix, and finding the maximum value of the matrix so as to locate the sea-sky-line position. The method is simple in implementation, is small in calculation amount, is clear and definite in physical significance, can detect horizontal and inclined sea-sky lines, and is easy to use in engineering. The test indicates that the method can effectively extract the sea-sky line in an infrared image, and has the certain anti-interference capability.