100results about How to "Avoid subjective influence" patented technology

The invention provides an automatic optical detection method, device and system of a mobile phone camera module. The method comprises the following steps of: a. connecting a mobile phone camera module with a control computer and reading a test picture shot by the mobile phone camera module; b. judging whether the mobile phone camera module has serious screen blurring or not according to the test picture; if not, turning to the step c; otherwise, judging that the mobile phone camera module is an unqualified product; c. sending a white board setting command to a control circuit by the control computer so that a white board is arranged at the shooting position for shooting a white board image; d. after reading the white board image, judging that the mobile phone camera module has no slight screen blurring by a slight screen blurring detection module; then carrying out lens smudginess detection by a smudginess detecting module on the basis of the white board image; and if the smudginess does not exist, the mobile phone camera module is a qualified product. The automatic optical detection method, device and system of a mobile phone camera module can carry out automatic detection on the serious screen blurring, slight screen blurring and smudginess which probably occur on the mobile phone camera module and have the advantages of accuracy, reliability, high detecting efficiency and the like.

The invention discloses a setting method of a content label, which includes steps of acquiring text information related to a multimedia content; performing word segmentation on the text information toobtain every word segment; gathering every word segment to obtain a first gathering result, wherein the first gathering result comprises a word segment group classified by every gathering type and formed by the word segment; extracting a target feature word from the first gathering result and inputting to a machine learning model; acquiring every probability value output by the machine learning model, wherein the machine learning model is acquired through performing semantic analysis training on a sample with corresponding relationship with the text information and the label; every probability value represents the probability of using every target feature word as the label of the text information; selecting the label in accordance with the probability condition according to every probability value; associating the selected label with the multimedia content. The invention further discloses a setting device of the content label and a storage medium at the same time.

The invention discloses a planting area remote sensing estimation method for grain subsidy accounting of the prefecture level and the county level. The method comprises the following steps: P1, conducting sample investigation on the planting area of grain corps in a province; P2, selecting and preprocessing a remote-sensing image; P3, conducting remote sensing recognition on the grain crops; P4, processing mixed pixels; P5, establishing a grain crop planting area estimation model, and estimating the planting area of large quantity of the grain crops in district-and-county-level administrative units.

The invention provides a time sequence image-based ship mark collision automatic detection method. The method comprises the steps of performing down-sampling processing, grayscale processing and binarization processing on a received video image in sequence; by using a Gaussian mixture model, performing foreground and background distinguishing on a current frame after the grayscale processing; performing difference operation with a background model of a continuous time sequence image sequence through the current frame, regarding a part greater than a difference threshold in a difference resultas a motion target region, otherwise, regarding the part as background information; by filtering the motion target region morphologically, selecting an object with a maximum connected region greater than a set value as a suspected ship motion object; and by taking a coincidence degree between the maximum connected region and an alarm region as a judgment standard, judging whether ship mark collision occurs or not. The target region can be automatically monitored and detected, so that the subjective influence is avoided.

The invention provides an automatic mine fault identification method. According to the method, spatial position clustering analysis is carried out on all micro-earthquake events according to mine micro-earthquake location result data, the mine fault surfaces are assumed to be planes, by means of a method of analyzing each cluster characteristic value, the geometrical parameters of the orientation, the length, the width and the like of each fault surface are obtained, faults possibly occurring are given, and the position pictures of the faults are drawn. The automatic mine fault identification method has the advantages of avoiding the subjectivity influence of manual interpretation, and being quantitative, automatic and semi-intelligent compared with a general method of manual interpretation after micro earthquake location.

The present invention discloses a system for observing a space target based on multiple channel cameras. The whole system is formed by an integrated information processing system, a servo control system, a tracking frame, a three-dimensional adjustment mechanism, a real-time image processing system, a display device, five channel cameras and five collection cards which are connected to the five channel cameras in one-to-one correspondence. By using the system, the observation of a dynamic target can be realized, the observation of a star can be realized, and the attitude angle solution methods of four channel cameras in star observation and a high-resolution post image processing algorithm are improved. According to the system, commercial binoculars and commercial camera are used as main components, and the live recording of the space dynamic target and the high-resolution and large vision field imaging of star target observation are realized with low cost.

The invention provides a zero sample image classification method based on a relative attribute random forest. An attribute ranking score model is built for an image with an unknown class according to the relative relationship between image classes and image attributes, the attribute ranking score models for all images serve as training samples to train a random forest classifier, and finally, according to an attribute ranking score for a tested image and the random forest classifier obtained through training, the label of the tested image is predicted. The method of the invention can realize zero sample image classification and has the advantages of high classification recognition rate, strong model stability and the like.

The present invention discloses a method for service quality evaluation of a customer service agent. The method comprises the steps of: extracting all the service dialogues of a customer service agentin a dialogue to be analyzed; obtaining process nodes being in one-to-one correspondence with all the service dialogues according to a process language database to obtain a service process of the customer service agent, wherein the process language database comprises all the process nodes of a standard service process, serial numbers of all the process nodes of the standard service process and language data corresponding to each process node of the standard service process; obtaining a sequence similarity of the service process of the customer service agent and the standard service process according to the obtained all the process nodes and the serial numbers of all the process nodes of the standard service process; and obtaining a process matching degree of the service process of the customer service agent and the standard service process according to the obtained all the process nodes. The method provided by the invention improves accuracy and efficiency of service quality evaluation of the customer service agent. The present invention further provides a system for service quality evaluation of a customer service agent.

Disclosed is a calculation method of permeability coefficient of a shield tunnel lining structure. The calculation method includes the steps of 1), determining geometrical features of a tunnel; 2), determining the permeability coefficient ks of surrounding strata of the tunnel; 2), determining leaking water amount of the tunnel, wherein underground water is leaked into the tunnel through a tunnel connector and is collected through a water-collecting well, and the leaking water amount V1 of the groundwater collected through the water-collecting well is measured; 3), determining exterior surface are S and water collecting time t of the tunnel in water-collecting sections of the tunnel; 4), calculating leakage Ql of the tunnel per unit area per unit time; 5), calculating height of water head HR of the interface of the tunnel and the strata; 6), calculating the average permeability coefficient kl of the tunnel. On the basis of the basic features of leakage water of the shield tunnel, the average permeability coefficient determination method capable of considering the permeability features of the shield tunnel connector is established, and force and deformation effects of tunnel leakage on surrounding environment and the shield tunnel can be analyzed more accurately according to the average shield tunnel permeability coefficient acquired in calculation.

The invention discloses an application type feature generation method, a mobile terminal and a readable storage medium. The application type feature generation method comprises the following steps: acquiring the operation number of times of a preset operation behavior of each single user terminal to target application, and calculating an integral interesting value of each single user terminal to the target application according to the operation number of times of the preset operation behavior and a corresponding weight of the preset operation behavior; carrying out classifying modeling by virtue of a preset modeling model according to the integral interesting value, so as to obtain a feature matrix of the target application; and acquiring type features of the target application according to feature values of different feature types of the application in the feature matrix. By virtue of statistics of operation behaviors of users to the application, a classifying modeling between the users and the application is executed so as to obtain the type features of the application, and the generation process does not need to be manually and subjectively judged, so that the application feature accuracy is improved.

The invention provides a quick nondestructive detection method for leather types based on a terahertz spectrum technology. The method comprises the following steps: S1, preparing a detection bracket suitable for detecting a leather sample by terahertz spectrum; S2, measuring the thickness of a to-be-detected leather sample; S3, measuring a blank sample detection bracket to acquire a reference terahertz time-domain spectral signal and measuring and acquiring a sample terahertz time-domain spectral signal of the detection bracket with the to-be-detected sample; S4, measuring different sites of a same sample for many times, calculating the characteristic parameters, taking the mean value to reduce error, and acquiring the mean value of the terahertz spectral characteristic parameters of the to-be-detected sample; and S5, comparing leather terahertz characteristic databases constructed to determine the types of leather. The method is quick, label-free, free of damage, simple to operate and widely suitable, can overcome subjective effect of an identifier, and can objectively reflect the types of detected leather and leather products.

The invention provides a ray detection method for noodle food, which comprises the following steps of: (1) emitting X-rays by an X-ray generating device, and making the X-rays pass through the detected food; (2) receiving the X-rays passing through the detected food by a CCD camera to form a CCD image of the detected food; (3) controlling an image input module in a computer to acquire the CCD image from the CCD camera; (4) processing the CCD image through an image information reorganization module; and (5) detecting and identifying the detected food, and sending corresponding alarm signals in the presence of bread crack, seasoning loss or foreign matter. The invention also provides a ray detection device for the noodle food for implementing the method. The device comprises a food conveying device, an imaging device, a control circuit and a control computer, wherein an intelligent foreign matter identification system is arranged in the control computer. The method and the device can detect, identify and alarm the bread crack, seasoning loss or foreign matter of the noodle food, and have the advantages of safety, accuracy, high efficiency and the like.

The invention discloses a magnetic resonance image analysis-based method for detecting P53 protein expression in a brain tumor. The method comprises the following steps: (1) collecting magnetic resonance images of a T1 weighted sequence, a T1 reinforced sequence and a fluid attenuated inversion recovery (FLAIR) sequence of a patient with brain tumor; (2) intercepting lesion area images in the magnetic resonance image from anyone of the sequences, combining the lesion area images into a lesion area image set, and marking the lesion area images as P53 protein expression positive or negative; (3) analyzing images of the lesion area image set, extracting image characteristics in the lesion area images, and combining the image characteristics into a lesion area sample set; (4) randomly selecting part of samples from the lesion area sample set as a training sample set, and combining other samples into a verification sample set, and training a classifier by using the training sample set; and(5) classifying the verification samples by using the trained classifier to obtain the situation of the brain tumor P53 protein expression of the verification samples.

The invention provides a sleep monitoring system. The sleep monitoring system comprises a main control module, a weak electrophysiology signal detection module, a photoelectric volumetric wave detection module and a respiration signal detection module, wherein the main control module is arranged in sleep monitoring equipment and used for transmitting a signal acquisition command to the weak electrophysiology signal detection module, the photoelectric volumetric wave detection module and the respiration signal detection module; the weak electrophysiology signal detection module is used for acquiring the weak electrophysiology signal of a user according to a command and transmitting the weak electrophysiology signal to the main control module; the photoelectric volumetric wave detection module is used for acquiring a blood oxygen signal of the user according to the command, transforming the blood oxygen signal to a pulse wave signal, and transmitting the pulse wave signal to the main control module; the respiration signal detection module is used for acquiring the respiration signal of the user according to the command, performing noise reduction amplifying, and transmitting the respiration signal after noise reduction amplifying to the main control module; and the main control module is used for performing analysis on the signals according to a pre-set sleep analysis method to generate a sleep analysis result. The sleep monitoring system disclosed by the invention is simple in structure, wide in application range and convenient to use, the function of automatically monitoring and analyzing sleep is realized, the influence of manual operation to the monitoring result is avoided, and the user can use the sleep monitoring system at any sleep place.

The invention provides a dynamic intuitionistic fuzzy cognitive map construction method, a time sequence prediction method and a time sequence prediction system. The dynamic intuitionistic fuzzy cognitive map construction method comprises the following steps: clustering standardized time sequence data into nodes of an intuitionistic fuzzy cognitive map by utilizing fuzzy C-means clustering; training a weight matrix of the intuitionistic fuzzy cognitive map to obtain an initial intuitionistic fuzzy cognitive map; receiving new standardized time sequence data and inputting the new standardized time sequence data into the initial intuitionistic fuzzy cognitive map; adjusting the position of a clustering center of the intuitionistic fuzzy cognitive map by utilizing dynamic fuzzy C-means clustering to obtain a dynamic intuitionistic fuzzy cognitive map; in the process of adjusting the position of the cluster center of the intuitionistic fuzzy cognitive map, judging whether the data at the current moment has concept drift or not by utilizing drift detection if the data at the current moment successfully fall into an existing cluster and an error between a real value solved by utilizing the adjusted intuitionistic fuzzy cognitive map and a predicted value is greater than a preset error threshold value and if so, adjusting a weight matrix of the intuitionistic fuzzy cognitive map; otherwise, not changing the intuitionistic fuzzy cognitive map.

The invention relates to an automatic index screening method for ecological health diagnosis. A netted ecological health diagnosis index system is introduced, and a new index weight determination method is combined. an automatic index screening method is provided; Namely, a 'net'-shaped ecological health diagnosis index system is constructed to replace a traditional 'crotch'-shaped index system toreflect the real cross connection of a complex system, and the problems that abstract layer index weights cannot be objectively obtained and weight distribution is unreasonable traditionally are solved through combination of principal component analysis, an entropy weight method and a secondary weight distribution method. And the problems of random and redundant setting of traditional indexes aresolved through an automatic index screening model. According to the method, the real cross connection of the ecosystem system can be reflected, the quantification and automation of the index screening process are realized, the subjective influence of people in the index weight calculation and index screening process is avoided, and the scientificity and objectivity of the index screening processare improved.

The invention relates to the technical field of musical instrument auxiliary devices, and particularly discloses a music score automatic page-turning method. The music automatic page-turning method includes the steps that a music score and an information stream of mechanical perform sound in correspondence with the music score are established; a page-turning mark is arranged at the former preset time position near the end of mechanical performing sound in correspondence with every page of the music score; the information stream is started to analyzed; and live performance sound is entered in contrast to the mechanical performing sound of the information stream, and when the live performance sound is performed to the page-turning mark, next page of the music score is automatically switched.The music score automatic page-turning method is realized by intelligent display equipment and is in contrast to previously entered mechanical performing sound while live performance sound which is processed by reception of present musical instruments is entered, next page is automatically switched when the live performance sound is performed to the page-turning mark, no artificial page-turning is really achieved, subjective affection to performers or audience can be avoided, and abundant labor cost is saved.

The embodiment of the invention discloses a house resource price determination method and device, equipment and a medium. The method comprises the steps of determining house resource feature information of a to-be-queried house resource; determining cell feature information of a cell where the to-be-queried house resource is located; inputting the housing resource feature information and the community feature information into a predetermined housing resource price prediction model to obtain a predicted price of the to-be-queried housing resource, wherein the house resource price prediction model is obtained by training a preprocessed house resource sample, and the predicted price comprises a selling price or a leasing price. According to the scheme, automatic estimation can be carried outthrough the predetermined house resource price prediction model according to the multi-dimensional feature information of the to-be-queried house resource, thereby achieving the objective prediction of the house resource price, avoiding the subjective impact of a house valuator during the valuation of the house resource, and guaranteeing that the predicted house resource price is as close as possible to the actual house resource price.

The invention provides an automatic focusing method of a cell phone camera module, an apparatus thereof and a system thereof. The method comprises the following steps that: a, initialization is carried out and an observation frame is arranged; a camera module enters a focusing station and equipment connection is established; b, a focusing value corresponded to the observation frame is calculated; c, a correct rotation direction of a focusing motor is determined; and d, according to the correct rotation direction of the focusing motor, rough focusing and fine focusing are carried out on the camera module so as to complete focusing. According to the invention, an automatic focusing process can be carried out on a cell phone camera module and the method has advantages of accuracy, reliability and high efficiency and the like. In addition, an image signal processing technology is employed; therefore, a focusing efficiency is improved; it can be avoided that a result by manual focusing is influenced by subjective consciousness of an operator; stability of product quality is enhanced; and burdens of the operator and labor costs of an enterprise are reduced.