CN104914082A - UV-induced fluorescence multi-spectral imaging ocean oil spill type identification method - Google Patents
UV-induced fluorescence multi-spectral imaging ocean oil spill type identification method Download PDFInfo
- Publication number
- CN104914082A CN104914082A CN201510260542.7A CN201510260542A CN104914082A CN 104914082 A CN104914082 A CN 104914082A CN 201510260542 A CN201510260542 A CN 201510260542A CN 104914082 A CN104914082 A CN 104914082A
- Authority
- CN
- China
- Prior art keywords
- oil
- image
- planted
- recognition methods
- oil spill
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Landscapes
- Investigating, Analyzing Materials By Fluorescence Or Luminescence (AREA)
Abstract
The invention discloses an ocean oil spill type fast identification method. The method utilizes a UV source-induced fluorescent light source system, a multi-spectral camera system and an image identification system. The UV source-induced fluorescent light source system comprises UV zone light sources with three wavelengths of 254nm, 302nm and 365nm. The multi-spectral camera system comprises 8 filters installed in front of a common camera. The filters have wavelengths of 365, 404, 410, 420, 435, 450, 546 and 577. The filter camera acquires images. The image identification system carries out oil type identification on image-acquired 6 oil products (comprising gasoline, diesel oil, kerosene, machine oil, crude oil and plant oil) according to acquired image spot RGBHSV color mean value as a characteristic. Oil type identification is carried out by K-mean clustering subjected to PCA optimization and ICA characteristic optimization and SVM identification. The method is economic and simple and can realize fast identification of ocean oil spill types.
Description
Technical field
The present invention relates to a kind of marine oil spill oil used in environmental area and plant discrimination method, relate to one specifically and carry out oil kind knowledge method for distinguishing by uviol lamp induced fluorescence multispectral imaging.
Background technology
In recent years, marine oil spill Frequent Accidents, planting oil spilling oil is promptly and accurately analyzed and differentiates, determine attribution of liability, trace illegal oil spilling source, correspondence takes suitable emergency response, is a complexity and the challenging work of tool.Existing Oil spill identification standard (GB/T 21247-2007), the laboratory chemical being representative mainly with phase Gas Chromatography/Mass Spectrometry Analysis differentiates that means are representative, although can plant accurate Analysis to oil, but detection speed is slow, and cost is high.
Spectral analysis in recent years becomes the emerging means of Oil spill identification, is subject to common concern.Especially, near infrared spectrum (NIR) is widely used in oily kind discrimination process, (spectrometer and the spectral analysis such as Wang Li, 2004.12), near-infrared spectrum technique is utilized to differentiate analog sea oil spilling, to the simulated seawater sample having prepared 56 gasoline, diesel oil, lubricating oil voluntarily, correct decision oil spilling classification.Oil product has fluorescence phenomenon under burst of ultraviolel, Wang Chunyan etc. (analytical test journal, 2014.3), proposes to use based on concentration parameter synchronous fluorescent spectrum technology, can realize the Accurate classification of the different oil spilling types in laboratory and different oil sources crude oil.But above-mentioned technical literature means all can not accomplish scene, fast in-situ investigation in essence, and equipment is heavy, and instrument precision costly.Marine oil spill is accident often, and the quick in situ detection being carried out marine oil spill by remote sensing is the important component part responded fast.But traditional remote sensing, as spaceborne, carried SAR etc., can only detect oil spill area, each oil can not be distinguished and plant, and affected by environment larger.
The invention provides a kind of kind by a kind of easy device realization oil and know method for distinguishing, the quick discriminating of oil spilling oil kind can be realized.
Summary of the invention
Technical matters to be solved by this invention is just to provide a kind of marine oil spill based on ultraviolet induced fluorescence multispectral imaging oil and plants recognition methods.
The present invention adopts following technical scheme:
A recognition methods planted by marine oil spill oil based on ultraviolet induced fluorescence multispectral imaging, and the method step is as follows:
1. the proving installation of this method: experimental provision used in the present invention comprises a lamp box, the ultraviolet lamp tube of 3 kinds of wavelength is installed in lamp box, centre wavelength is respectively 254,302,365nm, wherein the uviol lamp of 254nm and 365nm wavelength is above lamp box, 302 bottoms being contained in lamp box, often kind of wavelength is respectively the ultraviolet lamp tube of 2 15W, (certainly also can place the fluorescent tube of many different wattages), the filter plate of respective wavelength is placed in fluorescent tube front respectively, place quartz glass ware above the filter plate of wherein 302nm, innerly place oil spilling sample to be measured.Camera is arranged in the middle of the top of lamp box, installs the runner of 8 wave filters additional, lay the narrow band filter slice of 8 wavelength before camera, wavelength is respectively 365,404,410,420,435,450,546,577nm.Camera is connected by USB with computing machine, can by the Image Real-time Transmission of shooting to computer disposal.
2. image acquisition: image acquisition comprises two parts, training set image acquisition, preparing multiple sample (as prepared 1 gasoline, 2 diesel oil, 3 kerosene, 4 machine oil, 5 crude oil, 6 vegetable oil six kinds) is placed in circular glass ware respectively, each kind 50 samples, respectively under 254 reflections, 302 transmissions, 365 reflective light sources, rotate filter wheel under 577-365 totally 8 kinds of filter plates, two photos taken by each sample; After shooting, to image be numbered (as 5 machine oil, light source 254nm, filter plate 577nm, is numbered 5-254-577-1.jpg by the 1st), collected specimens photo is 1140 pairs (* 3 wavelength light source * 8 kinds of filter plate * 2 secondary * 50 samples planted by 6 oil) altogether, by Image Saving.
3. Image semantic classification: pre-service is mainly extracted the effective light spot information of oil sample, 1) pixel due to image acquisition is larger, for 3000*4000=12000 pixel, and oil sample is mainly distributed in centre position, so preferential, image is carried out 1500*1500 pixel in the middle of cutting, such compression of images is original 1/4th; 2) contrast is drawn enhancing, is got adaptive threshold, binary image, and carries out holes filling and opening and closing operation; 3), after district of image UNICOM mark, if the district of UNICOM being greater than 10000 pixels can be extracted, then extract maximum UNICOM district's boundary rectangle four angular coordinates, and as four points, subimage is plucked out.If can not extract, then do nothing; 4) by image stretch to 800*800,5) result of Image semantic classification is extracted by extraneous for image spot rectangular image, size is 800*800 pixel, but this light spot image includes the information such as background, glass dish, should remove, be slightly less than in hot spot circle that to connect rectangle be block of information 500*500 so first get; 6) name of light spot image according to original image is stored.
4. image characteristics extraction: because each oil sample is repeated twice shooting, so each kind 50 oil samples can obtain 100 secondary subimages, the characteristics of mean of RGB, HSV6 the component of (3 kinds of light sources, 8 filter plates) under asking 24 states of 50 samples respectively.Obtain 6 oil thus to plant, 50 samples planted by each oil, each sample 24 states, and the eigenmatrix Data of each state 6 color characteristics of mean, is saved in Excel form, in order to identifying use below.
5. characteristics of image optimization: identifying, first need to carry out standardization (centralization and albefaction) to data Data, then PCA conversion is carried out, sort according to the contribution rate of main point component, obtain according to Feature Mapping features, stochastic generation initial mixing matrix W between use K-ICA, obtains mixed vector x=features*W; Use K-ICA to try to achieve and separate hybrid matrix Wcca, try to achieve sestimate=Wcca*features', then use support vector machine to classify.
6. image recognition: image-recognizing method adopts supporting vector machine model, and the kernel function selected is radial basis RBF kernel function, and two parameter C and gamma wherein can be provided by the optimizing of grid method.Training pattern is obtained by said process.Testing sample (oil spilling source as illegal in certain) is repeated to the process of above-mentioned 1-5, to image acquisition, pre-service, feature extraction and optimization, obtain test data, test data is inputted the oil kind label that training pattern obtains this oil.
The invention has the beneficial effects as follows:
Recognition methods planted by a kind of oil of the marine oil spill based on ultraviolet induced fluorescence multispectral imaging disclosed in this invention, by the uviol lamp of three wavelength, oil spilling sample is irradiated, image is gathered respectively under 8 narrow band filter slices, by carrying out feature extraction, optimization and identification to image, and then achieve the identification of oil spill type.The device that the method uses is simple, cost is low, and detection speed is fast, quantizes to detect objective, science, is applied to spill response and detects, improve production efficiency.
Accompanying drawing explanation
Fig. 1 is ultraviolet induced fluorescence multi-optical spectrum image collecting device figure of the present invention.
Fig. 2 is that recognition methods process flow diagram planted by marine oil spill of the present invention oil.
Embodiment
In order to make object of the present invention, technical scheme and advantage clearly understand, below in conjunction with drawings and Examples, the present invention is further elaborated.Should be appreciated that specific embodiment described herein only in order to explain the present invention, be not intended to limit the present invention.
Embodiment, please refer to Fig. 1 and Fig. 2, and Fig. 1 schematically illustrates ultraviolet induced fluorescence multi-optical spectrum image collecting device figure, Fig. 2 disclosed in this invention and schematically illustrates marine oil spill disclosed in this invention oil kind recognition methods process flow diagram.
Composition graphs 1, the present embodiment illustrates the ultraviolet induced fluorescence multi-optical spectrum image collecting device of this method, this device comprises 1 lamp box, the ultraviolet lamp tube 2 of 3 kinds of wavelength is installed in lamp box, 6, 10, centre wavelength is respectively 254nm, 302, 365nm, the wherein uviol lamp 2 of 254nm and 365nm wavelength, 10 are contained in above lamp box, the uviol lamp 6 of 302nm is contained in the bottom of lamp box, often kind of wavelength is respectively the ultraviolet lamp tube of 2 15W, (certainly also can place the fluorescent tube of many different wattages), the filter plate 3 of respective wavelength is placed in fluorescent tube front respectively, 7, 9, 5 quartz glass wares are placed above the filter plate of wherein 302nm, the oil spilling sample that inner placement is to be measured.Camera 4 is arranged in the middle of the top of lamp box 1, installs 8 runners of 8 wave filters additional, lay the narrow band filter slice of 8 wavelength before camera, wavelength is respectively 365,404,410,420,435,450,546,577nm.Camera 4 is connected by USB with computing machine 11, can by the Image Real-time Transmission of shooting to computer disposal.
Composition graphs 2 illustrates that recognition methods planted by marine oil spill disclosed in this invention oil:
1. image acquisition: image acquisition comprises two parts, training set image acquisition, preparing multiple sample (as prepared 1 gasoline, 2 diesel oil, 3 kerosene, 4 machine oil, 5 crude oil, 6 vegetable oil six kinds) is placed in circular glass ware respectively, each kind 50 samples, respectively under 254 reflections, 302 transmissions, 365 reflective light sources, rotate filter wheel under 577-365 totally 8 kinds of filter plates, two photos taken by each sample; After shooting, to image be numbered (as 5 machine oil, light source 254nm, filter plate 577nm, is numbered 5-254-577-1.jpg by the 1st), collected specimens photo is 1140 pairs (* 3 wavelength light source * 8 kinds of filter plate * 2 secondary * 50 samples planted by 6 oil) altogether, by Image Saving.
2. Image semantic classification: pre-service is mainly extracted the effective light spot information of oil sample, 1) pixel due to image acquisition is larger, for 3000*4000=12000 pixel, and oil sample is mainly distributed in centre position, so preferential, image is carried out 1500*1500 pixel in the middle of cutting, such compression of images is original 1/4th; 2) contrast is drawn enhancing, is got adaptive threshold, binary image, and carries out holes filling and opening and closing operation; 3), after district of image UNICOM mark, if the district of UNICOM being greater than 10000 pixels can be extracted, then extract maximum UNICOM district's boundary rectangle four angular coordinates, and as four points, subimage is plucked out.If can not extract, then do nothing; 4) by image stretch to 800*800,5) result of Image semantic classification is extracted by extraneous for image spot rectangular image, size is 800*800 pixel, but this light spot image includes the information such as background, glass dish, should remove, be slightly less than in hot spot circle that to connect rectangle be block of information 500*500 so first get; 6) name of light spot image according to original image is stored.
3. image characteristics extraction: because each oil sample is repeated twice shooting, so each kind 50 oil samples can obtain 100 secondary subimages, the characteristics of mean of RGB, HSV6 the component of (3 kinds of light sources, 8 filter plates) under asking 24 states of 50 samples respectively.Obtain 6 oil thus to plant, 50 samples planted by each oil, each sample 24 states, and the eigenmatrix Data of each state 6 color characteristics of mean, is saved in Excel form, in order to identifying use below.
4. characteristics of image optimization: identifying, first need to carry out standardization (centralization and albefaction) to data Data, then PCA conversion is carried out, sort according to the contribution rate of main point component, obtain according to Feature Mapping features, stochastic generation initial mixing matrix W between use K-ICA, obtains mixed vector x=features*W; Use K-ICA to try to achieve and separate hybrid matrix Wcca, try to achieve sestimate=Wcca*features ', then use support vector machine to classify.
5. image recognition: image-recognizing method adopts supporting vector machine model, and the kernel function selected is radial basis RBF kernel function, and two parameter C and gamma wherein can be provided by the optimizing of grid method.Training pattern is obtained by said process.Testing sample (oil spilling source as illegal in certain) is repeated to the process of above-mentioned 1-5, to image acquisition, pre-service, feature extraction and optimization, obtain test data, test data is inputted the oil kind label that training pattern obtains this oil.
Claims (7)
1. recognition methods planted by the marine oil spill oil of a ultraviolet induced fluorescence multispectral imaging, it is characterized in that, the method comprises a set of ultraviolet induced fluorescence multispectral imaging device and recognition methods planted by marine oil spill oil, and described marine oil spill oil is planted recognition methods and comprised the collection of image, the pre-service of image, feature extraction, characteristic optimization and image recognition five steps.
2. ultraviolet induced fluorescence multispectral imaging device according to claim 1, it is characterized in that: comprise a lamp box, install 254 in lamp box, 302, the ultraviolet source of 365nm tri-kinds of wavelength, the filter plate of respective wavelength is placed in fluorescent tube front respectively, the runner of 8 wave filters is installed additional before camera, narrow band filter slice wavelength is respectively 365,404,410,420,435,450,546,577nm, pass to computer disposal by collected by camera image.
3. the collection of the first step image of recognition methods planted by marine oil spill oil according to claim 1, it is characterized in that: prepare more than 6 kinds oil samples, comprise gasoline, kerosene, diesel oil, machine oil, crude oil, vegetable oil, each kind 50 samples, be placed in circular glass ware respectively, respectively under three kinds of light sources, under installing 8 kinds of filter plate situations before camera respectively additional, two photos taken by each sample, be numbered, preserve image.
4. the pre-service of the second step image of recognition methods planted by marine oil spill oil according to claim 1, it is characterized in that: first by the pixel cutting of image middle oil oil sample position out, compression of images is original 1/4th, then contrast draws enhancing, adaptive threshold binary image, carry out holes filling and opening and closing operation, district of image UNICOM marks, if the district of UNICOM being greater than 10000 pixels can be extracted, extract UNICOM's district's boundary rectangle image, if without this district of UNICOM, then do nothing, by image stretch to formed objects, then be slightly less than to get in this image and connect the foursquare subimage of maximum inscribed, the name of light spot image according to original image is stored.
5. the 3rd step feature extraction of recognition methods is planted according to marine oil spill oil according to claim 1, it is characterized in that: because each oil sample is repeated twice shooting, so each kind 50 oil samples can obtain 100 secondary subimages, ask (3 kinds of light sources under 24 states of 50 samples respectively, 8 filter plates) the characteristics of mean of RGB, HSV6 component, obtain 6 oil thus to plant, 50 samples planted by each oil, each sample 24 states, the eigenmatrix Data of each state 6 color characteristics of mean, be saved in Excel form, in order to identifying use below.
6. the 4th step characteristic optimization of recognition methods is planted according to marine oil spill oil according to claim 1, it is characterized in that: first need to carry out standardization (centralization and albefaction) to data Data, then PCA conversion is carried out, sort according to the contribution rate of main point component, obtain according to Feature Mapping features, stochastic generation initial mixing matrix W between use K-ICA, obtains mixed vector x=features*W; Use K-ICA to try to achieve and separate hybrid matrix Wcca, try to achieve sestimate=Wcca*features', then use support vector machine to classify.
7. the 5th step image recognition of recognition methods is planted according to marine oil spill oil according to claim 1, it is characterized in that: image-recognizing method adopts supporting vector machine model, the kernel function selected is radial basis RBF kernel function, two parameter C and gamma wherein can be provided by the optimizing of grid method, training pattern is obtained by said process, testing sample (oil spilling source as illegal in certain) is repeated to the process of above-mentioned 1-5, to image acquisition, pre-service, feature extraction and optimization, obtain test data, test data is inputted the oil kind label that training pattern obtains this oil.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510260542.7A CN104914082A (en) | 2015-05-20 | 2015-05-20 | UV-induced fluorescence multi-spectral imaging ocean oil spill type identification method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510260542.7A CN104914082A (en) | 2015-05-20 | 2015-05-20 | UV-induced fluorescence multi-spectral imaging ocean oil spill type identification method |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN104914082A true CN104914082A (en) | 2015-09-16 |
Family
ID=54083335
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201510260542.7A Pending CN104914082A (en) | 2015-05-20 | 2015-05-20 | UV-induced fluorescence multi-spectral imaging ocean oil spill type identification method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN104914082A (en) |
Cited By (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105300617A (en) * | 2015-11-02 | 2016-02-03 | 国网福建省电力有限公司 | Method for rapidly determining oil leakage of oiling device of transformer station |
| CN108257119A (en) * | 2018-01-08 | 2018-07-06 | 浙江大学 | A kind of immediate offshore area floating harmful influence detection method for early warning based near ultraviolet image procossing |
| CN108663342A (en) * | 2018-03-05 | 2018-10-16 | 中国船舶工业系统工程研究院 | A kind of laser induced fluorescence system and method differentiated for oil kind |
| CN108916587A (en) * | 2018-07-05 | 2018-11-30 | 国网福建省电力有限公司 | Transformer oil oil stain fluorescent image camera system based on three-axis stabilization holder principle |
| CN109002859A (en) * | 2018-07-25 | 2018-12-14 | 郑州轻工业学院 | Sensor array feature selecting and array optimization method based on principal component analysis |
| CN109308498A (en) * | 2018-11-28 | 2019-02-05 | 安徽理工大学 | A kind of small thin rice gruel discrimination method of laser induced fluorescence vegetable oil doping |
| CN111141684A (en) * | 2020-02-19 | 2020-05-12 | 长春理工大学 | Ocean oil spill detection method based on visible light/infrared polarization characteristics |
| CN111492649A (en) * | 2017-12-22 | 2020-08-04 | 国立研究开发法人海洋研究开发机构 | Image recording method, image recording program, data processing apparatus, and image recording apparatus |
| CN111928944A (en) * | 2020-08-18 | 2020-11-13 | 深圳市汇投智控科技有限公司 | Laser ray detection method, device and system |
| CN113049787A (en) * | 2016-10-12 | 2021-06-29 | 纳普科有限责任公司 | Optical fluid sensor for cross-contamination control system |
| KR20230116234A (en) * | 2022-01-28 | 2023-08-04 | 한국해양대학교 산학협력단 | Ship oil leak detection system in low light using color sensor |
-
2015
- 2015-05-20 CN CN201510260542.7A patent/CN104914082A/en active Pending
Cited By (18)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105300617A (en) * | 2015-11-02 | 2016-02-03 | 国网福建省电力有限公司 | Method for rapidly determining oil leakage of oiling device of transformer station |
| CN113049787A (en) * | 2016-10-12 | 2021-06-29 | 纳普科有限责任公司 | Optical fluid sensor for cross-contamination control system |
| CN111492649B (en) * | 2017-12-22 | 2023-02-28 | 国立研究开发法人海洋研究开发机构 | Image recording method, image recording program, data processing apparatus, and image recording apparatus |
| US11423947B2 (en) | 2017-12-22 | 2022-08-23 | Japan Agency For Marine-Earth Science And Technology | Image recording method, image recording program, data processing apparatus, and image recording apparatus |
| CN111492649A (en) * | 2017-12-22 | 2020-08-04 | 国立研究开发法人海洋研究开发机构 | Image recording method, image recording program, data processing apparatus, and image recording apparatus |
| CN108257119A (en) * | 2018-01-08 | 2018-07-06 | 浙江大学 | A kind of immediate offshore area floating harmful influence detection method for early warning based near ultraviolet image procossing |
| CN108257119B (en) * | 2018-01-08 | 2020-09-01 | 浙江大学 | Near-shore sea area floating hazardous chemical detection early warning method based on near-ultraviolet image processing |
| CN108663342A (en) * | 2018-03-05 | 2018-10-16 | 中国船舶工业系统工程研究院 | A kind of laser induced fluorescence system and method differentiated for oil kind |
| CN108916587A (en) * | 2018-07-05 | 2018-11-30 | 国网福建省电力有限公司 | Transformer oil oil stain fluorescent image camera system based on three-axis stabilization holder principle |
| CN109002859A (en) * | 2018-07-25 | 2018-12-14 | 郑州轻工业学院 | Sensor array feature selecting and array optimization method based on principal component analysis |
| CN109002859B (en) * | 2018-07-25 | 2022-07-05 | 郑州轻工业学院 | Sensor array feature selection and array optimization method based on principal component analysis |
| CN109308498B (en) * | 2018-11-28 | 2021-01-01 | 安徽理工大学 | Identification method of laser-induced fluorescent vegetable oil-doped millet soup |
| CN109308498A (en) * | 2018-11-28 | 2019-02-05 | 安徽理工大学 | A kind of small thin rice gruel discrimination method of laser induced fluorescence vegetable oil doping |
| CN111141684B (en) * | 2020-02-19 | 2022-01-14 | 长春理工大学 | Ocean oil spill detection method based on visible light/infrared polarization characteristics |
| CN111141684A (en) * | 2020-02-19 | 2020-05-12 | 长春理工大学 | Ocean oil spill detection method based on visible light/infrared polarization characteristics |
| CN111928944A (en) * | 2020-08-18 | 2020-11-13 | 深圳市汇投智控科技有限公司 | Laser ray detection method, device and system |
| KR20230116234A (en) * | 2022-01-28 | 2023-08-04 | 한국해양대학교 산학협력단 | Ship oil leak detection system in low light using color sensor |
| KR102644188B1 (en) * | 2022-01-28 | 2024-03-05 | 국립한국해양대학교산학협력단 | Ship oil leak detection system in low light using color sensor |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN104914082A (en) | UV-induced fluorescence multi-spectral imaging ocean oil spill type identification method | |
| CN102854192B (en) | Method for detecting apple surface defect based on apple surface defect detecting system | |
| Ke et al. | A review of methods for automatic individual tree-crown detection and delineation from passive remote sensing | |
| CN102997856A (en) | Ocean spilt oil film thickness hyperspectral remote sensing estimation method based on parameter lookup table | |
| CN103900972B (en) | Multi-feature fusion-based meat freshness hyperspectral image visual detection | |
| CN102288613A (en) | Surface defect detecting method for fusing grey and depth information | |
| CN103984939B (en) | A kind of sample visible component sorting technique and system | |
| CN109766909B (en) | Analysis method for aging behavior of shore environment microplastic based on spectrogram fusion | |
| CN104181170B (en) | fruit appearance detection method based on spectral image analysis | |
| CN111339948B (en) | Automatic identification method for newly-added buildings of high-resolution remote sensing images | |
| CN103914707B (en) | Green channel product auxiliary discriminating method based on support vector machine | |
| CN107389638A (en) | A kind of microscopic fluorescent spectral imaging marine phytoplankton original position classifying identification method and device | |
| CN106251328B (en) | A method of pollutants on surface of whole chicken rinses is identified using two-dimentional high spectrum image | |
| CN103345617A (en) | Method and system for recognizing traditional Chinese medicine | |
| CN103839078A (en) | Hyperspectral image classifying method based on active learning | |
| CN106951863B (en) | Method for detecting change of infrared image of substation equipment based on random forest | |
| CN111007013B (en) | Crop rotation fallow remote sensing monitoring method and device for northeast cold region | |
| CN114898097B (en) | Image recognition method and system | |
| CN103413124A (en) | Method for detecting round traffic sign | |
| CN104990892A (en) | Spectrum image lossless identification model establishing method for seeds and seed identification method | |
| CN105044054A (en) | Ocean oil spill fine telemetering method based on blind source separation | |
| CN112730326A (en) | Intelligent identification device and method for rock slices | |
| Kamal et al. | Comparison of Google Earth Engine (GEE)-based machine learning classifiers for mangrove mapping | |
| CN113158969A (en) | Apple appearance defect identification system and method | |
| CN103745231A (en) | Teleutospore image identification method for Tillctia Controversa Kahn (TCK) and allied variety TCT (Tilletia caries (DC.) Tul.) of TCK |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| WD01 | Invention patent application deemed withdrawn after publication | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20150916 |