CN101584624B - Guideboard recognition blind-guide device and method thereof based on DSP - Google Patents
Guideboard recognition blind-guide device and method thereof based on DSP Download PDFInfo
- Publication number
- CN101584624B CN101584624B CN2009100533185A CN200910053318A CN101584624B CN 101584624 B CN101584624 B CN 101584624B CN 2009100533185 A CN2009100533185 A CN 2009100533185A CN 200910053318 A CN200910053318 A CN 200910053318A CN 101584624 B CN101584624 B CN 101584624B
- Authority
- CN
- China
- Prior art keywords
- guideboard
- character
- information
- unit
- image
- 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.)
- Expired - Fee Related
Links
Images
Abstract
The invention relates to a guideboard recognition blind-guide device and method thereof belonging to the field of pattern recognition technique based on DSP. The device includes: an outside information acquisition module for collecting information ahead of visual field and switching the information into greyscale information and outputting the information onto the guideboard detection module; a guideboard detection module for storing and processing the image information collected from the outside information acquisition module, and obtaining the zone related to the guideboard information in the image by image processing method, processing the zone where the guideboard is located using the pattern recognition method, looking for digital audio signal corresponding to the character and outputting the signal to the sound switch module; a sound switch module for switching the digital audio signal into a analog audio signal. The earphone switches the analog audio signal into a sound signal, and outputs to blindman. The device provided by the invention uses DSP for finishing character recognition, ultimately tells blindman by hearing instead of vision effect, and is used for guiding and directing blindman when walking on road.
Description
Technical field
The present invention relates to a kind of devices and methods therefor of mode identification technology.Particularly a kind of guideboard recognition blind-guide device and method thereof based on DSP.
Background technology
As the main thoroughfare that the people obtains external information, vision has played important effect in our life, can run into a lot of difficulties for the blind person of visual deprivation in life.The sidewalk for visually impaired people can indicate road direction and turning position for the blind person, but does not almost have due effect for the judgement of road and direction.Therefore the information that vision is needed offers the blind person by other way of feeling, provides great facility for blind person's democratic meeting.
Find by prior art documents, China's invention, number of patent application 200710039631.4, publication number 101040809, name is called: based on the vision alternative method of cognitive and target identification, this technology is converted into acoustical signal by the classification and the positional information of inspected object in environment with these information, thereby realizes the replacement of audition to vision.This technology based target identification is discerned and voice output at the position of object, but is difficult to identification and can not plays good guide effect for what determine in the road that article and background complexity cause.
" modern electronic technology " (Zhang Ming, 2008,31 (22): deliver " research of license plate automatic identification algorithm " 108-111), in this technology in the dynamic image character recognition only be confined to specific object---on the automotive license plate.And the quantity of information that automotive license plate itself comprises is less, and content is also more single.
Summary of the invention
The objective of the invention is to overcome the deficiencies in the prior art, a kind of guideboard recognition blind-guide device and method thereof based on DSP is provided.The present invention discerns the Chinese character in the common guideboard on the road by the method for pattern recognition, utilizes DSP to finish character recognition, and finally the blind person by the audition announcement is used for guiding and the guiding function of blind person in the road walking to replace visual effect.
The present invention is achieved through the following technical solutions,
A kind of guideboard recognition blind-guide device based on DSP involved in the present invention comprises: external information acquisition module, guideboard detection module and sound modular converter.The external information acquisition module is gathered the information in the place ahead, the visual field and is converted half-tone information to and outputs to the guideboard detection module; The image information that the guideboard detection module collects in the information pickup module is to external world stored and is handled, draw the zone of relevant guideboard information in the image by image processing method, use mode identification method the guideboard region to be handled and obtain character on the guideboard, search this character corresponding digital audio signal and it is outputed to the sound modular converter; The sound modular converter is converted to simulated audio signal with digital audio and video signals.
Described guideboard detection module comprises: guideboard information pretreatment unit, guideboard are sought unit, Character segmentation unit and condition code and template recognition unit.Wherein: guideboard information pretreatment unit, the half-tone information that obtains in the external information acquisition module is converted to gray matrix, carry out pretreatment successively with the method for histogram equalization, binaryzation, removal point-like noise.And the result is outputed to guideboard seek the unit.Guideboard is sought the unit, is used for searching out at pretreated gray matrix the position and the size of guideboard correspondence, and this information is outputed to the Character segmentation unit.The Character segmentation unit, thereby the gray scale difference that utilizes in the guideboard gray values of great majority point and character place dot matrix is handled and is obtained each character zone, to each character zone carry out that denoising, level and smooth, gradient sharpening, texture detect, image rotation, normalized, convert thereof into the character pattern of a normal size, and its result is outputed to condition code and template recognition unit.Condition code and template recognition unit, it is the numbering that character pattern is obtained a sequence by the method for feature identification or template matching, obtain corresponding characters or numeral and in the data base, search out corresponding digital audio and video signals by numbering, it is outputed to the sound modular converter.
A kind of guideboard recognition blind-guide method based on DSP involved in the present invention comprises the steps:
Step 1, discern, make labelling, carry out information processing and calculating for the Chinese and English position and the size characteristic data of guideboard of the same race for dissimilar guideboards for font in the dissimilar guideboards and character library;
Step 2, convert the information in the scene of extraneous the place ahead to digital information by analog digital conversion, the matrix of the half-tone information value of all pixels of the corresponding the place ahead of each two field picture by photographic head;
Step 3, employing histogram equalization, binaryzation, removal point-like noise, gradient sharpening, texture detect, the image rotation is handled and discerned the guideboard profile successively, analyze and labelling for the guideboard feature then;
Step 4, existing labelling rear number plate is carried out Character segmentation by its characteristics, utilize bilinear interpolation method that the figure after cutting apart is carried out standard normalization after obtaining the shared scope of each independent Chinese character, letter and number; The identification of carrying out Chinese character, letter and number by feature identification method and template matching method at last obtains the characteristic of correspondence sequence, notes its content;
Step 5, in the data base, seek corresponding Chinese character and letter, seek the digital audio and video signals in its system database simultaneously, at last this audio signal is converted to acoustical signal output for identifying characteristic sequence;
Described step 3, step 4 are all finished by DSP, seek the digital audio and video signals of Chinese character and letter correspondence in the step 5 and also finished by DSP in the data base.
Compared with prior art, the present invention has following beneficial effect: the present invention utilizes photographic head to carry out the collection of image, comparing existing OCR (Optical Character Recognition optical character recognition) system can make a response to the external world of real-time change and literal identification, improved the service efficiency and the scope of application greatly, also not only be confined to identification Chinese character on the paper and numeral.Existing OCR system major part is based on PC's, can't realize portable processing environment.And the present invention utilizes DSP to carry out portable processing, can reduce its volume and cost greatly based on the realization of DSP, and the higher arithmetic speed of DSP also can guarantee the correctness discerned.Character among the present invention, numeral and Chinese Character Recognition are at guideboard.Because guideboard has specificity and universality, and special labelling and regulation are all arranged on its structure, font, size and the frame.Can improve discrimination and the effectiveness that guideboard is discerned greatly so compare simple object identification or character recognition.
Description of drawings
Fig. 1 is a structured flowchart of the present invention;
Fig. 2 is the characteristic pattern of character in the present embodiment.
The specific embodiment:
Below in conjunction with accompanying drawing embodiments of the invention are elaborated: present embodiment is being to implement under the prerequisite with the technical solution of the present invention, provided detailed embodiment and concrete operating process, but protection scope of the present invention is not limited to following embodiment.
As shown in Figure 1, the blind-guide device of present embodiment comprises: black and white photographic head (being the external information acquisition module), guideboard detection module, sound modular converter and ancillary equipment earphone.The black and white photographic head is gathered the information in the place ahead, the visual field and is converted half-tone information to and outputs to the guideboard detection module; The guideboard detection module is stored and is handled the image information that collects in the black and white photographic head, draw the zone of relevant guideboard information in the image by image processing method, use mode identification method the guideboard region to be handled and obtain character on the guideboard, search this character corresponding digital audio signal and it is outputed to the sound modular converter; The sound modular converter is converted to simulated audio signal with digital audio and video signals; Earphone is converted to acoustical signal with simulated audio signal, exports to the blind person; The guideboard detection module is realized by TMS320DM642 in the present embodiment, and the sound modular converter is realized by the stereo digital signal codec of AIC23.
Described guideboard detection module comprises: guideboard information pretreatment unit, guideboard are sought unit, Character segmentation unit and condition code and template recognition unit.Wherein: guideboard information pretreatment unit, the half-tone information that obtains in the external information acquisition module is converted to gray matrix, carry out pretreatment successively with the method for histogram equalization, binaryzation, removal point-like noise.And the result is outputed to guideboard seek the unit.Guideboard is sought the unit, is used for searching out at pretreated gray matrix the position and the size of guideboard correspondence, and this information is outputed to the Character segmentation unit.The Character segmentation unit, thereby the gray values that utilize in the guideboard great majority point are handled with the gray scale difference of character place dot matrix and are obtained each character zone, each character zone are carried out denoising, level and smooth, gradient sharpening, texture detect, image rotates, normalized.Convert thereof into the character pattern of a normal size, and its result is outputed to condition code and template recognition unit.Condition code and template recognition unit, it is the numbering that character pattern is obtained a sequence by the method for feature identification or template matching, obtain corresponding characters or numeral and in the data base, search out corresponding digital audio and video signals by numbering, it is outputed to the sound modular converter.
The blind-guiding method of present embodiment comprises the steps:
Step 1, classify and sum up, determine font and character library in the guideboard for dissimilar guideboards.Carry out classification designator for dissimilar guideboards, estimate for the Chinese and English position and the size of guideboard of the same race;
Common guideboard is divided into two kinds of crossing direction board and road direction boards, wherein the crossing direction board generally uses at the crossing, be used to indicate each direction road corresponding title behind the bifurcated of crossing, its main feature has arrow in the middle of being, thus need in the identifying to identify earlier behind the arrow for its up and down four direction repartition.And the road direction board is used to indicate the road name and the direction (direction that specifically refers to the four corners of the world) of place road, its characteristics are for being divided into two row, behavior Chinese character indication wherein, below the English indication of a behavior, identify in the identifying and can be divided into a plurality of zones behind the demarcation line of up-downgoing correspondence and discern.
Reading in of step 2, external information: convert the information in the certain scene in extraneous the place ahead to digital information by analog digital conversion by photographic head, each two field picture corresponding expression the matrix of half-tone information value of all pixels of the place ahead.Be specially: according to the YCbCr video formats image is gathered (wherein Y represents luminance component, and Cb represents the chroma blue component, and Cr represents the red color component), cast out Cb, obtain simple gray value after the Cr value.The ranks number at respective pixel place is stored in the two-dimensional array.The image of different frame then corresponding stored in different arrays;
Step 3, the pretreatment of guideboard information: scene information carries out the contour area that analyzing and processing obtains guideboard thereby adopt histogram equalization, binaryzation, removal point-like noise, gradient sharpening, texture detection, image to rotate successively to external world.Analyze for the zone at guideboard place then and sort out.If a plurality of guideboards guideboard of priority treatment area occupied maximum then, its detailed process is: at first, thereby image is carried out the average gray value that histogram equalization obtains entire image; Directly utilize this gray scale to be threshold value for first frame, resulting target area was calculated average gray value and is set to threshold value after other frames utilized the previous frame Flame Image Process; By middle threshold value image is carried out binaryzation, obtain black white image and remove the point-like noise,, need carry out the gradient sharpening in order to reach the fine resolution of character edge energy; Do not have big information dropout in order to reach character itself, need carry out texture and detect; At last, image being carried out the integral body rotation drops to minimum so that photographic head is discerned issuable interference to next step;
The reading and discerning of step 4, character: the existing rear number plate of sorting out is carried out Character segmentation by its characteristics, utilize bilinear interpolation method that the figure after cutting apart is carried out standard normalization after obtaining the shared scope of each independent Chinese character, letter and number.The identification of carrying out Chinese character, letter and number by feature identification method and template matching method at last obtains the characteristic of correspondence sequence, notes its content.Its detailed process is:
(1) the guideboard information that step 3 is obtained is carried out Character segmentation in conjunction with the guideboard characteristics of storage, obtains the position and the edge at each character place in original image;
(2) utilize bilinear interpolation method that character matrix is carried out standard normalization, convert the character pattern of a normal size to, prepare for next step carries out feature identification or type matrix coupling;
(3) for finishing the normalized character pattern of standard, carry out feature identification and template matching.
Wherein characteristic recognition method is as follows: numeral or character for a fixed size, identical feature is arranged usually, and the feature of using in the present embodiment comprises:
1., character upper end has or not long horizontal about traversing, shown in Fig. 2 a;
2., the character lower end has or not the long horizontal stroke about traversing;
3., have or not the length about traversing horizontal in the middle of the character;
4., character left side or right side have or not the length about traversing perpendicular, shown in Fig. 2 b;
5., the number that from left to right can pass through different lines in the middle of the character, shown in Fig. 2 c;
6., the character centre is passed through a series of features such as different line segment numbers from top to bottom.
For the corresponding different characteristic of different characters, by asking for of character feature discerned corresponding character.
Template matching method is as follows: because character is a standard character on the guideboard, a standard character template is set then, by the primary standard character being carried out the sampling of horizontal 16 vertical 16 points of point, stores in the standard form.Real time imaging is respectively carried out the sampling of 16 points anyhow, obtain the character pattern of the standardized size of 16*16 particular point, as dot matrix to be measured.By tested point battle array and standard form dot matrix are compared, the Character mother plate that matching rate is the highest is thought the result of mating.
(4) characteristic sequence that above-mentioned two kinds of recognition methodss are obtained compares, if the characteristic sequence that two kinds of methods obtain is identical, then thinks to have obtained correct recognition result.
Step 5, in the data base, seek corresponding Chinese character and letter, seek the digital audio and video signals in its system database simultaneously, at last this audio signal is converted to acoustical signal output for identifying characteristic sequence.
In the present embodiment, the ICETEK-DM642-C evaluation board that has adopted the production of Beijing Rui Tai Creative Company is as development platform.This evaluation board is a low-power consumption stand-alone development platform.Equipment mainly comprises the TMS320DM642DSP chip of TIX (TI), independent HPI interface, 2 road video ports, 4M synchronous dynamic random access memory, the stereo digital signal codec of 32MFlash, AIC23, JTAG emulation interface and 4 user lamps on the plate.
Present embodiment is applied to assisting blind and obtains guideboard information and change sound into, is used for guiding and the guiding function of blind person in the road walking.The succinct easily row of present embodiment method, hardware is easy to carry, and the speed of service is fast, for blind person's daily life facilitates.
Claims (4)
1. the guideboard recognition blind-guide method based on DSP is characterized in that, comprises the steps:
1.. discern for font in the dissimilar guideboards and character library, make labelling, carry out information processing and calculating for the Chinese and English position and the size characteristic data of guideboard of the same race for dissimilar guideboards;
2.. convert the information in the scene of extraneous the place ahead to digital information by analog digital conversion by photographic head, the matrix of the half-tone information value of all pixels of the corresponding the place ahead of each two field picture;
3.. adopt histogram equalization, binaryzation, removal point-like noise, gradient sharpening, texture detection, image rotation successively the guideboard profile to be handled and discerned, analyze and labelling for the guideboard feature then;
4.. existing labelling rear number plate is carried out Character segmentation by its characteristics, utilize bilinear interpolation method that the figure after cutting apart is carried out standard normalization after obtaining the shared scope of each independent Chinese character, letter and number; The identification of carrying out Chinese character, letter and number by feature identification method and template matching method at last obtains the characteristic of correspondence sequence, notes its content;
5.. in the data base, seek corresponding Chinese character, letter and number for the characteristic sequence that has identified, seek the digital audio and video signals in its system database simultaneously, at last this audio signal is converted to acoustical signal output;
Described template matching method is: a Character mother plate is set, the sampling of the primary standard character being carried out horizontal 16 vertical 16 points of point is to obtain the Character mother plate dot matrix of a 16*16, store in the standard form, real time imaging is respectively carried out anyhow the sampling of 16 points, obtain the character pattern of the standardized size of 16*16 particular point, as dot matrix to be measured, by the Character mother plate dot matrix in tested point battle array and the standard form is compared, the Character mother plate that matching rate is the highest is thought the result of mating.
2. the guideboard recognition blind-guide device according to the described method of claim 1 is characterized in that, comprising: photographic head, guideboard detection module and sound modular converter; Photographic head is gathered the information in the place ahead, the visual field and is converted half-tone information to and outputs to the guideboard detection module; The guideboard detection module is stored and is handled the image information that collects in the photographic head, draw the zone of relevant guideboard information in the image by image processing method, use characteristic method of identification and template matching method are handled the guideboard region and are obtained the Chinese character on the guideboard, the characteristic sequence of letter and number, search this characteristic sequence corresponding digital audio signal and it is outputed to the sound modular converter; The sound modular converter is converted to simulated audio signal with digital audio and video signals.
3. guideboard recognition blind-guide device according to claim 2 is characterized in that, described guideboard detection module comprises: guideboard information pretreatment unit, guideboard are sought unit, Character segmentation unit and condition code and template recognition unit; Wherein: guideboard information pretreatment unit, the half-tone information that obtains in the photographic head is converted to gray matrix, adopt the method for histogram equalization, binaryzation, removal point-like noise, gradient sharpening, texture detection, image rotation to carry out pretreatment successively, and the result is outputed to guideboard searching unit; Guideboard is sought the unit, is used for searching out at pretreated gray matrix the position and the size of guideboard correspondence, and this information is outputed to the Character segmentation unit; The Character segmentation unit, thereby the gray scale difference that utilizes in the guideboard gray values of great majority point and character place dot matrix is handled and is obtained each character zone, to each character zone carry out that denoising, level and smooth, gradient sharpening, texture detect, image rotation, described standard normalization, convert thereof into the character pattern of a normal size, and its result is outputed to condition code and template recognition unit; Condition code and template recognition unit are that the method for character pattern by feature identification and template matching obtained an identical characteristic sequence, search out corresponding digital audio and video signals by characteristic sequence in the data base, and it is outputed to the sound modular converter.
4. according to claim 2 or 3 described guideboard recognition blind-guide devices, it is characterized in that described guideboard detection module adopts DSP to realize.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2009100533185A CN101584624B (en) | 2009-06-18 | 2009-06-18 | Guideboard recognition blind-guide device and method thereof based on DSP |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2009100533185A CN101584624B (en) | 2009-06-18 | 2009-06-18 | Guideboard recognition blind-guide device and method thereof based on DSP |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN101584624A CN101584624A (en) | 2009-11-25 |
| CN101584624B true CN101584624B (en) | 2011-01-19 |
Family
ID=41369201
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN2009100533185A Expired - Fee Related CN101584624B (en) | 2009-06-18 | 2009-06-18 | Guideboard recognition blind-guide device and method thereof based on DSP |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN101584624B (en) |
Families Citing this family (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101999972B (en) * | 2010-11-24 | 2013-07-03 | 上海理工大学 | Stereoscopic vision based auxiliary walking device for blindmen and auxiliary method thereof |
| CN102090947B (en) * | 2011-01-31 | 2013-09-11 | 中国科学院计算技术研究所 | Blind visual compensation method and system for implementing same |
| CN104778004A (en) * | 2015-03-24 | 2015-07-15 | 深圳市艾优尼科技有限公司 | Information content matching method |
| CN104777931A (en) * | 2015-03-24 | 2015-07-15 | 深圳市艾优尼科技有限公司 | Terminal |
| CN107085945A (en) * | 2016-02-16 | 2017-08-22 | 信亿科技股份有限公司 | Real-time Vehicle License Plate identification system and its method |
| CN107157651A (en) * | 2017-06-13 | 2017-09-15 | 浙江诺尔康神经电子科技股份有限公司 | A kind of visual pattern sensory perceptual system and method based on sonic stimulation |
| CN107320296A (en) * | 2017-06-23 | 2017-11-07 | 重庆锦上医疗器械有限公司 | The space three-dimensional acoustic expression system and method for visual signal |
| CN108009548A (en) * | 2018-01-09 | 2018-05-08 | 贵州大学 | A kind of Intelligent road sign recognition methods and system |
| CN108734167B (en) * | 2018-05-08 | 2022-05-17 | 湖南开启时代电子信息技术有限公司 | Method for recognizing characters on contaminated film |
| CN110222684A (en) * | 2019-04-19 | 2019-09-10 | 黑龙江大学 | A kind of blind person's " reading " system |
| CN212677349U (en) * | 2020-03-31 | 2021-03-09 | 易力声科技(深圳)有限公司 | Monitoring assembly, information processing assembly and earphone |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1746935A (en) * | 2005-07-28 | 2006-03-15 | 上海交通大学 | Monochip computer based on system as place name guide for blind person |
| CN1851731A (en) * | 2006-05-25 | 2006-10-25 | 电子科技大学 | Registration number character dividing method |
| CN201159903Y (en) * | 2008-03-14 | 2008-12-03 | 北京信路威科技发展有限公司 | License plate recognition device |
| CN101339601A (en) * | 2008-08-15 | 2009-01-07 | 张擎宇 | License plate Chinese character recognition method based on SIFT algorithm |
| CN101385677A (en) * | 2008-10-16 | 2009-03-18 | 上海交通大学 | Blind guiding method and device based on moving body track |
| CN101427969A (en) * | 2008-12-18 | 2009-05-13 | 苏州莱迪斯特电子有限公司 | Blind guiding system and method based on Chinese character recognition |
-
2009
- 2009-06-18 CN CN2009100533185A patent/CN101584624B/en not_active Expired - Fee Related
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1746935A (en) * | 2005-07-28 | 2006-03-15 | 上海交通大学 | Monochip computer based on system as place name guide for blind person |
| CN1851731A (en) * | 2006-05-25 | 2006-10-25 | 电子科技大学 | Registration number character dividing method |
| CN201159903Y (en) * | 2008-03-14 | 2008-12-03 | 北京信路威科技发展有限公司 | License plate recognition device |
| CN101339601A (en) * | 2008-08-15 | 2009-01-07 | 张擎宇 | License plate Chinese character recognition method based on SIFT algorithm |
| CN101385677A (en) * | 2008-10-16 | 2009-03-18 | 上海交通大学 | Blind guiding method and device based on moving body track |
| CN101427969A (en) * | 2008-12-18 | 2009-05-13 | 苏州莱迪斯特电子有限公司 | Blind guiding system and method based on Chinese character recognition |
Also Published As
| Publication number | Publication date |
|---|---|
| CN101584624A (en) | 2009-11-25 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN101584624B (en) | Guideboard recognition blind-guide device and method thereof based on DSP | |
| CN109840521B (en) | Integrated license plate recognition method based on deep learning | |
| CN109816644B (en) | Bearing defect automatic detection system based on multi-angle light source image | |
| CN101916364B (en) | Adaptive dial digital identification method | |
| CN102496157B (en) | Image detection method based on Gaussian multi-scale transform and color complexity | |
| CN109871780B (en) | Face quality judgment method and system and face identification method and system | |
| CN105844621A (en) | Method for detecting quality of printed matter | |
| CN103679168A (en) | Detection method and detection device for character region | |
| CN106529532A (en) | License plate identification system based on integral feature channels and gray projection | |
| CN103514448A (en) | Method and system for navicular identification | |
| CN109902618A (en) | A kind of sea ship recognition methods and device | |
| CN105046808A (en) | Banknote multi-spectral high-resolution image acquisition system and acquisition method | |
| CN104408449A (en) | Intelligent mobile terminal scene character processing method | |
| CN105426816A (en) | Method and device of processing face images | |
| Liu et al. | Stroke filter for text localization in video images | |
| CN115082776A (en) | Electric energy meter automatic detection system and method based on image recognition | |
| Chen et al. | Salient object detection: integrate salient features in the deep learning framework | |
| CN204029024U (en) | A kind of bank note face amount recognition system | |
| CN110866435A (en) | Far infrared pedestrian training method with self-similarity gradient oriented histogram | |
| CN113486712B (en) | Multi-face recognition method, system and medium based on deep learning | |
| CN111160080A (en) | Intelligent automobile appearance consistency detection system based on image recognition | |
| CN110222660A (en) | A kind of signature false distinguishing method and system merged based on dynamic with static nature | |
| CN109145912A (en) | A kind of digital instrument reading automatic identifying method | |
| CN111402189B (en) | Video image color cast detection device and method | |
| CN109784176B (en) | Vehicle-mounted thermal imaging pedestrian detection Rois extraction method and device |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant | ||
| C17 | Cessation of patent right | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20110119 Termination date: 20130618 |