CN113208882A - Blind person intelligent obstacle avoidance method and system based on deep learning - Google Patents
Blind person intelligent obstacle avoidance method and system based on deep learning Download PDFInfo
- Publication number
- CN113208882A CN113208882A CN202110465837.3A CN202110465837A CN113208882A CN 113208882 A CN113208882 A CN 113208882A CN 202110465837 A CN202110465837 A CN 202110465837A CN 113208882 A CN113208882 A CN 113208882A
- Authority
- CN
- China
- Prior art keywords
- preset
- picture
- blind
- deep learning
- distance
- 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
Images
Classifications
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61H—PHYSICAL THERAPY APPARATUS, e.g. DEVICES FOR LOCATING OR STIMULATING REFLEX POINTS IN THE BODY; ARTIFICIAL RESPIRATION; MASSAGE; BATHING DEVICES FOR SPECIAL THERAPEUTIC OR HYGIENIC PURPOSES OR SPECIFIC PARTS OF THE BODY
- A61H3/00—Appliances for aiding patients or disabled persons to walk about
- A61H3/06—Walking aids for blind persons
- A61H3/061—Walking aids for blind persons with electronic detecting or guiding means
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01C—MEASURING DISTANCES, LEVELS OR BEARINGS; SURVEYING; NAVIGATION; GYROSCOPIC INSTRUMENTS; PHOTOGRAMMETRY OR VIDEOGRAMMETRY
- G01C3/00—Measuring distances in line of sight; Optical rangefinders
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/80—Analysis of captured images to determine intrinsic or extrinsic camera parameters, i.e. camera calibration
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/40—Scenes; Scene-specific elements in video content
- G06V20/41—Higher-level, semantic clustering, classification or understanding of video scenes, e.g. detection, labelling or Markovian modelling of sport events or news items
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10016—Video; Image sequence
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Life Sciences & Earth Sciences (AREA)
- Health & Medical Sciences (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Software Systems (AREA)
- Artificial Intelligence (AREA)
- General Engineering & Computer Science (AREA)
- Evolutionary Biology (AREA)
- Bioinformatics & Computational Biology (AREA)
- Computational Linguistics (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Multimedia (AREA)
- Electromagnetism (AREA)
- Radar, Positioning & Navigation (AREA)
- Remote Sensing (AREA)
- Evolutionary Computation (AREA)
- Epidemiology (AREA)
- Pain & Pain Management (AREA)
- Physical Education & Sports Medicine (AREA)
- Rehabilitation Therapy (AREA)
- Animal Behavior & Ethology (AREA)
- General Health & Medical Sciences (AREA)
- Public Health (AREA)
- Veterinary Medicine (AREA)
- Image Analysis (AREA)
Abstract
The invention belongs to the technical field of computer deep learning, and provides an intelligent obstacle avoidance method for blind people based on deep learning, which comprises the following steps: acquiring pictures in a preset range through a preset camera; identifying an object existing in the acquired picture according to a preset training model and obtaining position information of the object; measuring the binocular distance of the object target in the acquired picture according to a preset distance measurement algorithm; and broadcasting the object position information and the measured object target distance through voice. The invention also provides an intelligent obstacle avoidance system for the blind based on deep learning, which has the advantages that a plurality of objects appearing in front can be detected and identified during walking, the change of the distance between the plurality of objects and the position of the plurality of objects can be respectively broadcasted, the intelligent obstacle avoidance system can be used for people with visual impairment to find routes indoors and outdoors, find objects indoors and the like, and helps more blind people open another door for knowing the real world.
Description
Technical Field
The invention relates to the technical field of computer deep learning, in particular to an intelligent blind-person obstacle avoidance method and system based on deep learning.
Background
According to the statistics of the world health organization, about 4 million to 4.5 million blind people exist all over the world, the low vision is 3 times of that of the blind people, about 1700 million blind people exist in China at present, the blind people are the most blind people all over the world, and 45 million blind people are newly increased every year, which means that one new blind person appears every minute almost every day. If this speed is left to develop without taking more active and effective measures, the blind becomes a serious social problem. Blind people lose light and suffer great pains in mind and body, and need more attention and help from society and countries. Nowadays, science and technology in the world develop rapidly, and new technical means emerge endlessly, for example, infrared rays and ultrasonic waves used by a blind guiding instrument can only detect the distance of an object, however, the object in front of eyes is still unknown, and in addition, in some underdeveloped cities, medium-sized blind guiding dogs have a plurality of potential safety hazards in the aspects of sitting on buses, sitting on subways, getting on cars and the like, and the technologies and the methods are still difficult to be accepted by the masses.
At present, artificial intelligence slowly permeates the public life, and the public is served by a machine, the core of an intelligent system of the machine is to continuously train, learn and improve through loading a deep learning model so as to achieve a more humanized effect, and the effect displayed by the deep learning model in the aspect of computer vision is more and more accepted by the public, such as object detection, classification and recognition, face recognition and the like. Therefore, how to apply the deep learning model to the blind guiding field to accurately identify the targets appearing around in real time and inform the blind of the targets is a key problem to be solved urgently.
Disclosure of Invention
The invention aims to provide an intelligent obstacle avoidance method and system for blind people based on deep learning, which are used for solving the problems of accurately identifying surrounding targets in real time and informing the blind people;
in order to achieve the purpose, the invention adopts the technical scheme that:
an intelligent obstacle avoidance method for blind people based on deep learning comprises the following steps:
acquiring pictures in a preset range through a preset camera;
identifying an object existing in the acquired picture according to a preset training model and obtaining position information of the object;
measuring the binocular distance of the object target in the acquired picture according to a preset distance measurement algorithm;
and broadcasting the object position information and the determined object target distance through voice, judging the direction of the blind person for avoiding the barrier according to the object position information and the determined object target distance, and reminding the blind person to avoid the corresponding barrier.
Further, the specific steps of collecting the pictures in the preset range through the preset camera include:
acquiring a video within a preset range through a preset camera;
and taking frames of the video to obtain a picture in a preset range.
Further, the specific steps of establishing the preset model include:
acquiring an object picture in a preset mode, and establishing a corresponding data set;
acquiring preset parameters of an object in a picture in a preset mode;
and training the picture through a preset network model, and continuously optimizing network parameters to obtain a preset training model.
Further, the preset parameter of the object is a position of the object, a name of the object, and a type of the object.
Further, the specific steps of performing distance measurement on the object target in the acquired picture according to a preset distance measurement algorithm include:
calibrating a preset camera;
dividing the collected pictures to obtain a left picture and a right picture;
performing coordinate mapping on each picture according to data obtained by camera calibration;
graying the two mapped pictures;
forming a parallax depth map from the two grayed pictures according to a preset function;
reconstructing the depth map to obtain a mapping map;
and calculating the distance between the object and a preset camera through the mapping according to the position information of the object.
The invention also aims to provide an intelligent obstacle avoidance system for blind people based on deep learning, which comprises:
the image acquisition module is used for acquiring pictures in a preset range through a preset camera;
the object detection module is used for identifying an object existing in the acquired picture according to a preset training model and obtaining position information of the object;
the binocular distance measurement module is used for measuring the binocular distance of the object target in the acquired picture according to a preset distance measurement algorithm;
and the voice conversion module is used for broadcasting the object position information and the determined object target distance through voice, judging the direction of the blind person for avoiding the barrier according to the object position information and the determined object target distance, and reminding the blind person of avoiding the corresponding barrier.
Further, the image acquisition module comprises:
the video acquisition unit is used for acquiring videos in a preset range through a preset camera;
and the frame taking unit is used for taking frames of the video to obtain a picture of each frame.
Further, the binocular range finding module includes:
the calibration unit is used for calibrating a preset camera;
the dividing unit is used for dividing the collected pictures to obtain a left picture and a right picture;
the coordinate mapping unit is used for mapping the coordinates of each picture according to the data obtained by camera calibration;
the gray processing unit is used for graying the two mapped pictures;
the depth map construction unit is used for forming a parallax depth map by the two grayed pictures according to a preset function;
the reconstruction unit is used for reconstructing the parallax depth map to obtain a mapping map;
and the measurement calculation unit is used for calculating the distance between the object and a preset camera according to the position information of the object through the mapping chart.
Compared with the prior art, the invention at least comprises the following beneficial effects:
(1) by adopting an artificial intelligence advanced technology and combining deep learning, the blind obstacle avoidance equipment becomes more intelligent and more convenient for a traditional physical obstacle avoidance system;
(2) the invention improves the existing confrontation network (GAN) according to the characteristic of the blind, and then performs autonomous learning according to the change of the environment, thereby achieving more humanized effect and leading the obstacle avoidance system to be more intelligent according to different daily behaviors of the blind;
(3) the invention can detect and identify a plurality of objects appearing in front when walking, and respectively broadcast the change of the distance between the plurality of objects and the positions of the plurality of objects, can be used for people with visual impairment to find routes indoors and outdoors, find objects indoors and the like, and helps more blind people open another door for knowing the real world.
Drawings
FIG. 1 is a general flow chart of a first embodiment of the present invention;
FIG. 2 is a flowchart of step S1 according to an embodiment of the present invention;
FIG. 3 is a flowchart illustrating steps of constructing a predetermined model according to a first embodiment of the present invention;
FIG. 4 is a flowchart of step S3 according to an embodiment of the present invention;
FIG. 5 is a block diagram of a second embodiment of the present invention.
Detailed Description
The following are specific embodiments of the present invention, and the technical solutions of the present invention will be further described with reference to the drawings, but the present invention is not limited to these embodiments.
Example one
As shown in FIG. 1, the invention relates to an intelligent obstacle avoidance method for blind people based on deep learning, which comprises the following steps:
s1, acquiring pictures in a preset range through a preset camera;
s2, identifying the object existing in the acquired picture according to a preset training model and obtaining the position information of the object;
s3, carrying out binocular distance measurement on the object target in the acquired picture according to a preset distance measurement algorithm;
s4, broadcasting the object position information and the measured object target distance through voice, judging the direction of the blind person avoiding the obstacle according to the object position information and the measured object target distance, and reminding the blind person to avoid the corresponding obstacle.
Before encountering an obstacle, the visual barrier detection device can help a person with visual barrier to make a decision on a walking route, namely help the person with visual barrier to avoid the corresponding obstacle by broadcasting the distance between the person and the obstacle, and after encountering the obstacle, the visual barrier detection device can judge the direction in which the person can avoid the obstacle according to the object position information and the measured object target distance, and remind the person with visual barrier of how to walk to avoid the corresponding obstacle.
As shown in fig. 2, step S1 specifically includes:
s11, acquiring a video within a preset range through a preset camera;
the video images of scenes in the peripheral range of the blind are acquired through the left camera and the right camera, so that the object recognition capability and the accuracy of position information can be further improved.
And S12, taking frames of the video to obtain pictures in a preset range.
As shown in fig. 3, the specific steps of establishing the preset model include:
s21, acquiring an object picture in a preset mode, and establishing a corresponding data set;
the method acquires the object picture in a live-action shooting mode or an on-line picture acquisition mode, and the object picture is used as a training sample required by model training and establishes a corresponding data set.
S22, acquiring preset parameters of the object in the picture in a preset mode;
the preset parameters of the object are the position of the object, the name of the object and the type of the object.
And drawing a square frame on an object existing in each picture in a manual labeling mode, writing an object name, providing parameters for model training, improving the recognition capability of the model, and enabling the model to recognize the type and position information of the object during recognition.
And S23, training the pictures through a preset network model, and continuously optimizing network parameters to obtain a preset training model.
The network model selected by the invention is yolov3 model, and is mainly responsible for carrying out object detection on the obtained picture, namely, the position of the object on the picture and the class predicted by the model are taken, the network model is trained through the training samples in the data set, and the network parameters are continuously optimized, so that the finally generated training model can accurately identify the object in the image and predict the type of the object.
Further, as shown in fig. 4, step S3 includes:
and S31, calibrating the preset camera.
The invention uses opencv to respectively intercept the images shot by the left camera and the right camera, during which the calibration plate needs to be held by hand and the angle is adjusted continuously to obtain the calibration data of the cameras, the data has a function similar to that of a scale, and the distance data on the images and the distance between an object and the cameras can be corresponded to obtain the internal and external parameters of the cameras.
And S32, dividing the collected pictures to obtain a left picture and a right picture.
And S33, performing coordinate mapping on each picture according to the data obtained by camera calibration.
And S34, graying the two mapped pictures.
S35, forming a parallax depth map by the two grayed pictures according to a preset function;
according to the invention, a depth map function is used to process the grayed left and right camera pictures to form a parallax depth map.
And S36, reconstructing the depth map to obtain a mapping map.
And S37, calculating the distance between the object and the preset camera according to the position information of the object through the mapping.
And extracting the position information of the object obtained before, calculating the central point of each object according to the position information, detecting the abscissa of the central point, and filtering the point of which the point deviates from the center of the image.
And screening the distances which obviously do not conform to the actual situation in the predicted distances to finally obtain the distance between the object target and the camera.
According to the method, through an artificial intelligence advanced technology and combined with deep learning, for a traditional physical obstacle avoidance system, the obstacle avoidance equipment for the blind becomes more intelligent and more convenient.
Example two
As shown in fig. 5, the blind intelligent obstacle avoidance system based on deep learning of the present invention includes an image acquisition module, an object detection module, a binocular distance measurement module and a voice conversion module.
The image acquisition module is used for through predetermineeing the camera collection and predetermineeing the picture in the scope, and wherein the image acquisition module includes:
the video acquisition unit is used for acquiring videos in a preset range through a preset camera;
and the frame taking unit is used for taking frames of the video to obtain a picture of each frame.
The object detection module is used for identifying an object existing in the acquired picture according to a preset training model and obtaining position information of the object;
the binocular distance measurement module is used for measuring the binocular distance of the object target in the acquired picture according to a preset distance measurement algorithm;
and the voice conversion module is used for broadcasting the object position information and the determined object target distance through voice, judging the direction of the blind person for avoiding the barrier according to the object position information and the determined object target distance, and reminding the blind person of avoiding the corresponding barrier.
Wherein, binocular range finding module includes:
the calibration unit is used for calibrating the pictures acquired by the preset camera;
the dividing unit is used for dividing the calibrated picture to obtain a left picture and a right picture;
the coordinate mapping unit is used for mapping the coordinates of each picture according to the data obtained by calibration;
the gray processing unit is used for graying the two mapped pictures;
the depth map construction unit is used for forming a parallax depth map by the two grayed pictures according to a preset function;
the reconstruction unit is used for reconstructing the parallax depth map to obtain a mapping map;
and the measurement calculation unit is used for calculating the distance between the object and a preset camera according to the position information of the object through the mapping chart.
The invention can detect and identify a plurality of objects appearing in front when walking, and respectively broadcast the change of the distance between the plurality of objects and the positions of the plurality of objects, can be used for people with visual impairment to find routes indoors/outdoors, find objects indoors and the like, and helps more people with visual impairment to open another door for knowing the real world.
The specific embodiments described herein are merely illustrative of the spirit of the invention. Various modifications or additions may be made to the described embodiments or alternatives may be employed by those skilled in the art without departing from the spirit or ambit of the invention as defined in the appended claims.
Claims (8)
1. An intelligent obstacle avoidance method for blind people based on deep learning is characterized by comprising the following steps:
acquiring pictures in a preset range through a preset camera;
identifying an object existing in the acquired picture according to a preset training model and obtaining position information of the object;
measuring the binocular distance of the object target in the acquired picture according to a preset distance measurement algorithm;
and broadcasting the object position information and the determined object target distance through voice, judging the direction of the blind person for avoiding the barrier according to the object position information and the determined object target distance, and reminding the blind person to avoid the corresponding barrier.
2. The blind intelligent obstacle avoidance method based on deep learning as claimed in claim 1, wherein the specific steps of collecting pictures in a preset range through a preset camera comprise:
acquiring a video within a preset range through a preset camera;
and taking frames of the video to obtain a picture in a preset range.
3. The blind intelligent obstacle avoidance method based on deep learning as claimed in claim 1, wherein the specific steps of establishing the preset model comprise:
acquiring an object picture in a preset mode, and establishing a corresponding data set;
acquiring preset parameters of an object in a picture in a preset mode;
and training the picture through a preset network model, and continuously optimizing network parameters to obtain a preset training model.
4. The blind intelligent obstacle avoidance method based on deep learning as claimed in claim 3, wherein the preset parameters of the object are the position of the object, the name of the object and the type of the object.
5. The blind intelligent obstacle avoidance method based on deep learning as claimed in claim 1, wherein the specific steps of performing distance measurement on the object target in the acquired picture according to a preset distance measurement algorithm comprise:
calibrating a preset camera;
dividing the collected pictures to obtain a left picture and a right picture;
performing coordinate mapping on each picture according to data obtained by camera calibration;
graying the two mapped pictures;
forming a parallax depth map from the two grayed pictures according to a preset function;
reconstructing the depth map to obtain a mapping map;
and calculating the distance between the object and a preset camera through the mapping according to the position information of the object.
6. The utility model provides a blind person intelligence keeps away barrier system based on degree of depth study which characterized in that includes:
the image acquisition module is used for acquiring pictures in a preset range through a preset camera;
the object detection module is used for identifying an object existing in the acquired picture according to a preset training model and obtaining position information of the object;
the binocular distance measurement module is used for measuring the binocular distance of the object target in the acquired picture according to a preset distance measurement algorithm;
and the voice conversion module is used for broadcasting the object position information and the determined object target distance through voice, judging the direction of the blind person for avoiding the barrier according to the object position information and the determined object target distance, and reminding the blind person of avoiding the corresponding barrier.
7. The blind intelligent obstacle avoidance system based on deep learning as claimed in claim 6, wherein the image acquisition module comprises:
the video acquisition unit is used for acquiring videos in a preset range through a preset camera;
and the frame taking unit is used for taking frames of the video to obtain a picture of each frame.
8. The blind intelligent obstacle avoidance system based on deep learning of claim 6, wherein the binocular ranging module comprises:
the calibration unit is used for calibrating a preset camera;
the dividing unit is used for dividing the collected pictures to obtain a left picture and a right picture;
the coordinate mapping unit is used for mapping the coordinates of each picture according to the data obtained by camera calibration;
the gray processing unit is used for graying the two mapped pictures;
the depth map construction unit is used for forming a parallax depth map by the two grayed pictures according to a preset function;
the reconstruction unit is used for reconstructing the parallax depth map to obtain a mapping map;
and the measurement calculation unit is used for calculating the distance between the object and a preset camera according to the position information of the object through the mapping chart.
Applications Claiming Priority (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2021102800124 | 2021-03-16 | ||
| CN202110280012 | 2021-03-16 |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN113208882A true CN113208882A (en) | 2021-08-06 |
Family
ID=77089552
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110465837.3A Pending CN113208882A (en) | 2021-03-16 | 2021-04-28 | Blind person intelligent obstacle avoidance method and system based on deep learning |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113208882A (en) |
Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105957116A (en) * | 2016-05-03 | 2016-09-21 | 大连理工大学 | Dynamic coding point designing and decoding method based on frequency |
| CN106389078A (en) * | 2016-11-24 | 2017-02-15 | 贵州大学 | Intelligent blind guiding glass system and blind guiding method thereof |
| CN106859929A (en) * | 2017-01-25 | 2017-06-20 | 上海集成电路研发中心有限公司 | A kind of Multifunctional blind person guiding instrument based on binocular vision |
| CN109106563A (en) * | 2018-06-28 | 2019-01-01 | 清华大学天津高端装备研究院 | A kind of automation blind-guide device based on deep learning algorithm |
| CN110109457A (en) * | 2019-04-29 | 2019-08-09 | 北方民族大学 | A kind of intelligent sound blind-guidance robot control method and control system |
| CN110132243A (en) * | 2019-05-31 | 2019-08-16 | 南昌航空大学 | A kind of modularization positioning system based on deep learning and ranging |
| CN111239684A (en) * | 2020-01-17 | 2020-06-05 | 中航华东光电（上海）有限公司 | Binocular fast distance measurement method based on YoloV3 deep learning |
| CN111260715A (en) * | 2020-01-20 | 2020-06-09 | 深圳市普渡科技有限公司 | Depth map processing method, small obstacle detection method and system |
| CN111815712A (en) * | 2020-06-24 | 2020-10-23 | 中国地质大学（武汉） | High-precision camera-single laser combined calibration method |
| CN112161572A (en) * | 2020-10-09 | 2021-01-01 | 上海飒智智能科技有限公司 | Object three-dimensional size measuring system based on fusion depth camera |
-
2021
- 2021-04-28 CN CN202110465837.3A patent/CN113208882A/en active Pending
Patent Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105957116A (en) * | 2016-05-03 | 2016-09-21 | 大连理工大学 | Dynamic coding point designing and decoding method based on frequency |
| CN106389078A (en) * | 2016-11-24 | 2017-02-15 | 贵州大学 | Intelligent blind guiding glass system and blind guiding method thereof |
| CN106859929A (en) * | 2017-01-25 | 2017-06-20 | 上海集成电路研发中心有限公司 | A kind of Multifunctional blind person guiding instrument based on binocular vision |
| CN109106563A (en) * | 2018-06-28 | 2019-01-01 | 清华大学天津高端装备研究院 | A kind of automation blind-guide device based on deep learning algorithm |
| CN110109457A (en) * | 2019-04-29 | 2019-08-09 | 北方民族大学 | A kind of intelligent sound blind-guidance robot control method and control system |
| CN110132243A (en) * | 2019-05-31 | 2019-08-16 | 南昌航空大学 | A kind of modularization positioning system based on deep learning and ranging |
| CN111239684A (en) * | 2020-01-17 | 2020-06-05 | 中航华东光电（上海）有限公司 | Binocular fast distance measurement method based on YoloV3 deep learning |
| CN111260715A (en) * | 2020-01-20 | 2020-06-09 | 深圳市普渡科技有限公司 | Depth map processing method, small obstacle detection method and system |
| CN111815712A (en) * | 2020-06-24 | 2020-10-23 | 中国地质大学（武汉） | High-precision camera-single laser combined calibration method |
| CN112161572A (en) * | 2020-10-09 | 2021-01-01 | 上海飒智智能科技有限公司 | Object three-dimensional size measuring system based on fusion depth camera |
Non-Patent Citations (1)
| Title |
|---|
| 李云鹏,侯凌燕,王超: "基于YOLOv3的自动驾驶中运动目标检测", 《计算机工程与设计》 * |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111209810B (en) | Boundary frame segmentation supervision deep neural network architecture for accurately detecting pedestrians in real time through visible light and infrared images | |
| CN108898085B (en) | Intelligent road disease detection method based on mobile phone video | |
| CN111563887B (en) | Intelligent analysis method and device for oral cavity image | |
| CN102389361B (en) | Blindman outdoor support system based on computer vision | |
| CN105654732A (en) | Road monitoring system and method based on depth image | |
| CN106886216A (en) | Robot automatic tracking method and system based on RGBD Face datections | |
| CN104408746B (en) | A kind of passenger flow statistical system based on depth information | |
| CN105225482A (en) | Based on vehicle detecting system and the method for binocular stereo vision | |
| CN110232389A (en) | A kind of stereoscopic vision air navigation aid based on green crop feature extraction invariance | |
| CN102609724B (en) | Method for prompting ambient environment information by using two cameras | |
| CN105243386A (en) | Face living judgment method and system | |
| CN109145708A (en) | A kind of people flow rate statistical method based on the fusion of RGB and D information | |
| CN113435336B (en) | Running intelligent timing system and method based on artificial intelligence | |
| CN105760846A (en) | Object detection and location method and system based on depth data | |
| CN109035307B (en) | Set area target tracking method and system based on natural light binocular vision | |
| CN113762009B (en) | Crowd counting method based on multi-scale feature fusion and double-attention mechanism | |
| CN106483130B (en) | A kind of detection method and its automatic detection device of rice disease | |
| CN105913013A (en) | Binocular vision face recognition algorithm | |
| CN104517095A (en) | Head division method based on depth image | |
| CN107241572A (en) | Student's real training video frequency tracking evaluation system | |
| CN112560716A (en) | High-resolution remote sensing image water body extraction method based on low-level feature fusion | |
| CN103481842B (en) | A kind of changing method of moving vehicles detection and tracking pattern | |
| CN110276276A (en) | The determination method and system of examinee's face direction of visual lines in a kind of Driving Test | |
| CN111444837B (en) | Temperature measurement method and temperature measurement system for improving face detection usability in extreme environment | |
| CN106709432A (en) | Binocular stereoscopic vision based head detecting and counting method |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |