CN108985343B - Automobile damage detection method and system based on deep neural network - Google Patents
Automobile damage detection method and system based on deep neural network Download PDFInfo
- Publication number
- CN108985343B CN108985343B CN201810653240.XA CN201810653240A CN108985343B CN 108985343 B CN108985343 B CN 108985343B CN 201810653240 A CN201810653240 A CN 201810653240A CN 108985343 B CN108985343 B CN 108985343B
- Authority
- CN
- China
- Prior art keywords
- automobile
- damage detection
- neural network
- damage
- photo
- 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.)
- Active
Links
- 230000006378 damage Effects 0.000 title claims abstract description 147
- 238000001514 detection method Methods 0.000 title claims abstract description 111
- 238000013528 artificial neural network Methods 0.000 title claims abstract description 36
- 238000013527 convolutional neural network Methods 0.000 claims abstract description 28
- 238000000034 method Methods 0.000 claims abstract description 28
- 238000011156 evaluation Methods 0.000 claims abstract description 16
- 230000004927 fusion Effects 0.000 claims abstract description 13
- 230000011218 segmentation Effects 0.000 claims description 23
- 238000000605 extraction Methods 0.000 claims description 10
- 240000004282 Grewia occidentalis Species 0.000 claims description 7
- 208000027418 Wounds and injury Diseases 0.000 claims description 6
- 208000014674 injury Diseases 0.000 claims description 6
- 238000010586 diagram Methods 0.000 description 4
- 238000013135 deep learning Methods 0.000 description 3
- 238000005516 engineering process Methods 0.000 description 2
- 230000003902 lesion Effects 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000012549 training Methods 0.000 description 2
- 230000006978 adaptation Effects 0.000 description 1
- 238000013459 approach Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 239000003086 colorant Substances 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 230000006735 deficit Effects 0.000 description 1
- 238000003709 image segmentation Methods 0.000 description 1
- 238000007689 inspection Methods 0.000 description 1
- 238000011835 investigation Methods 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
- 238000012552 review Methods 0.000 description 1
- 238000005070 sampling Methods 0.000 description 1
- 230000000007 visual effect Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/40—Extraction of image or video features
Abstract
The invention discloses an automobile damage detection method and system based on a deep neural network, wherein the automobile damage detection method comprises the following steps: acquiring an automobile damage detection photo at a specific angle; dividing appearance parts of the automobile damage detection photo, and determining the position and the type of each appearance part; carrying out damage detection on each appearance component by utilizing a deep convolutional neural network; and performing fusion evaluation on the damage detection results of all the appearance parts, and outputting the damaged parts and the confidence of the automobile. According to the technical scheme, the method and the device can obtain extremely high accuracy and recall rate, are convenient and simple to use, have no special requirements on front-end equipment, and improve the popularization and the efficiency of automobile damage detection.
Description
Technical Field
The invention relates to the technical field of vehicles, in particular to an automobile damage detection method based on a deep neural network and an automobile damage detection system based on the deep neural network.
Background
The automobile appearance damage investigation is common in automobile business, such as automobile insurance underwriting, time-sharing leasing and automobile taking and returning links of automobile daily leasing, and whether the automobile appearance is damaged or not needs to be confirmed. At present, two methods are generally adopted, one is that a worker surveys in the field and fills in a report, and the other is that a user takes a picture (video) and submits the picture to a server for processing.
The existing system has the following problems:
1. the method for submitting the manual review of the server by photographing by the user has the problems of large time delay, high cost and poor user experience, and cannot be expanded and used on a large scale.
2. The damage assessment technology is high in shooting requirement, multiple groups of photos need to be shot for each part, and the damage assessment technology is mainly applied to scenes of insurance claim settlement and cannot be applied to vehicle inspection and rental business.
Since 2012, deep learning has made great progress in the field of picture recognition. Compared with the traditional picture identification method, the method has the advantages that low-level visual characteristics such as colors, HOG and the like are used; deep neural networks can learn more advanced, abstract features, which makes the performance of deep neural networks far superior to traditional approaches. Since 2014, deep learning has begun to achieve excellent results in the fields of object detection, object segmentation and the like, and a series of methods such as deep lab, YOLO, fast RCNN and the like are developed, so that the recognition accuracy rate exceeds the level of human recognition on a specific task, and the method is widely used in a generation environment. The work of deep learning in the field of automobile appearance damage detection is relatively little, and about 2016, some attempts are made. However, due to the difficulty of data acquisition, the field is slow to develop, and no system or method which can be used on the ground exists at present.
In the prior art, an image-based vehicle damage assessment method is provided, wherein insurance claim settlement vehicle pictures are used as training identification data, and the data are harsh in adaptation to scenes; the method can only be used for appearance damage assessment and damage determination of the vehicles in danger, and cannot be popularized to the scenes of underwriting and renting which mainly comprise the vehicles without damage. The method is also suitable for confirming the damage degree and acquiring information, and cannot be popularized to the application scenes of underwriting and leasing mainly without damage vehicles.
Disclosure of Invention
In order to solve at least one of the problems, the invention provides a deep neural network-based automobile damage detection method and a deep neural network-based automobile damage detection system, wherein the damage detection is performed on the appearance parts in the automobile damage detection picture based on the deep neural network, and the damage detection structures of all the appearance parts are subjected to fusion evaluation to determine the damage condition of the automobile. The method based on the deep neural network can obtain extremely high accuracy and recall rate, each vehicle only needs to collect 4 or 6 pictures at a specific angle, the use is convenient and simple, no special requirement is required on front-end equipment, a mainstream smart phone can be directly used, and the popularization of the C end is facilitated.
In order to achieve the above object, the present invention provides an automobile damage detection method based on a deep neural network, comprising: acquiring an automobile damage detection photo at a specific angle; performing appearance part segmentation on the automobile damage detection photo, and determining the position and the type of each appearance part; carrying out damage detection on each appearance part by utilizing a deep convolutional neural network; and performing fusion evaluation on the damage detection results of all the appearance parts, and outputting the damaged parts and the confidence of the automobile.
In the above technical solution, preferably, the car damage detection photo is a four-corner photo or a hexagonal photo, and the four-corner photo is a photo of four angles of left front, left rear, right front and right rear relative to the car; the hexagonal photos are photos of any two angles of the left front angle, the left rear angle, the right front angle, the right rear angle and the front angle, the rear angle, the left angle and the right angle relative to the automobile.
In the foregoing technical solution, preferably, the dividing of the exterior parts of the automobile damage detection photograph, and the determining of the position and the type of each exterior part specifically include: and performing appearance part segmentation on the automobile damage detection photo by using a segmentation algorithm, and determining the position of the segmented appearance part in the automobile and the type of the appearance part.
In the foregoing technical solution, preferably, the performing damage detection on each appearance component by using a deep convolutional neural network specifically includes: cutting the appearance parts according to the position and the type of each cut appearance part; extracting abstract features of each appearance component by using a deep convolutional neural network; selecting a candidate damage area by using the area candidate network according to the extracted abstract characteristics; and returning a precise region of the damage from the candidate damage region.
In the above technical solution, preferably, the method for detecting automobile damage based on the deep neural network further includes: extracting characteristics of the automobile damage detection photo to obtain the vehicle posture and the vehicle type information of the automobile; and determining a corresponding deep convolutional neural network according to the vehicle posture and the vehicle type information of the vehicle, so as to perform damage detection on the damaged part by using the deep convolutional neural network.
In the foregoing technical solution, preferably, the performing feature extraction on the car damage detection photo and acquiring the car posture and car type information of the car specifically includes: extracting the characteristics of the automobile damage detection photo through a deep convolutional neural network; connecting a full connection layer network to classify the automobile damage detection photos and determine the automobile posture of the automobile; will the characteristic of car damage detection photo matches with the motorcycle type characteristic in predetermineeing the motorcycle type storehouse, confirms the motorcycle type information of car, wherein, when car damage detection photo was the four corners photo, full connection layer network carries out four classifications, when car damage detection photo was the hexagonal photo, full connection layer network carries out six classifications.
In the above technical solution, preferably, the process of performing fusion evaluation on the damage detection results of all the appearance parts and outputting the damaged parts and the confidence level of the automobile specifically includes: constructing a spatial relationship corresponding to the damage accurate area according to the automobile type information and the automobile posture of the automobile; and determining a damaged part of the automobile, and determining the confidence of the damaged part according to a confidence algorithm.
The invention also provides an automobile damage detection system based on the deep neural network, which comprises the following components: the photo acquisition module is used for acquiring an automobile damage detection photo at a specific angle; the photo segmentation module is used for segmenting appearance components of the automobile damage detection photo and determining the position and the type of each appearance component; the damage detection module is used for carrying out damage detection on each appearance component by utilizing a deep convolutional neural network; and the damage evaluation module is used for performing fusion evaluation on the damage detection results of all the appearance parts and outputting the damaged parts and the confidence coefficient of the automobile.
In the above technical solution, preferably, the car damage detection system based on the deep neural network further includes: the characteristic extraction module is used for extracting the characteristics of the automobile damage detection photo and acquiring the vehicle posture and the vehicle type information of the automobile; and the damage detection module determines a corresponding deep convolutional neural network according to the vehicle posture and the vehicle type information of the vehicle, which are acquired by the characteristic extraction module, so that the damage detection is carried out on the damaged part by utilizing the deep convolutional neural network.
Compared with the prior art, the invention has the beneficial effects that: and carrying out damage detection on the appearance parts in the automobile damage detection photo based on the deep neural network, and carrying out fusion evaluation on the damage detection structures of all the appearance parts to determine the damage condition of the automobile. The method based on the deep neural network can obtain extremely high accuracy and recall rate, each vehicle only needs to collect 4 or 6 pictures at a specific angle, the use is convenient and simple, no special requirement is required on front-end equipment, a mainstream smart phone can be directly used, and the popularization of the C end is facilitated.
Drawings
FIG. 1 is a schematic flowchart of a method for detecting vehicle damage based on a deep neural network according to an embodiment of the present invention;
FIG. 2 is a schematic flow chart of a method for detecting damage to a vehicle based on a deep neural network according to another embodiment of the present invention;
FIG. 3 is a schematic diagram illustrating a segmentation principle of a segmentation algorithm according to an embodiment of the present invention;
FIG. 4 is a schematic diagram illustrating a principle of performing an injury detection based on a deep neural network according to another embodiment of the present invention;
fig. 5 is a schematic block diagram of a deep neural network-based automobile injury detection system according to an embodiment of the present invention.
In the drawings, the correspondence between each component and the reference numeral is:
30. the automobile damage detection system based on the deep neural network comprises a 31 image acquisition module, a 32 image segmentation module, a 33 feature extraction module, a 34 damage detection module and a 35 damage assessment module.
Detailed Description
In order to make the objects, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are some, but not all, embodiments of the present invention. All other embodiments, which can be obtained by a person skilled in the art without any inventive step based on the embodiments of the present invention, are within the scope of the present invention.
The invention is described in further detail below with reference to the attached drawing figures:
as shown in fig. 1, the method for detecting the damage of the automobile based on the deep neural network provided by the invention comprises the following steps: step S11, obtaining an automobile damage detection photo at a specific angle; step S12, dividing the automobile damage detection photo into appearance parts, and determining the position and type of each appearance part; step S13, carrying out damage detection on each appearance component by using a deep convolutional neural network; and step S14, performing fusion evaluation on the damage detection results of all the appearance parts, and outputting the damaged parts and the confidence of the automobile.
In the above embodiment, preferably, the car damage detection photo is a four-corner photo or a hexagonal photo, and the four-corner photo is a photo of four angles with respect to the front left, the rear left, the front right and the rear right of the car; the hexagonal photograph is a photograph of four angles of left front, left rear, right front and right rear and any two angles of the four angles of front, rear, left and right with respect to the automobile.
In the above embodiment, preferably, the exterior part segmentation is performed on the automobile damage detection photo, and the position and the type of each exterior part are determined as follows: and performing appearance part segmentation on the automobile damage detection photo by using a segmentation algorithm, and determining the position of the segmented appearance part in the automobile and the type of the appearance part.
In the foregoing embodiment, preferably, the performing damage detection on each appearance component by using the deep convolutional neural network specifically includes: segmenting the appearance components according to the position and the type of each segmented appearance component; extracting abstract features of each appearance component by using a deep convolutional neural network; selecting a candidate damage area by using the area candidate network according to the extracted abstract characteristics; and returning the accurate region of the damage from the candidate damage region.
As shown in fig. 2, in the above embodiment, preferably, the method for detecting automobile damage based on a deep neural network further includes: extracting the characteristics of the automobile damage detection photo to obtain the vehicle posture and the vehicle type information of the automobile; and determining a corresponding deep convolutional neural network according to the vehicle posture and the vehicle type information of the vehicle so as to detect the damage of the damaged part by using the deep convolutional neural network.
In the foregoing embodiment, preferably, the performing feature extraction on the automobile damage detection photo and acquiring the vehicle posture and the vehicle type information of the automobile specifically includes: extracting the characteristics of the automobile damage detection photo through a deep convolutional neural network; connecting a full-connection layer network to classify the automobile damage detection photos and determine the automobile posture of the automobile; the method comprises the steps of matching the characteristics of an automobile damage detection photo with the characteristics of automobile types in a preset automobile type library and determining automobile type information of an automobile, wherein when the automobile damage detection photo is a four-corner photo, a full-connection layer network carries out four classifications, and when the automobile damage detection photo is a hexagonal photo, the full-connection layer network carries out six classifications.
In the above embodiment, preferably, the process of performing fusion evaluation on the damage detection results of all the appearance parts and outputting the damaged parts and the confidence degrees of the automobile specifically includes: constructing a spatial relation corresponding to the accurate damage area according to the vehicle type information and the vehicle posture of the vehicle; and determining a damaged part of the automobile, and determining the confidence of the damaged part according to a confidence algorithm.
Based on the description of the above embodiments, the following specific implementation methods are used as examples to further explain the method for detecting damage to an automobile provided by the present invention:
the implementation steps are as follows:
1. from the 4/6 angle photograph of the car obtained by photographing, the attitude V of the car is obtainediAnd vehicle type information Ti. In particular, the vehicle pose may use any general segmentation algorithm, such as: VGG, ResNet, GoogleNet, inclusion V3, NASNET, and the like. Vehicle attitude can be subdivided into 4 dimensions: left front, right front, left back, right back; optionally, one of the pairs of front, back, left and right may be added for the 6-dimension. After the picture features are extracted through the deep convolutional network, the full connected layers (FC) network is connected for 4(6) classification to obtain the posture Ti。
2. Performing appearance part segmentation on the automobile photo by using a deep convolutional network, and determining the position and the type of each part, which are recorded as Pij[n,l](jth part of ith diagram, part name n, location l). In particular, the vehicle component segmentation may use any general segmentation algorithm, such as: deeplab, PSPNet, DIS, IDW-CNN and the like, and after training, the vehicle picture is subjected to appearance component segmentation to obtain Pij. As shown in the figure3, the principle is illustrated by taking Deeplab as an example:
1) a deep convolution neural network, such as VGG-16 or ResNet-101, adopts a full convolution mode to reduce the degree of signal down-sampling (from 32x to 8x) by using porous convolution;
2) in a bilinear interpolation stage, increasing the resolution of the feature map to the original image;
3) and optimizing a segmentation result by using a conditional random field, and better grabbing the edge of the object to realize segmentation.
3. For each PijAccording to the vehicle attitude ViAnd vehicle type information TiUsing a uniform or corresponding deep convolutional network for impairment detection Dij. In particular, the lesion detection uses any general target detection algorithm, such as: faster RCNN, SSD, YOLO, etc. As shown in FIG. 4, the principle is illustrated below by taking the fast RCNN as an example:
1) according to the division result PijDividing the vehicle picture, each PijAs a detection input;
2) deep convolutional network (conv layers) pair PijExtracting feature picture abstract features (feature maps);
3) recommending a candidate damage area by using the area candidate network;
4) regression of the lesion from the candidate region to the precision region Dij。
4. According to the vehicle model information TiAnd attitude information V of vehicle pictureiConstruction of DijCorresponding spatial relationship to all DijAnd performing fusion evaluation, and outputting the part with the damage and the confidence coefficient.
As shown in fig. 5, the present invention further provides an automobile injury detection system based on a deep neural network, including: the photo acquisition module 31 is used for acquiring an automobile damage detection photo at a specific angle; the photo segmentation module 32 is used for segmenting appearance components of the automobile damage detection photo and determining the position and the type of each appearance component; a damage detection module 34, configured to perform damage detection on each appearance component by using a deep convolutional neural network; and the damage evaluation module 35 is configured to perform fusion evaluation on the damage detection results of all the appearance parts, and output the damaged parts and the confidence of the automobile.
In the above embodiment, preferably, the car damage detection system based on the deep neural network further includes: the characteristic extraction module 33 is used for extracting the characteristics of the automobile damage detection photo to obtain the vehicle posture and the vehicle type information of the automobile; the damage detection module 34 determines a corresponding deep convolutional neural network according to the vehicle posture and the vehicle type information of the vehicle acquired by the feature extraction module 33, so as to perform damage detection on the damaged component by using the deep convolutional neural network.
The foregoing is an embodiment of the present invention, and according to the method and system for detecting automobile damage based on the deep neural network provided by the present invention, the damage detection is performed on the appearance parts in the automobile damage detection photo based on the deep neural network, and the damage detection structures of all the appearance parts are subjected to fusion evaluation to determine the damage condition of the automobile. The method based on the deep neural network can obtain extremely high accuracy and recall rate, each vehicle only needs to collect 4 or 6 pictures at a specific angle, the use is convenient and simple, no special requirement is required on front-end equipment, a mainstream smart phone can be directly used, and the popularization of the C end is facilitated.
The above is only a preferred embodiment of the present invention, and is not intended to limit the present invention, and various modifications and changes will occur to those skilled in the art. Any modification, equivalent replacement, or improvement made within the spirit and principle of the present invention should be included in the protection scope of the present invention.
Claims (5)
1. A car damage detection method based on a deep neural network is characterized by comprising the following steps:
acquiring automobile damage detection photos at specific angles, wherein the automobile damage detection photos are four-corner photos or hexagonal photos, and the four-corner photos are photos at four angles of left front, left rear, right front and right rear relative to the automobile; the hexagonal photos are photos of four angles of left front, left back, right front and right back and any two angles of the four angles of front, back, left and right relative to the automobile;
performing appearance part segmentation on the automobile damage detection photo, and determining the position and the type of each appearance part;
carrying out feature extraction on the automobile damage detection photo to acquire the vehicle posture and the vehicle type information of the automobile, and specifically comprising the following steps of:
extracting the characteristics of the automobile damage detection photo through a deep convolutional neural network;
connecting a full connection layer network to classify the automobile damage detection photos and determine the automobile posture of the automobile;
matching the characteristics of the automobile damage detection photo with the automobile type characteristics of a preset automobile type library to determine the automobile type information of the automobile;
when the automobile damage detection photo is a quadrangular photo, the full connection layer network carries out four classifications, and when the automobile damage detection photo is a hexagonal photo, the full connection layer network carries out six classifications;
determining a corresponding deep convolutional neural network according to the vehicle posture and the vehicle type information of the vehicle;
carrying out damage detection on each appearance part by utilizing the deep convolutional neural network;
performing fusion evaluation on the damage detection results of all the appearance parts, and outputting the damaged parts and the confidence coefficient of the automobile, wherein the method specifically comprises the following steps:
constructing a spatial relationship corresponding to the damage accurate area according to the automobile type information and the automobile posture of the automobile;
and determining a damaged part of the automobile, and determining the confidence of the damaged part according to a confidence algorithm.
2. The method for detecting automobile damage based on the deep neural network of claim 1, wherein the step of performing appearance part segmentation on the automobile damage detection photo and determining the position and type of each appearance part specifically comprises the steps of: and performing appearance part segmentation on the automobile damage detection photo by using a segmentation algorithm, and determining the position of the segmented appearance part in the automobile and the type of the appearance part.
3. The method for detecting the damage of the automobile based on the deep neural network as claimed in claim 1, wherein the step of detecting the damage of each appearance part by using the deep convolutional neural network specifically comprises the following steps:
cutting the appearance parts according to the position and the type of each cut appearance part;
extracting abstract features of each appearance component by using a deep convolutional neural network;
selecting a candidate damage area by using the area candidate network according to the extracted abstract characteristics;
and returning a precise region of the damage from the candidate damage region.
4. A deep neural network-based automobile injury detection system is characterized in that the deep neural network-based automobile injury detection method according to any one of claims 1 to 3 is applied, and comprises the following steps:
the photo acquisition module is used for acquiring an automobile damage detection photo at a specific angle;
the photo segmentation module is used for segmenting appearance components of the automobile damage detection photo and determining the position and the type of each appearance component;
the damage detection module is used for carrying out damage detection on each appearance component by utilizing a deep convolutional neural network;
and the damage evaluation module is used for performing fusion evaluation on the damage detection results of all the appearance parts and outputting the damaged parts and the confidence coefficient of the automobile.
5. The deep neural network-based automotive injury detection system of claim 4, further comprising:
the characteristic extraction module is used for extracting the characteristics of the automobile damage detection photo and acquiring the vehicle posture and the vehicle type information of the automobile;
and the damage detection module determines a corresponding deep convolutional neural network according to the vehicle posture and the vehicle type information of the vehicle, which are acquired by the characteristic extraction module, so that the damage detection is carried out on the damaged part by utilizing the deep convolutional neural network.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810653240.XA CN108985343B (en) | 2018-06-22 | 2018-06-22 | Automobile damage detection method and system based on deep neural network |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810653240.XA CN108985343B (en) | 2018-06-22 | 2018-06-22 | Automobile damage detection method and system based on deep neural network |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN108985343A CN108985343A (en) | 2018-12-11 |
| CN108985343B true CN108985343B (en) | 2020-12-25 |
Family
ID=64538340
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810653240.XA Active CN108985343B (en) | 2018-06-22 | 2018-06-22 | Automobile damage detection method and system based on deep neural network |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108985343B (en) |
Families Citing this family (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109684956A (en) * | 2018-12-14 | 2019-04-26 | 深源恒际科技有限公司 | A kind of vehicle damage detection method and system based on deep neural network |
| CN109697429A (en) * | 2018-12-27 | 2019-04-30 | 睿驰达新能源汽车科技（北京）有限公司 | A kind of method and device of determining vehicle damage |
| CN110569865B (en) * | 2019-02-01 | 2020-07-17 | 阿里巴巴集团控股有限公司 | Method and device for recognizing vehicle body direction |
| DE102019204346A1 (en) * | 2019-03-28 | 2020-10-01 | Volkswagen Aktiengesellschaft | Method and system for checking a visual complaint on a motor vehicle |
| CN110084806A (en) * | 2019-05-06 | 2019-08-02 | 深源恒际科技有限公司 | A kind of interface alternation method and device |
| CN110135437B (en) | 2019-05-06 | 2022-04-05 | 北京百度网讯科技有限公司 | Loss assessment method and device for vehicle, electronic equipment and computer storage medium |
| CN110781770B (en) * | 2019-10-08 | 2022-05-06 | 高新兴科技集团股份有限公司 | Living body detection method, device and equipment based on face recognition |
| CN110895814B (en) * | 2019-11-30 | 2023-04-18 | 南京工业大学 | Aero-engine hole-finding image damage segmentation method based on context coding network |
| CN111259969A (en) * | 2020-01-19 | 2020-06-09 | 上海钧正网络科技有限公司 | Failure reporting identification method, device, server and medium |
| CN111798452A (en) * | 2020-07-06 | 2020-10-20 | 北京小白世纪网络科技有限公司 | Carotid artery handheld ultrasonic image segmentation method, system and device |
| CN112270370B (en) * | 2020-11-06 | 2023-06-02 | 北京环境特性研究所 | Vehicle apparent damage assessment method |
| CN113538293B (en) * | 2021-08-20 | 2022-09-13 | 爱保科技有限公司 | Method and device for enhancing vehicle damage image |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107392218A (en) * | 2017-04-11 | 2017-11-24 | 阿里巴巴集团控股有限公司 | A kind of car damage identification method based on image, device and electronic equipment |
| CN107403424A (en) * | 2017-04-11 | 2017-11-28 | 阿里巴巴集团控股有限公司 | A kind of car damage identification method based on image, device and electronic equipment |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| KR101735874B1 (en) * | 2013-10-21 | 2017-05-15 | 한국전자통신연구원 | Apparatus and method for detecting vehicle number plate |
-
2018
- 2018-06-22 CN CN201810653240.XA patent/CN108985343B/en active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107392218A (en) * | 2017-04-11 | 2017-11-24 | 阿里巴巴集团控股有限公司 | A kind of car damage identification method based on image, device and electronic equipment |
| CN107403424A (en) * | 2017-04-11 | 2017-11-28 | 阿里巴巴集团控股有限公司 | A kind of car damage identification method based on image, device and electronic equipment |
Also Published As
| Publication number | Publication date |
|---|---|
| CN108985343A (en) | 2018-12-11 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108985343B (en) | Automobile damage detection method and system based on deep neural network | |
| CN108921068B (en) | Automobile appearance automatic damage assessment method and system based on deep neural network | |
| CN110988912B (en) | Road target and distance detection method, system and device for automatic driving vehicle | |
| Uittenbogaard et al. | Privacy protection in street-view panoramas using depth and multi-view imagery | |
| CN110245678B (en) | Image matching method based on heterogeneous twin region selection network | |
| CN111104867B (en) | Recognition model training and vehicle re-recognition method and device based on part segmentation | |
| CN108846333B (en) | Method for generating landmark data set of signpost and positioning vehicle | |
| CN111461170A (en) | Vehicle image detection method and device, computer equipment and storage medium | |
| CN111928842B (en) | Monocular vision based SLAM positioning method and related device | |
| CN111027481A (en) | Behavior analysis method and device based on human body key point detection | |
| CN111143489B (en) | Image-based positioning method and device, computer equipment and readable storage medium | |
| CN111837158A (en) | Image processing method and device, shooting device and movable platform | |
| CN110826415A (en) | Method and device for re-identifying vehicles in scene image | |
| CN109784171A (en) | Car damage identification method for screening images, device, readable storage medium storing program for executing and server | |
| CN111928857B (en) | Method and related device for realizing SLAM positioning in dynamic environment | |
| CN110120013A (en) | A kind of cloud method and device | |
| CN112598743B (en) | Pose estimation method and related device for monocular vision image | |
| CN108447084B (en) | Stereo matching compensation method based on ORB characteristics | |
| CN113793251A (en) | Pose determination method and device, electronic equipment and readable storage medium | |
| CN113033517B (en) | Vehicle damage assessment image acquisition method and device and storage medium | |
| CN112200850B (en) | ORB extraction method based on mature characteristic points | |
| Gong et al. | Joint view-identity manifold for target tracking and recognition | |
| CN112288817A (en) | Three-dimensional reconstruction processing method and device based on image | |
| Rasyidy et al. | A Framework for Road Boundary Detection based on Camera-LIDAR Fusion in World Coordinate System and Its Performance Evaluation Using Carla Simulator | |
| CN115049731B (en) | Visual image construction and positioning method based on binocular camera |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| CP03 | Change of name, title or address | ||
| CP03 | Change of name, title or address |
Address after: Room 203, Floor 2, Building 6, Qinghe Xisanqi East Road, Haidian District, Beijing 100,089 Patentee after: Beijing Shenzhi Hengji Technology Co.,Ltd. Address before: 0706-003, 113 Zhichun Road, Haidian District, Beijing 100086 Patentee before: SHENYUAN HENGJI TECHNOLOGY CO.,LTD. |