CN106548474A - A kind of micro-structure surface detection method - Google Patents
A kind of micro-structure surface detection method Download PDFInfo
- Publication number
- CN106548474A CN106548474A CN201610980599.9A CN201610980599A CN106548474A CN 106548474 A CN106548474 A CN 106548474A CN 201610980599 A CN201610980599 A CN 201610980599A CN 106548474 A CN106548474 A CN 106548474A
- Authority
- CN
- China
- Prior art keywords
- matching
- micro
- measured
- value
- gray
- 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
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/0002—Inspection of images, e.g. flaw detection
- G06T7/0004—Industrial image inspection
- G06T7/0006—Industrial image inspection using a design-rule based approach
Landscapes
- Engineering & Computer Science (AREA)
- Quality & Reliability (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Image Analysis (AREA)
Abstract
A kind of micro-structure surface detection method, it is characterised in that comprise the following steps：S1：Image acquisition；S2：Matching order S3：Gray-scale Matching；S4：Matching step-length is set；S5：Sensitizing range is arranged；S6：Labelling and output.Can quickly be detected for micro-structure surface geometric feature sizes, spacing and surface defect by this method.
Description
Technical field
The present invention relates to a kind of micro-structure surface detection method.
Technical background
Micro-structure surface has the irregular geometric properties of various complexity, due on machine components shape face, different
The micro structure of functional characteristic is different to the demand of geometric properties.The three-dimensional appearance of micro structure is complicated, with miniature scale feature, institute
It is bigger than the measurement difficulty of conventional ultraprecise part with the measurement of micro structure.Traditional small metering system, such as contact method
Measurement, according to morphology sampling theory (Morphological Extraction Theorem), will truly measure measured surface,
The minimum profile curvature radius of measured surface optional position are needed to have to be larger than the Probe-radius for measuring contact pilotage.At the same time, between sampling
Every Probe-radius can not be more than.So, the measuring method would generally be limited by small gauge head size so that precision is not reached
Require, while measurement efficiency is low.And non-contact optical measuring method is essentially all to be entered based on the reflection case of measured surface light
Row measurement, but micro-structure surface generally has precipitous convex-concave structure, and three-dimensional appearance is complicated, so using the side of optical measurement
Method also arrives restriction.It is SEM and its measuring method phase in Altitude that scanning electron microscope (SEM) method is measured to measured surface
Than with high lateral resolution (about 5nm), being to tested micro-structure surface by the stereoscopic measurement method for gradually obtaining image
Evaluated, but longitudinal resolution is depended on lateral resolution and angle of inclination, and detection efficiency is slower, so SEM
The measurement requirement of micro-structure surface cannot be met.
Microstructure unit in micro structure function surface typically has the characteristics that：Size is little, and quantity is more, distribution density
Greatly, three-dimensional appearance is complicated.These features so that the measurement to micro structure has larger difficulty.First, micro structure function surface
Size range it is general all in submillimeter level, size is less, it is difficult to conventional measurement facility measurement；It is large number of, if used
The mode for measuring one by one, then efficiency is low；Three-dimensional appearance is complicated, and the feature with geometric catastrophe then more increases measurement difficulty.Institute
With, common survey tool or instrument be all difficult in precision and efficiency meet the detection requirement of micro structure.
In view of above-mentioned defect, the design people, positive research and innovation in addition, to founding a kind of novel Microstructure surface
Detection method so as to more the value in industry.
The content of the invention
To solve above-mentioned technical problem, it is an object of the invention to provide a kind of detection method of micro-structure surface.
Technical scheme is as follows：
A kind of micro-structure surface detection method, it is characterised in that comprise the following steps：
S1：Image acquisition, acquisition target include detection template and micro-measuring structure to be checked, the detection template and micrometer to be checked
Structure is acquired under identical environment；
S2：Matching order, the detection template for collecting and microstructure image to be detected are input in computer, according to from
Order on the micro-measuring structure to be checked from left to right from top to bottom carries out progressive search, intercepts and institute on micro-measuring structure to be checked
The subgraph to be measured for stating detection template same size is matched；
S3：Gray-scale Matching, by the gray scale difference value of each pixel of correspondence position on calculation template picture and subgraph to be measured, then
Relevance degree is calculated by normalization crossover algorithm, the Gray-scale Matching value of each subgraph to be measured and detection template is obtained；
S4：Matching step-length is set, the maximum threshold values and minimum threshold values of Gray-scale Matching value is set, when Gray-scale Matching value is near most
During little threshold values, the step-length of progressive search is 10 pixels；When Gray-scale Matching value is near maximum threshold values, the step-length of progressive search is 1
Individual pixel；
S5：Sensitizing range is arranged, and arranges sensitizing range matching degree threshold values, when Gray-scale Matching value is more than the sensitizing range of setting
During matching degree threshold values, sensitizing range is set in this one region comprising subgraph to be measured of position planning, and in the sensitizing range of setting
Gray-scale Matching is carried out in domain, maximum match value is obtained, the maximum match value for obtaining is set to into the true match value of the subgraph to be measured,
After the matching of a subgraph to be measured is completed, mobile detection module to complete to match misaligned next to be measured of subgraph to be measured
Subgraph；
S6：The matching value of the subgraph to be measured for determining is carried out matching degree classification by labelling and output, and to different matching degrees
Grade arranges different colors and is distinguish between, by subgraph to be measured according to the grade of its matching for detecting in subgraph frame to be measured in addition
Corresponding color.
Further, in step S1, identical environment is identical photoenvironment, identical illumination visual angle.
Further, Gray-scale Matching is carried out in the sensitizing range for setting described in step S5, specially by the inspection
The order that surveying template carries out the 1 pixel movement of each step according to order from left to right from top to bottom in the sensitizing range is entered
Row matching.
Further, in step S6, matching degree classification is Pyatyi altogether, respectively matching degree below 0.65,
Matching degree 0.65-0.70, matching degree 0.70-0.80, matching degree 0.80-0.88, matching degree more than 0.88.
By such scheme, the present invention at least has advantages below：
(1) in this method, detection template and micro-measuring structure to be checked are gathered under same environment, it is ensured that this method is examined
Survey without loss of generality；
(2) this method arranges different matching step-lengths according to matching value, improves detection efficiency, saves what detection spent
Time；
(3) this method is provided with sensitizing range, it is to avoid easily produce asking for repeated matching using normalization matching algorithm
Topic；
(4) this method will detect that matching angle value is fed back on micro-measuring structure to be checked, beneficial to right by the frame of different colours
Micro-structure surface defect is shown.
Description of the drawings
Fig. 1, method of the present invention overhaul flow chart；
Fig. 2, micro-measuring structure schematic diagram to be checked of the invention；
Fig. 3, detection template structural representation of the present invention；
Fig. 4, sensitizing range structural representation of the present invention.
Specific embodiment
With reference to the accompanying drawings and examples, the specific embodiment of the present invention is described in further detail.Hereinafter implement
Example is for illustrating the present invention, but is not limited to the scope of the present invention.
Referring to Fig. 1 to Fig. 4, a kind of micro-structure surface detection method described in a preferred embodiment of the present invention, this method bag
Image acquisition is included, by fine detection, selects geometrical morphology preferably, scale error is little, and the flawless microstructure unit in surface is made
For detection template.Image acquisition is carried out to template and micro-measuring structure to be checked respectively.Detection template needs to ensure with micro-measuring structure to be checked
In identical photoenvironment, identical illumination carries out image acquisition under visual angle, gray scale is produced with eliminating illumination, visual angle factor
Affect, without loss of generality.
Matching order, is matched to image using program, due to developing computer reading graph order be from left to right, on to
Under.So, matching order is that the upper left corner on testing image starts, so that from left to right, order from top to bottom is passed according to step-length
Enter search, intercepted on testing image every time and matched with template with the subgraph of template image same size.
Gray-scale Matching, obtains the Gray-scale Matching value of template and each subgraph to be measured using normalization crossover algorithm (NCC).It is logical
The gray scale difference value of calculation template picture and each pixel of subgraph correspondence position to be measured is crossed, recycles normalization cross correlation algorithm to obtain
To relevance degree, as Gray-scale Matching value.
Matching step-length is set, and when matching template moves closer to microstructure unit position, Gray-scale Matching value can be gradually
Increase, if the progressive step-length of matching template is 1, a pixel of only advancing after the completion of matching every time.So matching efficiency
Low, speed is slow.As the characteristic distributions of micro structure are regular arrays, each micro structure is mutually separated with certain spacing.When template with
The actual position of to-be-measured cell farther out when, as image difference is huge, the dependency of respective pixel is little, so Gray-scale Matching value
It is little；And when template is nearer with the actual position of to-be-measured cell, Gray-scale Matching value can be incrementally increased, being incremented by property increases.
So, when template is near microstructure unit to be detected, matching value increase, when reaching, microstructure unit to be measured is true
During position, matching value reaches maximum, and when template is away from microstructure unit to be measured, matching value reduces again.I.e. in each micro- knot
Structure unit near zone, matching value is the trend of first increases and then decreases, has a matching value maximum, the maximum of the matching value
The true Gray-scale Matching value of microstructure unit as to be detected.
According to matching value and the relationship characteristic for treating micro-measuring structure distance, different restriction threshold values are set.When matching value hour,
Then represent that template is remote apart from the distance of microstructure unit to be measured, it is possible to search for using big step-length progressive (10 pixel), when
Matching value is big, then it represents that template is near apart from the distance of microstructure unit to be measured, so pass (1 pixel) using little step-length to enter to search
Rope.
Sensitizing range is arranged, and with the progressive matching microstructure unit of little step length searching, detects matching template and treats that micrometer is tied
The actual position of structure unit, Gray-scale Matching value now is matching value maximum in the field, attached in the position for treating micro-measuring structure
Closely there is many and close Gray-scale Matching value of maximum match value, easily produced using the search strategy of traditional normalization matching algorithm
Raw repeated matching.
When matching template progressive search is matched, it may appear that many with what maximum match value in the microstructure positional neighborhood was close to
There are multiple close Gray-scale Matching values around the microstructure unit of individual matching value, i.e., a micro structure testing goes out multiple matching degrees
Situation.Further, since in order to improve rate matched, and larger step size is provided with, in the progressive matching process of micro structure substep length
In, it is possible to the maximum match point of matching degree is skipped, the situation for causing the microstructure unit matching degree more low than actual value is caused
Matching result error.
Method in order to solve the two problems, using matching sensitizing range is arranged.By arranging sensitizing range matching degree
Threshold value.When Gray-scale Matching value is more than sensitizing range matching degree threshold value, i.e., near this position, certain limit, the scope delimited
It is sufficiently large, the position at the place of microstructure unit to be measured can be included.Then, in sensitizing range, fine match is carried out, is obtained
Maximum match value, treat the true match value of micro-measuring structure as the position.A matching sensitizing range is completed in matching template
Maximum match value after, carry out once relatively larger transverse progressive so that matching template leaves the region of the microstructure unit, promptly
Enter in the matching area of next microstructure unit.By this method, can avoid many near a microstructure unit
Secondary matching, can be accurately obtained the true match degree of the microstructure unit, again while accelerating matching speed.
Labelling and output, after the Gray-scale Matching value and matched position for searching out each microstructure unit to be measured.By
The mode for identifying is done in the position, and position and the Gray-scale Matching degree grade of microstructure unit intuitively can be recognized from image.With
This reaches the purpose for quickly recognizing each micro structure similarity degree.
In order to be able to more intuitively embody the difference of the matching degree of tested microstructure unit and template micro structure, according to obtaining
The matching value of the micro structure, sets matching degree grade.By arranging different matching degree grades, and to different matching degree etc.
The in addition different color of the Matching sub-image of level is distinguishing.Matching degree grade is divided into into following Pyatyi：Matching degree is less than 0.65,
Identified with red frame；Matching degree is 0.65 to 0.70, is indicated with purple frame；Matching degree is 0.70 to 0.80, with Huang
Color frame is indicated；Matching degree is 0.80 to 0.88, is indicated with green frame；Matching degree is more than 0.88, uses blue border
Sign.According to match grade to being passed judgment on by the matched position of microstructure unit and whole matching degree.
The matching value of each tested microstructure unit is depicted as into matching degree rectangular histogram, and according to different matching degree etc.
Level is represented with different colours.The pixel coordinate of each microstructure unit to be measured is converted into into line number and columns simultaneously, has been reached more
Intuitively, rapidly recognize position and the Gray-scale Matching value for being detected microstructure unit.
Can quickly be examined for micro-structure surface geometric feature sizes, spacing and surface defect by this method
Survey.
Size detecting method is as follows：Template micro structure is made into the template picture of different proportion size, with each different chi
Very little template picture is matched with each microstructure unit, compares Gray-scale Matching value of the microstructure unit under different size,
Find the size when Gray-scale Matching value is maximum.Then think the proportional sizes closest to the full-size(d) of the microstructure unit, institute
Using the size as the microstructure unit to be measured.The size of the microstructure unit is detected by this method.
Spacing detection is as follows：The micro structure list for obtaining is detected in sensitizing range according to hereinbefore matching in each micro structure
First position, by the position for comparing adjacent microstructures unit, you can the location of pixels for obtaining two adjacent microstructure units is poor, then leads to
Cross and calculate full-size(d) representated by each pixel, that is, obtain the true spacing of two micro structures.It is possible thereby to detect two
The spacing of adjacent microstructures.
Detection to surface defect：Two steps are broadly divided into, the first step is that matching template is first matched with picture to be measured,
Detect the position of microstructure unit in each sensitizing range；Second step is will to detect from each Gray-scale Matching sensitizing range
Microstructure unit, all carry out piecemeal division with detection template, be divided into several fritters.Matched based on Normalized Grey Level again
Method, is respectively compared the Gray-scale Matching value of detected microstructure unit and each corresponding little piecemeal in matching template.Correspondence point
The matching value of block is less, then illustrate that difference of the microstructure unit with template microstructure unit in the position is bigger, if difference
Larger fritter distribution is concentrated, then illustrate that tested micro-measuring structure has larger defect in the position, such as larger burr, draw
The manufacturing deficiencies such as trace, chipping.Testing result is detected in the segmented areas of the micro structure differed greatly with template, and is marked
Know, and pass through Program Generating X-Y scheme audio-visual picture.
The present invention at least has advantages below：
(1) in this method, detection template and micro-measuring structure to be checked are gathered under same environment, it is ensured that this method is examined
Survey without loss of generality；
(2) this method arranges different matching step-lengths according to matching value, improves detection efficiency, saves what detection spent
Time；
(3) this method is provided with sensitizing range, it is to avoid easily produce asking for repeated matching using normalization matching algorithm
Topic；
(4) this method will detect that matching angle value is fed back on micro-measuring structure to be checked, beneficial to right by the frame of different colours
Micro-structure surface defect is shown.
The above is only the preferred embodiment of the present invention, is not limited to the present invention, it is noted that for this skill
For the those of ordinary skill in art field, on the premise of without departing from the technology of the present invention principle, can also make it is some improvement and
Modification, these improve and modification also should be regarded as protection scope of the present invention.
Claims (4)
1. a kind of micro-structure surface detection method, it is characterised in that comprise the following steps：
S1：Image acquisition, acquisition target include detection template and micro-measuring structure to be checked, the detection template and micro-measuring structure to be checked
It is acquired under identical environment；
S2：Matching order, the detection template for collecting and microstructure image to be detected are input in computer, according to from described
Order on micro-measuring structure to be checked from left to right from top to bottom carries out progressive search, intercepts and the inspection on micro-measuring structure to be checked
The subgraph to be measured for surveying template same size is matched；
S3：Gray-scale Matching, by the gray scale difference value of each pixel of correspondence position on calculation template picture and subgraph to be measured, then passes through
Normalization crossover algorithm is calculated relevance degree, obtains the Gray-scale Matching value of each subgraph to be measured and detection template；
S4：Matching step-length is set, the maximum threshold values and minimum threshold values of Gray-scale Matching value is set, when Gray-scale Matching value is near minimum valve
During value, the step-length of progressive search is 10 pixels；When Gray-scale Matching value is near maximum threshold values, the step-length of progressive search is 1 picture
Element；
S5：Sensitizing range is arranged, and arranges sensitizing range matching degree threshold values, when Gray-scale Matching value is more than the sensitizing range matching of setting
During bottom valve value, sensitizing range is set in this one region comprising subgraph to be measured of position planning, and in the sensitizing range of setting
Gray-scale Matching is carried out, maximum match value is obtained, the maximum match value for obtaining the true match value of the subgraph to be measured is set to into, when complete
Into after the matching of a subgraph to be measured, mobile detection module to complete to match the misaligned son next to be measured of subgraph to be measured
Figure；
S6：The matching value of the subgraph to be measured for determining is carried out matching degree classification by labelling and output, and to different matching degree grades
Arrange different colors to be distinguish between, subgraph to be measured is corresponded in subgraph frame to be measured according to the grade of its matching for detecting
Color.
2. a kind of micro-structure surface detection method according to claim 1, it is characterised in that：Identical in step S1
Environment is identical photoenvironment, identical illumination visual angle.
3. a kind of micro-structure surface detection method according to claim 2, it is characterised in that：Set described in step S5
Gray-scale Matching is carried out in fixed sensitizing range, specially by the detection template in the sensitizing range according to from left to right from
The order of top to bottm carries out the order of 1 pixel movement of each step and is matched.
4. a kind of micro-structure surface detection method according to claim 3, it is characterised in that：It is in step S6, described
Matching degree classification is altogether Pyatyi, respectively matching degree below 0.65, matching degree 0.65-0.70, matching degree 0.70-0.80,
With degree 0.80-0.88, matching degree more than 0.88.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610980599.9A CN106548474A (en) | 2016-11-08 | 2016-11-08 | A kind of micro-structure surface detection method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610980599.9A CN106548474A (en) | 2016-11-08 | 2016-11-08 | A kind of micro-structure surface detection method |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN106548474A true CN106548474A (en) | 2017-03-29 |
Family
ID=58394827
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201610980599.9A Pending CN106548474A (en) | 2016-11-08 | 2016-11-08 | A kind of micro-structure surface detection method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106548474A (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110836641A (en) * | 2019-11-25 | 2020-02-25 | 江苏集萃精凯高端装备技术有限公司 | Detection method and detection equipment for three-dimensional size of part special-shaped surface microstructure |
| CN113837184A (en) * | 2021-09-27 | 2021-12-24 | 常州市宏发纵横新材料科技股份有限公司 | Mosquito detection method, device and storage medium |
| CN114091620A (en) * | 2021-12-01 | 2022-02-25 | 常州市宏发纵横新材料科技股份有限公司 | Template matching detection method, computer equipment and storage medium |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104034266A (en) * | 2014-06-16 | 2014-09-10 | 中国科学院长春光学精密机械与物理研究所 | Surface microstructure based high-accuracy length detection method |
| CN104732476A (en) * | 2015-03-23 | 2015-06-24 | 中国民航大学 | Microstructure low-overlapping-degree three-dimensional splicing method based on optical nondestructive testing |
-
2016
- 2016-11-08 CN CN201610980599.9A patent/CN106548474A/en active Pending
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104034266A (en) * | 2014-06-16 | 2014-09-10 | 中国科学院长春光学精密机械与物理研究所 | Surface microstructure based high-accuracy length detection method |
| CN104732476A (en) * | 2015-03-23 | 2015-06-24 | 中国民航大学 | Microstructure low-overlapping-degree three-dimensional splicing method based on optical nondestructive testing |
Non-Patent Citations (1)
| Title |
|---|
| 赵亮 等: "基于灰度匹配理论的跨尺度微结构表面检测方法研究", 《制造技术与机床》 * |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110836641A (en) * | 2019-11-25 | 2020-02-25 | 江苏集萃精凯高端装备技术有限公司 | Detection method and detection equipment for three-dimensional size of part special-shaped surface microstructure |
| CN110836641B (en) * | 2019-11-25 | 2024-05-28 | 江苏集萃精凯高端装备技术有限公司 | Method and equipment for detecting three-dimensional size of special-shaped surface microstructure of part |
| CN113837184A (en) * | 2021-09-27 | 2021-12-24 | 常州市宏发纵横新材料科技股份有限公司 | Mosquito detection method, device and storage medium |
| CN113837184B (en) * | 2021-09-27 | 2022-06-10 | 常州市宏发纵横新材料科技股份有限公司 | Mosquito detection method, device and storage medium |
| CN114091620A (en) * | 2021-12-01 | 2022-02-25 | 常州市宏发纵横新材料科技股份有限公司 | Template matching detection method, computer equipment and storage medium |
| CN114091620B (en) * | 2021-12-01 | 2022-06-03 | 常州市宏发纵横新材料科技股份有限公司 | Template matching detection method, computer equipment and storage medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN112907528B (en) | Point cloud-to-image-based composite material laying wire surface defect detection and identification method | |
| CN105913093B (en) | A kind of template matching method for Text region processing | |
| CN104331699B (en) | A kind of method that three-dimensional point cloud planarization fast search compares | |
| CN109064461A (en) | A kind of detection method of surface flaw of steel rail based on deep learning network | |
| CN102135417B (en) | Full-automatic three-dimension characteristic extracting method | |
| CN110490874A (en) | Weaving cloth surface flaw detecting method based on YOLO neural network | |
| CN108319949A (en) | Mostly towards Ship Target Detection and recognition methods in a kind of high-resolution remote sensing image | |
| CN108765389A (en) | A kind of microcosmic wafer surface defects image detecting method | |
| CN109711288A (en) | Remote sensing ship detecting method based on feature pyramid and distance restraint FCN | |
| CN107274393B (en) | Monocrystaline silicon solar cell piece detection method of surface flaw based on grid line detection | |
| CN101598677B (en) | Online three-dimensional recognition method of fabric defects | |
| CN104597057B (en) | A kind of column Diode facets defect detecting device based on machine vision | |
| CN105787486A (en) | Method for detecting girder cracks based on image processing | |
| CN105574527A (en) | Quick object detection method based on local feature learning | |
| Di Angelo et al. | Automatic dimensional characterisation of pottery | |
| CN114066820A (en) | Fabric defect detection method based on Swin-transducer and NAS-FPN | |
| CN106548474A (en) | A kind of micro-structure surface detection method | |
| CN113222982A (en) | Wafer surface defect detection method and system based on improved YOLO network | |
| CN109949287A (en) | A kind of fabric defects detection method based on adaptivenon-uniform sampling and template correction | |
| CN109141384A (en) | Acquisition and preprocess method to data before the detection after subway tunnel completion | |
| CN103839271A (en) | Method for image texture synthesis based on optimum matching | |
| CN109410241A (en) | The metamorphic testing method of image-region growth algorithm | |
| CN105740859B (en) | A kind of three-dimensional interest point detecting method based on geometric measures and sparse optimization | |
| CN112233104B (en) | Real-time displacement field and strain field detection method, system, device and storage medium | |
| CN109447968A (en) | The metamorphic testing system of image-region growth algorithm |
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 | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20170329 |