CN105954292B - Underwater works surface crack detection device based on the bionical vision of compound eye and method - Google Patents
Underwater works surface crack detection device based on the bionical vision of compound eye and method Download PDFInfo
- Publication number
- CN105954292B CN105954292B CN201610282173.6A CN201610282173A CN105954292B CN 105954292 B CN105954292 B CN 105954292B CN 201610282173 A CN201610282173 A CN 201610282173A CN 105954292 B CN105954292 B CN 105954292B
- Authority
- CN
- China
- Prior art keywords
- compound eye
- image
- underwater
- underwater works
- eye array
- 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
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
- G01N21/84—Systems specially adapted for particular applications
- G01N21/88—Investigating the presence of flaws or contamination
- G01N21/8851—Scan or image signal processing specially adapted therefor, e.g. for scan signal adjustment, for detecting different kinds of defects, for compensating for structures, markings, edges
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
- G01N21/84—Systems specially adapted for particular applications
- G01N21/88—Investigating the presence of flaws or contamination
- G01N21/8851—Scan or image signal processing specially adapted therefor, e.g. for scan signal adjustment, for detecting different kinds of defects, for compensating for structures, markings, edges
- G01N2021/8887—Scan or image signal processing specially adapted therefor, e.g. for scan signal adjustment, for detecting different kinds of defects, for compensating for structures, markings, edges based on image processing techniques
Abstract
The invention discloses a kind of underwater works surface crack detection device and method based on the bionical vision of compound eye, the contradiction of single aperture imaging system visual field and resolution ratio in imaging process can effectively be overcome, the complementary information in every image can be utilized to the super-resolution rebuilding process of same target multiple image, successfully manage underwater complex light environment and noise jamming, obtain the abundant clearly high-definition picture of details, utilize the underwater works surface crack self-adapting detecting method based on gray-scale map, foreground information in image can more effectively be extracted, reduce Crack Detection range, the influence of underwater noise fracture detection is greatly reduced, improve Crack Detection precision, it has a good application prospect.
Description
Technical field
The present invention relates to a kind of underwater works surface crack detection device and method based on the bionical vision of compound eye, belong to
Bionical and Video Analysis Technology field.
Background technology
Underwater works, including reservoir dam, bridge pier etc. are a states as the important products in process of economic development
The capital construction facility of family, is the important engineering for being related to national economy, is played an important role in agricultural, traffic etc..
For these underwater works while playing enormous benefits, there is also the risks that accident occurs.To find out its cause, crack is to cause
The arch-criminal of underwater works major accident.The crack on underwater works structure surface is the accumulation of internal injury during use
External concentrated expression after reaching a certain level, if it is possible to the feature in crack is understood in time and is subject to tracking and monitoring, it will
Effective protection public property and people life property safety.
Currently, underwater picture is the important channel for carrying out nondestructive inspection to underwater works.But due to Underwater Optical wire loop
The complexity in border, light scattering absorb more serious, and noise is also more difficult to remove, so to the vision system of underwater computer
System proposes stern challenge.Therefore, traditional air imaging system is difficult to directly apply to underwater environment.With bionical skill
The multiple aperture Vision imaging system of art, the especially development of vision bionics technology, imitative Compound Eye of Insects has been applied to radar, leads
The fields such as bullet can effectively solve the contradiction of visual field and resolution ratio.
Therefore, how the multiple aperture imaging system based on the bionical vision mechanism of compound eye to be applied in underwater environment detection
It goes, to cope with underwater complex luminous environment, reduces underwater noise interference, image resolution ratio is promoted, to improve underwater works
Crack Detection precision is current urgent problem.
Invention content
The light that the invention aims to be easy to meet in existing atmospheric optics imaging system under water environment detection
Scattering, decaying are serious, the difficulties such as noise difficulty removal, and caused by image resolution ratio the problem of declining.
In order to achieve the above object, the technical solution adopted in the present invention is：
A kind of underwater works surface crack detection device based on the bionical vision of compound eye, it is characterised in that：Including encapsulation
Compound eye array, LED secondary light sources inside anti-sump and Embedded Processor,
The compound eye array, for carrying out video image shooting, including seven CCD cameras to underwater works, wherein
Six CCD cameras are arranged in regular hexagon, and the 7th CCD camera is located at the center of equilateral hexagon, and seven CCD take the photograph
The equal position of camera is fixed, and to same direction and is carried out at the same time video image shooting,
The LED secondary light sources, for providing illumination compensation, including several LED lamp beads, each LED light for compound eye array
Pearl surrounds circular ring shape, and is fixed on the border of the 7th CCD camera,
The Embedded Processor, the underwater works image for being shot to compound eye array are handled,
The output end of each CCD camera in the compound eye array is connected with Embedded Processor, the LED auxiliary
Light source is connected with Embedded Processor, and the break-make of each LED lamp bead in it is controlled by Embedded Processor.
Underwater works surface crack detection device above-mentioned based on the bionical vision of compound eye, it is characterised in that：It is described anti-
The one side of sump is flat glass face, and lap is steel plate surface, and the compound eye array is obtained by flat glass face and constructed under water
The anti-water of the video image of object, the anti-sump is 100 meters.
Underwater works surface crack detection device above-mentioned based on the bionical vision of compound eye, it is characterised in that：It is described multiple
Six CCD cameras are arranged in regular hexagon in eye array, and are arranged according to the regular hexagon of Compound Eye of Insects, adjacent two
Spacing between CCD camera optical axis is 10 centimetres.
Underwater works surface crack detection device above-mentioned based on the bionical vision of compound eye, it is characterised in that：The LED
The quantity of lamp bead is 16.
Underwater works surface crack detection device above-mentioned based on the bionical vision of compound eye, it is characterised in that：It is described embedding
It includes ARM subsystems, DSP subsystems, power module, video image interface module, DDR memory modules, Flash to enter formula processor
Memory module, data outputting module, debugging interface module, the power module are entire underwater works surface crack detection dress
Power supply is set, the video image interface module is used to provide the data-interface between compound eye array and Embedded Processor, described
DDR modules are used to store the video image information of compound eye array acquisition, and provide data source for ARM subsystems, DSP subsystems,
The Flash memory modules are used for export place from Embedded Processor for storing program code, the data outputting module
Data after reason, the debugging interface module use when being debugged for program code, and the ARM subsystems, DSP subsystems are constituted
The host processing system of Embedded Processor, the host processing system are carried out for the video image information to compound eye array acquisition
LED lamp bead is turned on and off in super-resolution reconstruction, the analysis of underwater works Crack Detection and LED secondary light sources.
Underwater works surface crack detection device above-mentioned based on the bionical vision of compound eye, it is characterised in that：It is described anti-
Sump is sunk under the water or is equipped on underwater robot using towing cable and enters under water.
Underwater works table based on the above-mentioned underwater works surface crack detection device based on the bionical vision of compound eye
Face crack detection method, it is characterised in that：Include the following steps,
Step (A) initializes underwater works surface crack detection device
Compound eye array, Embedded Processor are subjected to Initialize installation, and entire underwater works surface crack detection dress
It sets and is sealed in anti-sump, underwater robot is sunk under the water or be equipped on by towing cable and is entered under water, underwater works table is made
Face crack detecting device is started to work；
Step (B) obtains compound eye array image sequence and storage
Compound eye array image sequence is obtained by compound eye array, is transmitted, and is stored to the DDR in Embedded Processor
In memory module；
Step (C), the pretreatment of compound eye array image
Compound eye array image is obtained from DDR memory modules, and compound eye array image is surpassed by Embedded Processor
Resolution reconstruction obtains the high-definition picture of underwater works；
Step (D), the detection and analysis of underwater works surface crack and data storage
The high-definition picture of underwater works is obtained according to step (C), utilizes gray scale morphology gradient and adaptive ash
The method of degree threshold value is split extraction to doubtful crack area, and the high-definition picture for marking doubtful crack area is deposited
It stores up in DDR memory modules；
Step (F), the output of underwater works surface crack detection data
After underwater works surface crack detection device completes underwater operation, user is read original by data outputting module
Compound eye array image and mark the high-definition picture of doubtful crack area, carry out underwater works surface crack detection data
Output, complete underwater works surface crack detection.
It splits on the underwater works surface of underwater works surface crack detection device above-mentioned based on the bionical vision of compound eye
Stitch detection method, it is characterised in that：Step (C) carries out super-resolution reconstruction by Embedded Processor to compound eye array image,
Include the following steps,
(C1) compound eye array image sequence is read from DDR memory modules；
(C2) compound eye array image sequence is registrated, makes each CCD camera acquisition of synchronization in compound eye array
Image has been aligned, if reference picture is f (x ', y '), compound eye array image to be registered is g (x, y), x, y) and (x ', y ')
It is the pixel containing identical information respectively in compound eye array image to be registered is g (x, y) and reference picture f (x ', y ')
Coordinate, image caused by the 7th CCD camera is image produced by sequence synchronization other six CCD cameras
Reference picture, compound eye array image to be registered are remaining all images in image sequence in addition to reference picture；
Then both sides relation formula, as shown in formula (1),
G (x, y)=f (x+a1x+a2y+a3,y+a1y-a2x+a4) (1)
Wherein, x '=x+a1x+a2y+a3, y '=y+a1y-a2x+a4a1, a2, a3, a4It is four parameters to be determined, a3For
Horizontal displacement, a4For vertical displacement amount, rotation angle isTo reference
Image f (x ', y ') does Taylor expansion and retains first item, error function E is acquired, to error function E about a1, a2, a3, a4Point
Local derviation is not sought, and it is zero to enable its local derviation, obtains a1, a2, a3, a4Corresponding value, you can obtain the level of compound eye array image sequence
The estimated value of displacement, vertical displacement amount and rotation angle；
(C3) image reconstruction, the compound eye array image obtained according to (C2) are carried out to the compound eye array image sequence after registration
The estimated value of corresponding displacement component and rotation angle waits for the input quantity of reconstructed image as compound eye array image；
(C4) convolution method is normalized using structure adaptive when image reconstruction, multiframe low-resolution image is reconstructed,
To obtain high-definition picture.
It splits on the underwater works surface of underwater works surface crack detection device above-mentioned based on the bionical vision of compound eye
Stitch detection method, it is characterised in that：(C4) convolution method is normalized using structure adaptive when image reconstruction, to multiframe low resolution
Image is reconstructed, and includes the following steps,
(1) multiframe low-resolution image is analyzed according to noise robustness, estimates noise pixel；
(2) after obtaining noise pixel, the height by the size of multiframe low-resolution image is changed and for normalizing convolution
This filter direction obtains the sharp keen high-resolution reconstruction image of details.
It splits on the underwater works surface of underwater works surface crack detection device above-mentioned based on the bionical vision of compound eye
Stitch detection method, it is characterised in that：Step (D) is split using gray scale morphology gradient and the method for adaptive gray threshold to doubtful
Seam region is split extraction, and by the high-definition picture storage to DDR memory modules for marking doubtful crack area, wraps
Include following steps,
(D1) Double Thresholding Segmentation is carried out to the high-definition picture of underwater works, obtains corresponding target area MhighWith
Mlow, be partitioned into underwater works main body from high-definition picture, remove background area, be further reduced aquatic organism and
The influence of rubbish noise fracture identification；
(D2) two are calculated into column hisgram back projection to the high-definition picture of the underwater works after Double Thresholding Segmentation
Target area MhighAnd MlowGray value overlap ratio, when ratio be higher than threshold xi when, then the corresponding pixel of the gray scale be figure
The foreground area of picture, as underwater works main body F, shown in calculation formula such as formula (2),
Wherein, Hhigh(r) and Hlow(r) it is respectively two target area MhighAnd MlowGrey level histogram, r is gray value,
Threshold xi is to overlap ratio to gray value during the experiment to carry out observing obtained empirical value, takes 0.8；
(D3) it combines and is partitioned into underwater works main body F, according to the characteristic features automatic identification high-definition picture in crack
Doubtful crack area, doubtful crack area is detected according to the low frequency strip connected domain in high-definition picture, detects to doubt
After crack area, with red minimum external oval marks；
(D4) high-definition picture of doubtful crack area will be marked to export, and carries out storage and arrives DDR memory modules.
The beneficial effects of the invention are as follows：The underwater works surface crack detection dress based on the bionical vision of compound eye of the present invention
It sets and method, can effectively overcome the contradiction of single aperture imaging system visual field and resolution ratio in imaging process, it is more to same target
The super-resolution rebuilding process of width image can utilize the complementary information in every image, successfully manage underwater complex light environment and
Noise jamming obtains the abundant clearly high-definition picture of details, certainly using the underwater works surface crack based on gray-scale map
Detection method is adapted to, foreground information in image can be more effectively extracted, Crack Detection range is reduced, underwater noise is greatly reduced
The influence of fracture detection, improves Crack Detection precision, has a good application prospect.
Description of the drawings
Fig. 1 is the system block diagram of the underwater works surface crack detection device based on the bionical vision of compound eye of the present invention.
Fig. 2 is the compound eye array of the present invention and the arrangement mode schematic diagram of secondary light source.
Fig. 3 is the system block diagram of the Embedded Processor of the present invention.
Fig. 4 is the flow chart of the underwater works surface crack detection method based on the bionical vision of compound eye of the present invention.
Specific implementation mode
Below in conjunction with Figure of description, the present invention will be further described.Following embodiment is only used for clearly
Illustrate technical scheme of the present invention, and not intended to limit the protection scope of the present invention.
As shown in Figure 1, the underwater works surface crack detection device based on the bionical vision of compound eye of the present invention, including envelope
Compound eye array, LED secondary light sources inside anti-sump and Embedded Processor,
The compound eye array, for carrying out video image shooting, including seven CCD cameras to underwater works, wherein
Six CCD cameras are arranged in regular hexagon, and the 7th CCD camera is located at the center of equilateral hexagon, and seven CCD take the photograph
The equal position of camera is fixed, and to same direction and is carried out at the same time video image shooting,
The LED secondary light sources, for providing illumination compensation, including several LED lamp beads, each LED light for compound eye array
Pearl surrounds circular ring shape, and is fixed on the border of the 7th CCD camera,
The Embedded Processor, the underwater works image for being shot to compound eye array are handled,
The output end of each CCD camera in the compound eye array is connected with Embedded Processor, the LED auxiliary
Light source is connected with Embedded Processor, and the break-make of each LED lamp bead in it is controlled by Embedded Processor.
The one side of the anti-sump is flat glass face, and lap is steel plate surface, and the compound eye array passes through flat glass
Face obtains the video image of underwater works, and the anti-water of the anti-sump is 100 meters, and the anti-sump is sunk to using towing cable
Under water or underwater robot is equipped under water, can be selected according to the environment of water body, if the water that mobility is bigger
Body (water body environment is complicated severe) may be selected to be equipped on underwater robot under water.
As shown in Fig. 2, six CCD cameras are arranged in regular hexagon in the compound eye array, and according to Compound Eye of Insects
Regular hexagon arranges, and the spacing d between adjacent two CCD camera optical axises is 10 centimetres, and the quantity of the LED lamp bead is 16
, compound eye array can be made to remain to obtain the suitable image of brightness in insufficient light.
It is connect as shown in figure 3, the Embedded Processor includes ARM subsystems, DSP subsystems, power module, video image
Mouth mold block, DDR memory modules, Flash memory modules, data outputting module, debugging interface module, the power module are entire
Underwater works surface crack detection device is powered, and the video image interface module is for providing compound eye array and embedded place
Data-interface between reason machine, the DDR modules are used to store the video image information of compound eye array acquisition, and are ARM subsystems
System, DSP subsystems provide data source, and for storing program code, the data outputting module is used for the Flash memory modules
Export treated the data from Embedded Processor, the debugging interface module uses when being debugged for program code, described
ARM subsystems, DSP subsystems constitute the host processing system of Embedded Processor, the host processing system, for compound eye array
The video image information of acquisition carries out super-resolution reconstruction, LED in the analysis of underwater works Crack Detection and LED secondary light sources
Lamp bead is turned on and off.
As shown in figure 4, the underwater works surface crack detection device based on the bionical vision of compound eye of the present invention is underwater
Structures surface crack detection method, includes the following steps,
Step (A), initialization underwater works surface crack detection device (hardware initialization)
Compound eye array, Embedded Processor are subjected to Initialize installation and (complete setting, the inspection of each intrinsic parameters of the camera
Whether normal look into the work of Embedded Processor modules, prepare for compound eye imaging), and entire underwater works surface crack
Detection device is sealed in anti-sump, and underwater robot is sunk under the water or be equipped on by towing cable and is entered under water, underwater structure is made
Build the start-up operation of object surface crack detection device；
Step (B) obtains compound eye array image sequence and storage
Compound eye array image sequence is obtained by compound eye array, is transmitted, and is stored to the DDR in Embedded Processor
In memory module；
Step (C), the pretreatment of compound eye array image
Compound eye array image is obtained from DDR memory modules, and compound eye array image is surpassed by Embedded Processor
Resolution reconstruction obtains the high-definition picture of underwater works, includes the following steps,
(C1) compound eye array image sequence is read from DDR memory modules；
(C2) compound eye array image sequence is registrated, makes each CCD camera acquisition of synchronization in compound eye array
Image has been aligned, if reference picture is f (x ', y '), compound eye array image to be registered is g (x, y), (x, y) and (x ', y ')
It is the pixel containing identical information respectively in compound eye array image to be registered is g (x, y) and reference picture f (x ', y ')
Coordinate, image caused by the 7th CCD camera is image produced by sequence synchronization other six CCD cameras
Reference picture, compound eye array image to be registered are remaining all images in image sequence in addition to reference picture；
Then both sides relation formula, as shown in formula (1),
G (x, y)=f (x+a1x+a2y+a3,y+a1y-a2x+a4) (1)
Wherein, x '=x+a1x+a2y+a3, y '=y+a1y-a2x+a4, a1, a2, a3, a4It is four parameters to be determined, a3For
Horizontal displacement, a4For vertical displacement amount, rotation angle isTo reference
Image f (x ', y ') does Taylor expansion and retains first item, error function E is acquired, to error function E about a1, a2, a3, a4Point
Local derviation is not sought, and it is zero to enable its local derviation, obtains a1, a2, a3, a4Corresponding value, you can obtain the level of compound eye array image sequence
The estimated value of displacement, vertical displacement amount and rotation angle；
Here it is to improve registration accuracy, reference picture f (x ', y ') can also be joined according to the displacement and rotation that have acquired
Number is converted, and new reference picture f ' is obtained, and is substituted the reference picture f (x ', y ') in above-mentioned steps, is iterated, and more
New displacement and rotation parameter of the image subject to registration relative to original reference image f, the suspension of iterative process is by preset threshold value
It is controlled with maximum iteration, if the sum of absolute value of kinematic parameter of estimation is more than preset threshold value or iterations reach
To maximum iteration, then stop iterative process, in of the invention, threshold value is set to be 10-1, maximum iteration 102。
In traditional super-resolution reconstruction algorithm, " reference picture " is any one width in image sequence, to calculate
It is convenient, piece image is generally selected, in the present invention, because there are multiple camera shooting machine simultaneously imagings, selects the 7th CCD
Image sequence caused by video camera is the reference picture of image produced by synchronization other six CCD cameras, " to be registered
Compound eye array image " is that remaining all image in image sequence in addition to reference picture refer in the present invention synchronization,
Image caused by other six CCD cameras, process of image registration will obtain image subject to registration relative to reference picture
Displacement (including horizontal displacement, vertical displacement amount) and rotation amount, so after the completion of registration, what compound eye image sequence generated
For image there is no changing, what is obtained is displacement and rotation amount of each image subject to registration relative to reference picture, the two parameters
Will be in super-resolution reconstruction step, that is, used in (C3), it is used to compound eye image sequence being aligned to reference picture, then into
Row reconstruct；
(C3) image reconstruction, the compound eye array image obtained according to (C2) are carried out to the compound eye array image sequence after registration
The estimated value of corresponding displacement component and rotation angle waits for the input quantity of reconstructed image as compound eye array image；
(C4) convolution method is normalized using structure adaptive when image reconstruction, multiframe low-resolution image is reconstructed,
To obtain high-definition picture comprising following steps,
(1) multiframe low-resolution image is analyzed according to noise robustness, estimates noise pixel, can greatly reduces to noise
Influence；
(2) after obtaining noise pixel, the height by the size of multiframe low-resolution image is changed and for normalizing convolution
This filter direction obtains the sharp keen high-resolution reconstruction image of details；
Step (D), the detection and analysis of underwater works surface crack and data storage
The high-definition picture of underwater works is obtained according to step (C), utilizes gray scale morphology gradient and adaptive ash
The method of degree threshold value is split extraction to doubtful crack area, and the high-definition picture for marking doubtful crack area is deposited
It stores up in DDR memory modules, includes the following steps,
(D1) Double Thresholding Segmentation is carried out to the high-definition picture of underwater works, obtains corresponding target area MhighWith
Mlow, be partitioned into underwater works main body from high-definition picture, remove background area, be further reduced aquatic organism and
The influence of rubbish noise fracture identification；
(D2) two are calculated into column hisgram back projection to the high-definition picture of the underwater works after Double Thresholding Segmentation
Target area MhighAnd MlowGray value overlap ratio, when ratio be higher than threshold xi when, then the corresponding pixel of the gray scale be figure
The foreground area of picture, as underwater works main body F, shown in calculation formula such as formula (2),
Wherein, Hhigh(r) and Hlow(r) it is respectively two target area MhighAnd MlowGrey level histogram, r is gray value,
Threshold xi is to overlap ratio to gray value during the experiment to carry out observing obtained empirical value, takes 0.8；
(D3) it combines and is partitioned into underwater works main body F, according to the characteristic features automatic identification high-definition picture in crack
Doubtful crack area, doubtful crack area is detected according to the low frequency strip connected domain in high-definition picture, detects to doubt
After crack area, with red minimum external oval marks；
(D4) high-definition picture of doubtful crack area will be marked to export, and carries out storage and arrives DDR memory modules；
Step (F), the output of underwater works surface crack detection data
After underwater works surface crack detection device completes underwater operation, user is read original by data outputting module
Compound eye array image and mark the high-definition picture of doubtful crack area, carry out underwater works surface crack detection data
Output, complete underwater works surface crack detection.
In conclusion the underwater works surface crack detection device and method based on the bionical vision of compound eye of the present invention,
The contradiction that single aperture imaging system visual field and resolution ratio in imaging process can effectively be overcome surpasses same target multiple image
Resolution reconstruction process can utilize the complementary information in every image, successfully manage underwater complex light environment and noise jamming,
The abundant clearly high-definition picture of details is obtained, the underwater works surface crack self-adapting detecting side based on gray-scale map is utilized
Method can more effectively extract foreground information in image, reduce Crack Detection range, and the detection of underwater noise fracture is greatly reduced
Influence, improve Crack Detection precision, have a good application prospect.
The basic principles and main features and advantage of the present invention have been shown and described above.The technical staff of the industry should
Understand, the present invention is not limited to the above embodiments, and the above embodiments and description only describe the originals of the present invention
Reason, without departing from the spirit and scope of the present invention, various changes and improvements may be made to the invention, these changes and improvements
It all fall within the protetion scope of the claimed invention.The claimed scope of the invention is by appended claims and its equivalent circle
It is fixed.
Claims (10)
1. the underwater works surface crack detection device based on the bionical vision of compound eye, it is characterised in that：Including being encapsulated in waterproof
Compound eye array, LED secondary light sources inside storehouse and Embedded Processor,
The compound eye array, for carrying out video image shooting, including seven CCD cameras, six of which to underwater works
CCD camera is arranged in regular hexagon, and the 7th CCD camera is located at the center of equilateral hexagon, seven CCD cameras
Equal position is fixed, and to same direction and is carried out at the same time video image shooting,
The LED secondary light sources, for providing illumination compensation for compound eye array, including several LED lamp beads, each LED lamp bead are enclosed
At circular ring shape, and it is fixed on the border of the 7th CCD camera,
The Embedded Processor, the underwater works image for being shot to compound eye array are handled,
The output end of each CCD camera in the compound eye array is connected with Embedded Processor, the LED secondary light sources
It is connected with Embedded Processor, the break-make of each LED lamp bead in it is controlled by Embedded Processor.
2. the underwater works surface crack detection device according to claim 1 based on the bionical vision of compound eye, feature
It is：The one side of the anti-sump is flat glass face, and lap is steel plate surface, and the compound eye array is obtained by flat glass face
The video image of underwater works is taken, the anti-water of the anti-sump is 100 meters.
3. the underwater works surface crack detection device according to claim 1 based on the bionical vision of compound eye, feature
It is：Six CCD cameras are arranged in regular hexagon in the compound eye array, and are arranged according to the regular hexagon of Compound Eye of Insects,
Spacing between two adjacent CCD camera optical axises is 10 centimetres.
4. the underwater works surface crack detection device according to claim 1 based on the bionical vision of compound eye, feature
It is：The quantity of the LED lamp bead is 16.
5. the underwater works surface crack detection device according to claim 1 based on the bionical vision of compound eye, feature
It is：The Embedded Processor includes that ARM subsystems, DSP subsystems, power module, video image interface module, DDR are deposited
Module, Flash memory modules, data outputting module, debugging interface module are stored up, the power module is entire underwater works table
Face crack detecting device power supply, the video image interface module are used to provide the number between compound eye array and Embedded Processor
According to interface, the DDR modules are used to store the video image information of compound eye array acquisition, and are ARM subsystems, DSP subsystems
Data source is provided, the Flash memory modules are used for for storing program code, the data outputting module from embedded processing
Export treated data in machine, the debugging interface module use when being debugged for program code, the ARM subsystems, DSP
Subsystem constitutes the host processing system of Embedded Processor, the host processing system, for the video figure to compound eye array acquisition
As information carry out super-resolution reconstruction, underwater works Crack Detection analysis and LED secondary light sources in LED lamp bead unlatching or
It closes.
6. the underwater works surface crack detection device according to claim 1 or 2 based on the bionical vision of compound eye, special
Sign is：The anti-sump is sunk under the water or is equipped on underwater robot using towing cable and enters under water.
7. the underwater structure based on the underwater works surface crack detection device described in claim 1 based on the bionical vision of compound eye
Build object surface crack detection method, it is characterised in that：Include the following steps,
Step (A) initializes underwater works surface crack detection device
Compound eye array, Embedded Processor are subjected to Initialize installation, and entire underwater works surface crack detection device is close
It is enclosed in anti-sump, underwater robot is sunk under the water or be equipped on by towing cable and is entered under water, underwater works surface is made to split
Detection device is stitched to start to work；
Step (B) obtains compound eye array image sequence and storage
Compound eye array image sequence is obtained by compound eye array, is transmitted, and is stored to the DDR storages in Embedded Processor
In module；
Step (C), the pretreatment of compound eye array image
Compound eye array image is obtained from DDR memory modules, and super-resolution is carried out to compound eye array image by Embedded Processor
Rate reconstructs, the process of the super-resolution reconstruction be by image caused by the 7th CCD camera be sequence synchronization its
The reference picture of image produced by his six CCD cameras, compound eye array image to be registered are that reference chart is removed in image sequence
Remaining all image as outside, and the specific method using Taylor expansion carries out image registration, and in image reconstruction using knot
Structure adaptively normalizes convolution method and is reconstructed, and obtains the high-definition picture of underwater works；
Step (D), the detection and analysis of underwater works surface crack and data storage
The high-definition picture of underwater works is obtained according to step (C), utilizes gray scale morphology gradient and adaptive gray scale threshold
The method of value is split extraction to doubtful crack area, and the high-definition picture storage for marking doubtful crack area is arrived
In DDR memory modules；
Step (F), the output of underwater works surface crack detection data
After underwater works surface crack detection device completes underwater operation, user reads original answer by data outputting module
Eye array image and the high-definition picture for marking doubtful crack area carry out the defeated of underwater works surface crack detection data
Go out, completes the detection of underwater works surface crack.
8. the underwater structure of the underwater works surface crack detection device according to claim 7 based on the bionical vision of compound eye
Build object surface crack detection method, it is characterised in that：Step (C) carries out oversubscription by Embedded Processor to compound eye array image
Resolution reconstructs, and includes the following steps,
(C1) compound eye array image sequence is read from DDR memory modules；
(C2) compound eye array image sequence is registrated, makes the image that each CCD camera of synchronization obtains in compound eye array
Be aligned, if reference picture is f (x ', y '), compound eye array image to be registered is g (x, y), (x, y) and (x ', y ') for containing
There is the seat of the pixel of identical information respectively in compound eye array image to be registered is g (x, y) and reference picture f (x ', y ')
Mark, image caused by the 7th CCD camera are the reference of image produced by sequence synchronization other six CCD cameras
Image, compound eye array image to be registered are remaining all images in image sequence in addition to reference picture；
Then both sides relation formula, as shown in formula (1),
G (x, y)=f (x+a1x+a2y+a3, y+a1y-a2x+a4) (1)
Wherein, x '=x+a1x+a2y+a3, y '=y+a1y-a2x+a4, a1, a2, a3, a4It is four parameters to be determined, a3For level
Displacement, a4For vertical displacement amount, rotation angle isTo reference chart
Picture f (x ', y ') does Taylor expansion and retains first item, error function E is acquired, to error function E about a1, a2, a3, a4Respectively
Local derviation is sought, and it is zero to enable its local derviation, obtains a1, a2, a3, a4Corresponding value, you can obtain the horizontal position of compound eye array image sequence
The estimated value of shifting amount, vertical displacement amount and rotation angle；
(C3) image reconstruction is carried out to the compound eye array image sequence after registration, is corresponded to according to the compound eye array image that (C2) is obtained
Displacement component and rotation angle estimated value, the input quantity of reconstructed image is waited for as compound eye array image；
(C4) convolution method is normalized using structure adaptive when image reconstruction, multiframe low-resolution image is reconstructed, to
Obtain high-definition picture.
9. the underwater structure of the underwater works surface crack detection device according to claim 8 based on the bionical vision of compound eye
Build object surface crack detection method, it is characterised in that：(C4) convolution method is normalized using structure adaptive when image reconstruction, to more
Frame low-resolution image is reconstructed, and includes the following steps,
(1) multiframe low-resolution image is analyzed according to noise robustness, estimates noise pixel；
(2) after obtaining noise pixel, the size for changing multiframe low-resolution image and the Gauss for normalizing convolution are filtered
Wave device direction obtains the sharp keen high-resolution reconstruction image of details.
10. the underwater works surface crack detection device according to claim 7 based on the bionical vision of compound eye is underwater
Structures surface crack detection method, it is characterised in that：Step (D) utilizes gray scale morphology gradient and adaptive gray threshold
Method is split extraction to doubtful crack area, and the high-definition picture for marking doubtful crack area is stored to DDR
In memory module, include the following steps,
(D1) Double Thresholding Segmentation is carried out to the high-definition picture of underwater works, obtains corresponding target area MhighAnd Mlow,
It is partitioned into underwater works main body from high-definition picture, removes background area, is further reduced aquatic organism and rubbish
The influence of noise fracture identification；
(D2) two targets are calculated into column hisgram back projection to the high-definition picture of the underwater works after Double Thresholding Segmentation
Region MhighAnd MlowGray value overlap ratio, when ratio is higher than threshold xi, then the gray scale corresponding pixel is image
Foreground area, as underwater works main body F, shown in calculation formula such as formula (2),
Wherein, Hhigh(r) and Hlow(r) it is respectively two target area MhighAnd MlowGrey level histogram, r is gray value, threshold value
ξ is to overlap ratio to gray value during the experiment to carry out observing obtained empirical value, takes 0.8；
(D3) it combines and is partitioned into underwater works main body F, according to doubting for the characteristic features automatic identification high-definition picture in crack
Like crack area, doubtful crack area is detected according to the low frequency strip connected domain in high-definition picture, detects doubtful split
After stitching region, with red minimum external oval marks；
(D4) high-definition picture of doubtful crack area will be marked to export, and carries out storage and arrives DDR memory modules.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610282173.6A CN105954292B (en) | 2016-04-29 | 2016-04-29 | Underwater works surface crack detection device based on the bionical vision of compound eye and method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610282173.6A CN105954292B (en) | 2016-04-29 | 2016-04-29 | Underwater works surface crack detection device based on the bionical vision of compound eye and method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN105954292A CN105954292A (en) | 2016-09-21 |
| CN105954292B true CN105954292B (en) | 2018-09-14 |
Family
ID=56913129
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201610282173.6A Active CN105954292B (en) | 2016-04-29 | 2016-04-29 | Underwater works surface crack detection device based on the bionical vision of compound eye and method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN105954292B (en) |
Families Citing this family (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107576665A (en) * | 2017-10-12 | 2018-01-12 | 上海遨拓深水装备技术开发有限公司 | A kind of underwater structure crack detection system based on machine vision |
| CN108537728A (en) * | 2018-03-05 | 2018-09-14 | 中国地质大学（武汉） | High spectrum image super-resolution forming method and system based on spectrum fidelity |
| WO2019214064A1 (en) * | 2018-05-11 | 2019-11-14 | 苏州源泽光电科技有限公司 | Fly-eye detection device |
| CN110118788A (en) * | 2019-04-30 | 2019-08-13 | 东南大学 | Submerged structure Defect inspection contact visual sensing equipment and image processing method |
| CN115346127B (en) * | 2022-10-20 | 2023-01-24 | 成都大汇物联科技有限公司 | Dam safety detection method and system |
Family Cites Families (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| DE3644354A1 (en) * | 1985-12-27 | 1987-07-09 | Canon Kk | OPTICAL PROJECTION SYSTEM |
| CN100373394C (en) * | 2005-10-28 | 2008-03-05 | 南京航空航天大学 | Petoscope based on bionic oculus and method thereof |
| JP2012185149A (en) * | 2011-02-17 | 2012-09-27 | Ricoh Co Ltd | Defect inspection device and defect inspection processing method |
| CN102507592B (en) * | 2011-11-01 | 2014-05-28 | 河海大学常州校区 | Fly-simulation visual online detection device and method for surface defects |
| CN103048330B (en) * | 2012-12-18 | 2015-12-23 | 河海大学常州校区 | The visual sniffer of a kind of surface crack of underwater structure |
| CN103051887A (en) * | 2013-01-23 | 2013-04-17 | 河海大学常州校区 | Eagle eye-imitated intelligent visual sensing node and work method thereof |
| CN103900696B (en) * | 2014-03-07 | 2015-09-30 | 河海大学 | The method of polarization imaging under water of a kind of imitative squill vision polarization antagonism perception |
| CN104948916A (en) * | 2015-05-15 | 2015-09-30 | 厦门大学 | Circularly-surrounding underwater fish-shaped robot underwater pipeline detecting device and method |
-
2016
- 2016-04-29 CN CN201610282173.6A patent/CN105954292B/en active Active
Also Published As
| Publication number | Publication date |
|---|---|
| CN105954292A (en) | 2016-09-21 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN105954292B (en) | Underwater works surface crack detection device based on the bionical vision of compound eye and method | |
| CN106952257B (en) | A kind of curved surface label open defect detection method based on template matching and similarity calculation | |
| CN105894499B (en) | A kind of space object three-dimensional information rapid detection method based on binocular vision | |
| CN106897698B (en) | Classroom people number detection method and system based on machine vision and binocular collaborative technology | |
| CN110211093A (en) | A kind of water outlet control method for automatic drinking water apparatus | |
| CN107340546A (en) | A kind of undersea detection divides the double CCD real-time polarizations imaging devices in aperture and method | |
| CN108564065A (en) | A kind of cable tunnel open fire recognition methods based on SSD | |
| CN110335245A (en) | Cage netting damage monitoring method and system based on monocular space and time continuous image | |
| CN110706197A (en) | Railway foreign matter intrusion detection method based on transfer learning in special scene | |
| CN109389630A (en) | Visible images and the determination of Infrared Image Features point set, method for registering and device | |
| CN109145734A (en) | Algorithm is captured in IPC Intelligent human-face identification based on 4K platform | |
| CN109688382A (en) | A kind of underwater picture processing system of underwater robot | |
| CN110060304A (en) | A kind of organism three-dimensional information acquisition method | |
| CN108921881A (en) | A kind of across camera method for tracking target based on homography constraint | |
| CN106228541A (en) | The method and device of screen location in vision-based detection | |
| Ji et al. | Real-time detection of underwater river crab based on multi-scale pyramid fusion image enhancement and MobileCenterNet model | |
| CN106530315B (en) | Target extraction system and method for medium and small objects under full angle | |
| CN105919543A (en) | Multispectral endoscope imaging image acquisition method and system | |
| CN111079733A (en) | Data augmentation method based on underwater imaging environment model | |
| CN110210292A (en) | A kind of target identification method based on deep learning | |
| CN207301366U (en) | A kind of undersea detection divides aperture double CCD real-time polarization imaging devices | |
| CN110348344A (en) | A method of the special facial expression recognition based on two and three dimensions fusion | |
| CN115457437A (en) | Crop identification method, device and system and pesticide spraying robot | |
| Chen et al. | Underwater image restoration and target detection based on monocular depth estimation | |
| Deng et al. | Underwater circular object positioning system based on monocular vision |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |