CN104243776A - Image processing method - Google Patents
Image processing method Download PDFInfo
- Publication number
- CN104243776A CN104243776A CN201410205894.8A CN201410205894A CN104243776A CN 104243776 A CN104243776 A CN 104243776A CN 201410205894 A CN201410205894 A CN 201410205894A CN 104243776 A CN104243776 A CN 104243776A
- Authority
- CN
- China
- Prior art keywords
- image
- area
- interest
- image processing
- identification
- 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
Landscapes
- Image Processing (AREA)
- Processing Or Creating Images (AREA)
Abstract
The invention provides an image processing method applied to an image processing device, the image processing method includes following steps: receiving an original wide-angle image, pre-processing the original wide-angle image and capturing at least a region of interesting (ROI); executing anti-distorting processing on the ROI to generate a local correction image; and executing image processing on the local correction image. In the image processing method, parts of ROI of the original wide-angle image is captured, anti-distorting processing and image processing are executed, which significantly improves the image processing efficiency and reduces the time consumption instead of executing anti-distorting processing on the ROI directly.
Description
Technical field
The invention relates to a kind of image processing technique, and relate to a kind of wide-angle image processing method especially.
Background technology
The visual angle of general video camera between 60 degree to 90 degree, the aerial image Limited information of acquisition.In comparison, bugeye lens such as fish eye lens then can obtain the wide-angle image containing the significantly visual field.But because the visual angle of bugeye lens extensively can be multiplied by 180 degree to 360 degree, the wide-angle image captured is generally the form of torsional deformation.Therefore, after acquisition wide-angle image, as carrying out further image processing, often first panorama correction must be carried out.But, current wide-angle image treatment mechanism needs to expend sizable operand in the process corrected, in addition, and the algorithm that follow-up image processing is normally complicated, therefore carry out image processing computing in full-view image after calibration, efficiency cannot reach gratifying performance.
Summary of the invention
The invention provides a kind of image treatment method, be applied to image processor.Image treatment method comprises: receive original wide-angle image by pre-treatment module, and carries out pre-treatment to capture at least one area-of-interest; By adjustment of image module, anti-twist process is carried out to area-of-interest, to produce partial correction image; And by image processing module, image processing is carried out to described partial correction image.
Apply image treatment method of the present invention, only need to capture the area-of-interest of part in original wide-angle image, carry out anti-twist process and image processing, and do not need directly to carry out anti-twist process to the entirety of original wide-angle image, significantly promote the efficiency of image processing, and reduce the time cost consumed.
Accompanying drawing explanation
Figure 1 shows that the calcspar of the image processor according to one embodiment of the invention;
Figure 2 shows that the flow chart of the image treatment method according to one embodiment of the invention.
Embodiment
Please refer to Fig. 1, Figure 1 shows that the calcspar of the image processor according to one embodiment of the invention.Image processor comprises: pre-treatment module 10, adjustment of image module 12, image processing module 14 and database 16.
Pre-treatment module 10 receives original wide-angle image 101, and carries out pre-treatment to capture at least one area-of-interest (region of interest; ROI) 13.Original wide-angle image 101 for utilizing bugeye lens, such as but not limited to fish eye lens, acquired image information.In an embodiment, this original wide-angle image 101 can comprise the image information that 360 degree are multiplied by 180 degree.Therefore due to visual angle relation, original wide-angle image 101 is by the image for torsional deformation.
In an embodiment, pre-treatment module 10 searches the multiple characteristic points in original wide-angle image 101, is greater than at least one region of critical value as area-of-interest 103 using the density capturing characteristic point.Wherein, on the characteristic point border that is arranged in original wide-angle image 101 or texture.
In an embodiment, characteristic point can utilize the technology such as, but not limited to edge detection (edge detection), and search in original wide-angle image 101, color or GTG have the pixel of notable difference as characteristic point.In other embodiments, also can carry out the search of characteristic point by other technologies.
Therefore, when pre-treatment module 10 judges in original wide-angle image 101, when the density of the characteristic point in a region is greater than the critical value of acquiescence, described region will be captured as area-of-interest 103.In different embodiment, all qualified regions of pre-treatment module 10 fechtable are as area-of-interest 103, or only in qualified region, the maximum or area the maximum of acquisition characteristic point density person is as area-of-interest 103 and carry out follow-up process.
In another embodiment, the color in the original wide-angle image 101 of pre-treatment module 10 identification or grey decision-making, to capture color or close at least one region of grey decision-making as area-of-interest 103.In different embodiments, all qualified regions of pre-treatment module 10 fechtable are as area-of-interest 103, or only in qualified region, acquisition area the maximum is as area-of-interest 103 and carry out follow-up process.
In an embodiment again, at least one moving area in the original wide-angle image 101 of pre-treatment module 10 identification is as area-of-interest 103.In an embodiment, pre-treatment module 10 need compare the image of this original wide-angle image 101 and previous time point, utilize the technology such as but not limited to mobile detection (motion detection), the moving area in the original wide-angle image 101 of identification.In different embodiments, all qualified regions of pre-treatment module 10 fechtable are as area-of-interest 103, or only in qualified region, acquisition area the maximum is as area-of-interest 103 and carry out follow-up process.
In different embodiments, pre-treatment module 10 can carry out the search of characteristic point, color or the identification of grey decision-making and the identification of moving area at least both, to capture dissimilar area-of-interest 103.Further, in an embodiment, pre-treatment module 10 can carry out such as but not limited to erosion (erosion) to area-of-interest 103 or expand (dilation), reach the object removing noise.In other embodiments, pre-treatment module 10 also on demand, can carry out other pre-treatment program to area-of-interest 103.
After acquisition area-of-interest 103, adjustment of image module 12 pairs of area-of-interests 103 carry out anti-twist (unwrapping) process, to produce partial correction image 105.In an embodiment, adjustment of image module 12 can according to the position of area-of-interest 103 at original wide-angle image 101, anti-twist process is carried out, the area-of-interest 103 of script distortion to be expanded into the image of plane such as but not limited to the angle of relative centre or side and distance.
Image processing module 14, further according to partial correction image 105, is compared with the data 107 in database 16, to carry out scenery identification, personage's identification, process identification or its combination, and produces identification result 109.
For example, when pre-treatment module 10 captures the region with scenario objects such as building according to characteristic point, and after carrying out anti-twist process by adjustment of image module 12, can with stored contextual data (not drawing in the drawings) comparison in database 16, why to judge this building, and after confirming as specific buildings, judge further the scene of this original wide-angle image 101 specific buildings periphery for this reason.
In an embodiment, scenery identification by such as but not limited to carrying out normalization, feature extraction (feature extraction) to different brightness, to carry out according to database that feature troops that (clustering) matches with ballot (voting), descriptor (descriptor), geometric verification or its technology combined reach.In other embodiments, other technology also may be adopted to reach the object of scenery identification.
Similarly, the region of the tool colour of skin in pre-treatment module 10 is with the original wide-angle image 101 of technology identification of such as color filter device (skin filter), and after carrying out anti-twist process by adjustment of image module 12, image processing module 14 can with stored human face data (not drawing in the drawings) comparison in database 16, to judge whether this region is the personage that face and face are corresponding, reaches effect of personage's identification.
And when the moving area in the original wide-angle image 101 of pre-treatment module 10 identification, and after carrying out anti-twist process by adjustment of image module 12, image processing module 14 can with stored personage in database 16 or human face data (not drawing in the drawings) comparison, to judge that whether this region is personage or the face of humanoid and humanoid correspondence, reach effect of personage's identification.
Similarly, process identification also can carry out via above-mentioned mode, therefore repeats no more.
As previously mentioned, in some embodiments, pre-treatment module 10 fechtable is all meets the region of the condition of characteristic point density, color or grey decision-making or moving area as area-of-interest 103, and is processed by adjustment of image module 12 and image processing module 14.
And in section Example, pre-treatment module 10 is only in the region of condition meeting above-mentioned arbitrary or its combination, acquisition characteristic point density the maximum or area the maximum carry out follow-up process as area-of-interest 103.When image processing module 14 and this region of unsuccessful identification time, pre-treatment module 10 can in qualified region, acquisition characteristic point density or size is next cis-position person, as area-of-interest 103, and being processed by adjustment of image module 12 and image processing module 14, until successful identification or all qualified regions all process.
In an embodiment, image processing module 14 can pick out conform to scene, personage or object time, return the data number that it is corresponding in database 16, to confirm identification success.
And when pre-treatment module 10 cannot capture any area-of-interest 103, such as do not have obvious target edges or area of skin color can supply to capture, adjustment of image module 12 directly can carry out anti-twist process to original wide-angle image 101, to produce panorama correction image 111, carry out image processing by image processing module 14 pairs of panorama correction images 111.
Wide-angle image processing unit 1 of the present invention only can capture the area-of-interest 103 of part in original wide-angle image 101, carries out anti-twist process and image processing, and does not need directly to carry out anti-twist process to the entirety of original wide-angle image 101.Therefore, wide-angle image processing unit 1 of the present invention significantly can promote the efficiency of image processing, and reduces the time cost consumed.
Please refer to Fig. 2.Figure 2 shows that the flow chart of the image treatment method according to one embodiment of the invention.Image treatment method 200 can be applicable to image processor 1 as shown in Figure 1.Image treatment method 200 comprises the following step (should be appreciated that, step mentioned in the present embodiment, except chatting its order person bright especially, all can adjust its tandem according to actual needs, even can perform simultaneously or partly simultaneously).
In step 201, receive original wide-angle image 101 by pre-treatment module 10, and carry out pre-treatment to capture area-of-interest 103.
In step 202, pre-treatment module 10 judges whether that acquisition is to area-of-interest 103.
When pre-treatment module 10 captures area-of-interest 103, perform step 203, carry out anti-twist process, to produce partial correction image 105 by adjustment of image module 12 pairs of area-of-interests 103.
Then, flow process carries out scene or process identification by image processing module 14 to local correcting image 105 in step 204, or carries out personage or process identification by image processing module 14 to local correcting image 105 in step 205.
And ought be in step 202., when pre-treatment module 10 judges not capture area-of-interest 103, adjustment of image module 12 carries out anti-twist process in step 206 to original wide-angle image 101, to produce panorama correction image 111, and then carry out scene or process identification or personage or process identification in step 204 or 205 by image processing module 14 pairs of panorama correction images 111.
Although the present invention discloses as above with preferred embodiment; so itself and be not used to limit the present invention; have in any art and usually know the knowledgeable; without departing from the spirit and scope of the present invention; when doing a little change and retouching, therefore protection scope of the present invention is when being as the criterion depending on those as defined in claim.
Claims (7)
1. an image treatment method, is applied to image processor, it is characterized in that, described image treatment method comprises:
Receive original wide-angle image, carry out pre-treatment and capture at least one area-of-interest;
Anti-twist process is carried out to described area-of-interest, to produce partial correction image; And
Image processing is carried out to described partial correction image.
2. image treatment method according to claim 1, is characterized in that, wherein said pre-treatment also comprises:
Search multiple characteristic points of described original wide-angle image; And
The density of acquisition these characteristic points described is greater than at least one region of critical value as described area-of-interest.
3. image treatment method according to claim 2, is characterized in that, at least one border that these characteristic points wherein said are arranged in described original wide-angle image or at least one texture.
4. image treatment method according to claim 1, is characterized in that, wherein said pre-treatment also comprises:
Color in original wide-angle image described in identification or grey decision-making; And
Capture described color or close at least one region of described grey decision-making as described area-of-interest.
5. image treatment method according to claim 1, is characterized in that, wherein said pre-treatment also comprises:
At least one moving area in original wide-angle image described in identification is as described area-of-interest.
6. image treatment method according to claim 1, is characterized in that, wherein said image processing also comprises:
Compare according to described partial correction image and database, to carry out scenery identification, personage's identification, process identification or its combination.
7. image treatment method according to claim 1, is characterized in that, also comprises:
When judging to capture described area-of-interest, described anti-twist process is carried out to described original wide-angle image, to produce panorama correction image; And
Described image processing is carried out to described panorama correction image.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| US14/307,484 US20140369625A1 (en) | 2013-06-18 | 2014-06-17 | Image processing method |
Applications Claiming Priority (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| US201361836649P | 2013-06-18 | 2013-06-18 | |
| US61/836,649 | 2013-06-18 |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN104243776A true CN104243776A (en) | 2014-12-24 |
Family
ID=52231056
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201410205894.8A Pending CN104243776A (en) | 2013-06-18 | 2014-05-15 | Image processing method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN104243776A (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109903227A (en) * | 2019-02-21 | 2019-06-18 | 武汉大学 | Full-view image joining method based on camera geometry site |
| CN112153251A (en) * | 2019-06-27 | 2020-12-29 | 多方科技(广州)有限公司 | Image processing device and method |
| CN113674138A (en) * | 2020-05-14 | 2021-11-19 | 杭州海康威视数字技术股份有限公司 | Image processing method, device and system |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US5904725A (en) * | 1995-04-25 | 1999-05-18 | Matsushita Electric Industrial Co., Ltd. | Local positioning apparatus |
| US20100002071A1 (en) * | 2004-04-30 | 2010-01-07 | Grandeye Ltd. | Multiple View and Multiple Object Processing in Wide-Angle Video Camera |
| CN102096898A (en) * | 2009-11-17 | 2011-06-15 | 三星电子株式会社 | Method and apparatus for image processing |
-
2014
- 2014-05-15 CN CN201410205894.8A patent/CN104243776A/en active Pending
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US5904725A (en) * | 1995-04-25 | 1999-05-18 | Matsushita Electric Industrial Co., Ltd. | Local positioning apparatus |
| US20100002071A1 (en) * | 2004-04-30 | 2010-01-07 | Grandeye Ltd. | Multiple View and Multiple Object Processing in Wide-Angle Video Camera |
| CN102096898A (en) * | 2009-11-17 | 2011-06-15 | 三星电子株式会社 | Method and apparatus for image processing |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109903227A (en) * | 2019-02-21 | 2019-06-18 | 武汉大学 | Full-view image joining method based on camera geometry site |
| CN112153251A (en) * | 2019-06-27 | 2020-12-29 | 多方科技(广州)有限公司 | Image processing device and method |
| CN113674138A (en) * | 2020-05-14 | 2021-11-19 | 杭州海康威视数字技术股份有限公司 | Image processing method, device and system |
| CN113674138B (en) * | 2020-05-14 | 2023-09-01 | 杭州海康威视数字技术股份有限公司 | Image processing method, device and system |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| EP3740897B1 (en) | License plate reader using optical character recognition on plural detected regions | |
| EP3327668B1 (en) | Image-processing device and method for foreground mask correction for object segmentation | |
| US20230260321A1 (en) | System And Method For Scalable Cloud-Robotics Based Face Recognition And Face Analysis | |
| KR101858646B1 (en) | Apparatus and method for fusion of image | |
| US8913792B2 (en) | Method of determining reference features for use in an optical object initialization tracking process and object initialization tracking method | |
| CN107908195B (en) | Target tracking method, target tracking device, tracker and computer-readable storage medium | |
| US20200143549A1 (en) | Image processing method and device, and unmanned aerial vehicle | |
| US11042725B2 (en) | Method for selecting frames used in face processing | |
| CN111160202B (en) | Identity verification method, device, equipment and storage medium based on AR equipment | |
| US20180176461A1 (en) | System and Method for Intelligent Camera Control | |
| US9344629B2 (en) | Method of capturing images and image capturing device using the method | |
| EP3481052B1 (en) | Electronic device for processing image based on priority and method for operating thereof | |
| US20140037212A1 (en) | Image processing method and device | |
| CN112272832A (en) | Method and system for DNN-based imaging | |
| CN111079613B (en) | Gesture recognition method and device, electronic equipment and storage medium | |
| US9466095B2 (en) | Image stabilizing method and apparatus | |
| US20110074965A1 (en) | Video processing system and method | |
| CN104243776A (en) | Image processing method | |
| EP4266250A1 (en) | Image processing method and chip, and electronic device | |
| CN114913338A (en) | Segmentation model training method and device, and image recognition method and device | |
| CN111666869B (en) | Face recognition method and device based on wide dynamic processing and electronic equipment | |
| CN101859370A (en) | Imaging system and imaging method thereof | |
| WO2014165159A1 (en) | System and method for blind image deconvolution | |
| US20140369625A1 (en) | Image processing method | |
| CN112529928A (en) | Part assembly detection method, computer device and storage medium |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| WD01 | Invention patent application deemed withdrawn after publication | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20141224 |