CN111368854A - Method for batch extraction of same-class target contour with single color in aerial image - Google Patents
Method for batch extraction of same-class target contour with single color in aerial image Download PDFInfo
- Publication number
- CN111368854A CN111368854A CN202010140878.0A CN202010140878A CN111368854A CN 111368854 A CN111368854 A CN 111368854A CN 202010140878 A CN202010140878 A CN 202010140878A CN 111368854 A CN111368854 A CN 111368854A
- Authority
- CN
- China
- Prior art keywords
- image
- contour
- color
- target
- aerial
- 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
- 238000000034 method Methods 0.000 title claims abstract description 36
- 238000000605 extraction Methods 0.000 title claims abstract description 23
- 238000004422 calculation algorithm Methods 0.000 claims abstract description 20
- 238000007781 pre-processing Methods 0.000 claims abstract description 8
- 230000002708 enhancing effect Effects 0.000 claims abstract description 4
- 239000003086 colorant Substances 0.000 claims abstract 3
- 238000013135 deep learning Methods 0.000 claims description 7
- 230000001965 increasing effect Effects 0.000 claims description 7
- 238000005260 corrosion Methods 0.000 claims description 5
- 230000007797 corrosion Effects 0.000 claims description 5
- 238000010801 machine learning Methods 0.000 claims description 4
- 230000000007 visual effect Effects 0.000 claims description 4
- 230000001131 transforming effect Effects 0.000 claims description 2
- 238000004364 calculation method Methods 0.000 abstract description 4
- 238000012545 processing Methods 0.000 description 12
- 238000012544 monitoring process Methods 0.000 description 6
- 238000009499 grossing Methods 0.000 description 4
- 239000002023 wood Substances 0.000 description 4
- 235000008331 Pinus X rigitaeda Nutrition 0.000 description 3
- 235000011613 Pinus brutia Nutrition 0.000 description 3
- 241000018646 Pinus brutia Species 0.000 description 3
- 238000006243 chemical reaction Methods 0.000 description 3
- 230000009466 transformation Effects 0.000 description 3
- 238000010586 diagram Methods 0.000 description 2
- 241000607479 Yersinia pestis Species 0.000 description 1
- 230000003044 adaptive effect Effects 0.000 description 1
- 230000007123 defense Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000018109 developmental process Effects 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 239000000284 extract Substances 0.000 description 1
- 238000005286 illumination Methods 0.000 description 1
- 229910052500 inorganic mineral Inorganic materials 0.000 description 1
- 238000011835 investigation Methods 0.000 description 1
- 238000007726 management method Methods 0.000 description 1
- 239000011707 mineral Substances 0.000 description 1
- 230000008569 process Effects 0.000 description 1
- 230000011218 segmentation Effects 0.000 description 1
- 238000007619 statistical method Methods 0.000 description 1
- 238000012546 transfer Methods 0.000 description 1
- 230000007704 transition Effects 0.000 description 1
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/40—Extraction of image or video features
- G06V10/44—Local feature extraction by analysis of parts of the pattern, e.g. by detecting edges, contours, loops, corners, strokes or intersections; Connectivity analysis, e.g. of connected components
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/40—Extraction of image or video features
- G06V10/56—Extraction of image or video features relating to colour
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/10—Terrestrial scenes
- G06V20/13—Satellite images
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Multimedia (AREA)
- Theoretical Computer Science (AREA)
- Astronomy & Astrophysics (AREA)
- Remote Sensing (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Image Analysis (AREA)
Abstract
The invention provides a method for extracting similar target contours with single colors in aerial images in batches, which solves the problem of low efficiency of extracting similar target contours which repeatedly appear in aerial photos in batches under the condition of similar target colors. The method comprises the following steps: (1) reading a small image of a target from an original image, and enhancing the image saturation difference; (2) converting the color space, generating a binary image, and carrying out contour extraction preprocessing on the binary image; (3) and extracting the object with the maximum outline in the binary image as the outline of the target. The invention has the following advantages: the method has strong universality; the algorithm is simple; the calculation efficiency is high; the result is concise.
Description
Technical Field
The invention belongs to the field of image processing, and relates to a method for extracting the contour of a same-purpose logo with a single color in aerial images in batches.
Background
The unmanned aerial vehicle aerial photography technology can be widely applied to the fields of national ecological environment protection, mineral resource exploration, marine environment monitoring, land utilization investigation, water resource development, crop growth monitoring and yield estimation, agricultural operation, natural disaster monitoring and assessment, urban planning and municipal management, forest pest protection and monitoring, public safety, national defense industry, digital earth, advertisement photography and the like. The unmanned aerial vehicle aerial image has the advantages of high definition, large scale, small area and high availability. And the pilotless plane provides a remote sensing platform which is convenient to operate and easy to transfer for aerial photography. The take-off and landing are less limited by the field, and the landing can be carried out on playgrounds, highways or other wider ground, so that the stability and the safety are good, and the transition is very easy. At present, the mode of monitoring objects based on aerial images of unmanned aerial vehicles is gradually replacing the mode of manual field operation.
The monitoring and identification of objects in the aerial images mainly depend on a deep learning algorithm, and some enterprises can also adopt a manual identification mode. In practical application, the contour of the identified target is often extracted, and the extracted contour can be used for realizing requirements of example segmentation of a three-dimensional object, clustering of the target from two dimensions to three dimensions, statistical analysis and the like. However, in the process of extracting the target contour, extracting the contour while identifying the target by using the deep learning algorithm greatly reduces the working efficiency, the consumed time is approximately increased to twice that of identifying the target, and the adoption of the manual contour delineation mode is not only inefficient, but also the quality of the obtained result is uneven.
Disclosure of Invention
In order to solve the problem of high time cost of obtaining the target contour from the aerial image, the invention designs a method for extracting the contour of the same kind of target with single color in the aerial image in batches, and realizes a batch processing mode for fully automatically extracting the contour of the same kind of target with single color. The method combines the recognition result of the target, calculates an enhanced image through RGB three channels, transforms the color space of the image into a visual color space, selects a binary image of a saturation channel from an adaptive threshold value and extracts the contour of the target. The method comprises the following specific steps:
1) reading a small image of a target from an original image, and enhancing the image saturation difference;
2) converting the color space, generating a binary image, and carrying out contour extraction preprocessing on the binary image;
3) and extracting the object with the maximum outline in the binary image as the outline of the target.
The method for extracting the similar target contour with single color in the aerial image depends on the target identification result, image enhancement is carried out according to the color characteristics of the target and the background, and the target contour is rapidly extracted by utilizing color space transformation and image binarization processing.
In the method for extracting the contour of the same target with a single color in the aerial image in batches, the position of the target in the aerial image is determined manually or by a deep learning algorithm or a machine learning algorithm.
In the method for extracting the contour of the same type of object with single color in the aerial image in batches, in the step (1), the difference of the color saturation of the object and the background in the image is increased according to the characteristic difference of the object color and the background color in the small image. Specifically, a formula for converting RGB three-channel numerical values is designed, so that three-channel numerical values of a single pixel point in a target area and a background area after conversion are close to each other, and at least one of three-channel numerical values of the other single pixel point is obviously larger than or smaller than other numerical values.
In the method for extracting the contour of the same type of target with a single color in the aerial image in batches, the image with the enhanced image saturation difference is subjected to smoothing processing. Further, the smoothing of the image is to smooth the image by using a two-dimensional gaussian filter.
In the above method for batch extraction of the contour of the same kind of object with single color in the aerial image, the image color space is converted into the visual color space in the step (2).
In the method for extracting the similar target contour with a single color in the aerial image in batches, the saturation channel image of the image after the color space conversion is converted into the threshold-adaptive binary image according to the average value of the saturation channel.
In the above method for batch extraction of the same-class object contour with a single color in an aerial image, the binary image contour extraction preprocessing step in step (2) includes: reverse color treatment, hole filling treatment, expansion treatment and corrosion treatment.
In the method for extracting the contour of the same type of target with a single color in the aerial image in batches, the ellipse is used for checking the image, and the image is subjected to expansion processing twice and then is subjected to corrosion processing once again.
In the method for extracting the contour of the same kind of target with a single color in the aerial image in batches, after the contour point of the object with the maximum contour is obtained in the step (3), the contour point is simplified by using the douglas-podcast algorithm, and the specific steps include:
setting an initial threshold value and a contour point constraint value, simplifying contour points by using a Douglas-Pock algorithm, if the number of the obtained contour points is still larger than the contour point constraint value, increasing the threshold value, and continuing to simplify the contour until the number of the obtained contour points is smaller than or equal to the contour point constraint value.
In the method for extracting the contour of the same target with a single color in the aerial image in batches, the position of the target in the aerial image is determined manually or by a deep learning algorithm or a machine learning algorithm.
The invention has the advantages that: the contour is extracted based on the small image where the target is located, the requirement on the image quality is reduced, and the accuracy of the result is improved; the used image enhancement method is simple, easy to realize and high in calculation efficiency; a threshold self-adaptive binarization method is adopted to improve the automation degree; an algorithm which can be copied in the same group of aerial photos is designed, and the image batch processing is realized.
Drawings
FIG. 1 is a flow chart of the present invention
FIG. 2 is an input image of an embodiment
FIG. 3 is a binarized image of an embodiment
FIG. 4 is a schematic diagram of a final profile of an embodiment
Detailed Description
The invention is further explained by taking the outline extraction of the dying or dead wood of the pine trees in the aerial image as an example with reference to the attached drawings:
a method for batch extraction of homogeneous object contours with single color in aerial images comprises the following steps:
1) reading a small image of a target from an original image, and enhancing the image saturation difference;
2) converting the color space, generating a binary image, and carrying out contour extraction preprocessing on the binary image;
3) and extracting the object with the maximum outline in the binary image as the outline of the target.
The method for extracting the similar target contour with single color in the aerial image depends on the target identification result, image enhancement is carried out according to the color characteristics of the target and the background, and the target contour is rapidly extracted by utilizing color space transformation and image binarization processing. Firstly, the positions of the small images of the target in the aerial image are determined manually or by a deep learning algorithm or a machine learning algorithm, and fig. 2 shows three small images read in the example, which are obtained from the recognition result of the deep learning algorithm on the infected pine trees with red, yellow or fallen leaves in the forest.
In the step (1), the difference of the color saturation of the target and the background in the image is increased according to the characteristic difference of the target color and the background color in the small image. Specifically, a formula for converting RGB three-channel numerical values is designed, so that three-channel numerical values of a single pixel point in a target area and a background area after conversion are close to each other, and at least one of three-channel numerical values of the other single pixel point is obviously larger than or smaller than other numerical values.
Preferably, the formula for transforming the RGB three-channel values can be designed as follows:
1. comparing the difference between the RGB three-channel values of the target and the background to find two channels with the largest numerical difference, wherein an R channel and a G channel are selected according to the color characteristics that healthy pine wood is green, and dying wood or withered wood is red, yellow or grey-white;
2. note that the channels found in step 1 are channel one and channel two, the remaining channels are channel three, and the three channels are transformed as follows
Wherein the value of α is selected artificially, and the values of all channels are constrained to be integers of 0-255, in this example, α is 1, and the transformation formula is
And smoothing the image with the enhanced image saturation difference. Preferably, the smoothing of the image is performed by using a two-dimensional gaussian filter.
Converting the color space in step (2) to generate a binary image refers to converting the RGB color space into a visual color space (including HSV color space, HSL color space, HSI color space, and HSB color space). And then converting the saturation channel image of the image after the color space is converted into a threshold-value-adaptive binary image according to the average value of the saturation channel (S channel).
Preferably, the average value of the S-channel values is calculated, and a fixed value (which is selected artificially according to the model of the photographing device and the illumination condition of the aerial photography operation) is added or subtracted to the average value as the threshold value for the image binarization.
The binary image contour extraction preprocessing step in the step (2) comprises the following steps: reverse color treatment, hole filling treatment, expansion treatment and corrosion treatment.
Further, the binary image contour extraction preprocessing specifically comprises the following operations:
acquiring all the contours in the binary image, filling the contours with the contour areas smaller than a fixed value, wherein the fixed value is selected within a range of 0.1-0.5 times of the area of the small image, performing image color reversal processing, and repeating the operation once; and performing expansion processing twice by using the ellipse to check the image, and performing corrosion processing once again. In this example, an elliptical kernel is used
Fig. 3 is a binary image obtained by processing the three small images in fig. 2 through the above steps.
After the contour point of the object with the maximum contour is obtained in the step (3), simplifying the contour point by using a Douglas-Puke algorithm, and the method specifically comprises the following steps:
setting an initial threshold value and a contour point constraint value, simplifying contour points by using a Douglas-Pock algorithm, if the number of the obtained contour points is still larger than the contour point constraint value, increasing the threshold value, and continuing to simplify the contour until the number of the obtained contour points is smaller than or equal to the contour point constraint value.
Fig. 4 is a schematic diagram of a simplified contour result after extracting the contour from the three binary images in fig. 3.
The technical conception of the invention is as follows: the accuracy is improved by combining the target identification result; fully utilizing the color characteristics of the target and the background to enhance the image; obtaining a threshold self-adaptive binary image according to the numerical characteristics of the saturation channel; calculating by using a binary image to improve the calculation efficiency; and the rapid simplification of the outline by using the Douglas-Puke algorithm is convenient for practical application of the result.
The method and the device have the advantages that the similar target contour with single color in the aerial images is automatically extracted in batches by the machine, labor is saved, calculation efficiency is improved, and time cost is reduced.
The embodiments described in this specification are merely illustrative of implementations of the inventive concept and the scope of the present invention should not be considered limited to the specific forms set forth in the embodiments but rather by the equivalents thereof as may occur to those skilled in the art upon consideration of the present inventive concept.
Claims (10)
1. A method for batch extraction of homogeneous object contours with single color in aerial images is characterized by comprising the following steps:
1) reading a small image of a target from an original image, and enhancing the image saturation difference;
2) converting the color space, generating a binary image, and carrying out contour extraction preprocessing on the binary image;
3) and extracting the object with the maximum outline in the binary image as the outline of the target.
2. The method for batch extraction of homogeneous object contour with single color in aerial image as claimed in claim 1, wherein in step (1), the difference in color saturation between object and background in image is increased according to the characteristic difference between object color and background color in small image.
3. The method for batch extraction of the similar target contour with single color in the aerial image as claimed in claim 2, wherein a formula for transforming RGB three-channel values is designed, so that in both the transformed target region and the background region, the three-channel values of one single pixel point are close, and at least one of the three-channel values of the other single pixel point is obviously greater than or less than the other values.
4. The method for batch extraction of homogeneous object contours with single color in aerial images as claimed in claim 1, wherein the images with enhanced image saturation difference are smoothed.
5. The method for batch extraction of color-single homogeneous object contours in aerial images according to claim 1, characterized in that in step (2) the image color space is transformed into a visual color space.
6. The method according to claim 5, wherein the saturation channel image of the image after converting the color space is converted into the threshold-adaptive binary image according to the average value of the saturation channel.
7. The method for batch extraction of homogeneous object contours with single color in aerial images according to claim 1, wherein the binary image contour extraction preprocessing step in step 2 comprises: reverse color treatment, hole filling treatment, expansion treatment and corrosion treatment.
8. The method for batch extraction of homogeneous object contours with single color in aerial images as claimed in claim 7, wherein the image is dilated twice by ellipse check and then eroded once again.
9. The method for batch extraction of the contour of the same kind of target with a single color in the aerial image as claimed in claim 1, wherein after the contour point of the object with the maximum contour is obtained in step 3, the contour point is simplified by using the douglas-podocar algorithm, and the specific steps include:
setting an initial threshold value and a contour point constraint value, simplifying contour points by using a Douglas-Pock algorithm, if the number of the obtained contour points is still larger than the contour point constraint value, increasing the threshold value, and continuing to simplify the contour until the number of the obtained contour points is smaller than or equal to the contour point constraint value.
10. The method for batch extraction of homogeneous object outlines with single colors in aerial images according to claim 1, wherein the small image positions of the objects in the aerial images are determined manually or by a deep learning algorithm or a machine learning algorithm.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010140878.0A CN111368854A (en) | 2020-03-03 | 2020-03-03 | Method for batch extraction of same-class target contour with single color in aerial image |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010140878.0A CN111368854A (en) | 2020-03-03 | 2020-03-03 | Method for batch extraction of same-class target contour with single color in aerial image |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN111368854A true CN111368854A (en) | 2020-07-03 |
Family
ID=71211660
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010140878.0A Pending CN111368854A (en) | 2020-03-03 | 2020-03-03 | Method for batch extraction of same-class target contour with single color in aerial image |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111368854A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112200846A (en) * | 2020-10-23 | 2021-01-08 | 东北林业大学 | Forest stand factor extraction method fusing unmanned aerial vehicle image and ground radar point cloud |
| CN112750162A (en) * | 2020-12-29 | 2021-05-04 | 北京电子工程总体研究所 | Target identification positioning method and device |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| ES2481347A1 (en) * | 2012-12-26 | 2014-07-29 | Universidad De Almeria | Method for the automatic interpretation of images for the quantification of nucelar tumour markers |
| CN104932683A (en) * | 2015-05-28 | 2015-09-23 | 重庆大学 | Game motion sensing control method based on vision information |
| CN105023014A (en) * | 2015-08-21 | 2015-11-04 | 马鞍山市安工大工业技术研究院有限公司 | Method for extracting tower target in unmanned aerial vehicle routing inspection power transmission line image |
| CN105825505A (en) * | 2016-03-14 | 2016-08-03 | 北京航空航天大学 | Vision measurement method facing boom air refueling |
| CN106780438A (en) * | 2016-11-11 | 2017-05-31 | 广东电网有限责任公司清远供电局 | Defects of insulator detection method and system based on image procossing |
| CN110287822A (en) * | 2019-06-10 | 2019-09-27 | 浙江大学城市学院 | The snail pest control method of view-based access control model target detection in a kind of officinal dendrobium stem plantation |
-
2020
- 2020-03-03 CN CN202010140878.0A patent/CN111368854A/en active Pending
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| ES2481347A1 (en) * | 2012-12-26 | 2014-07-29 | Universidad De Almeria | Method for the automatic interpretation of images for the quantification of nucelar tumour markers |
| CN104932683A (en) * | 2015-05-28 | 2015-09-23 | 重庆大学 | Game motion sensing control method based on vision information |
| CN105023014A (en) * | 2015-08-21 | 2015-11-04 | 马鞍山市安工大工业技术研究院有限公司 | Method for extracting tower target in unmanned aerial vehicle routing inspection power transmission line image |
| CN105825505A (en) * | 2016-03-14 | 2016-08-03 | 北京航空航天大学 | Vision measurement method facing boom air refueling |
| CN106780438A (en) * | 2016-11-11 | 2017-05-31 | 广东电网有限责任公司清远供电局 | Defects of insulator detection method and system based on image procossing |
| CN110287822A (en) * | 2019-06-10 | 2019-09-27 | 浙江大学城市学院 | The snail pest control method of view-based access control model target detection in a kind of officinal dendrobium stem plantation |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112200846A (en) * | 2020-10-23 | 2021-01-08 | 东北林业大学 | Forest stand factor extraction method fusing unmanned aerial vehicle image and ground radar point cloud |
| CN112750162A (en) * | 2020-12-29 | 2021-05-04 | 北京电子工程总体研究所 | Target identification positioning method and device |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108009542B (en) | Weed image segmentation method in rape field environment | |
| CN109558806B (en) | Method for detecting high-resolution remote sensing image change | |
| CN107862667B (en) | Urban shadow detection and removal method based on high-resolution remote sensing image | |
| CN109636784B (en) | Image saliency target detection method based on maximum neighborhood and super-pixel segmentation | |
| CN111027497B (en) | Weak and small target rapid detection method based on high-resolution optical remote sensing image | |
| CN112488046B (en) | Lane line extraction method based on high-resolution images of unmanned aerial vehicle | |
| CN111080696B (en) | Underwater sea cucumber identification and positioning method based on computer vision | |
| CN107992856B (en) | High-resolution remote sensing building shadow detection method under urban scene | |
| CN107358161B (en) | Coastline extraction method and coastline extraction system based on remote sensing image classification | |
| CN111368854A (en) | Method for batch extraction of same-class target contour with single color in aerial image | |
| CN115578660B (en) | Land block segmentation method based on remote sensing image | |
| CN105405138A (en) | Water surface target tracking method based on saliency detection | |
| CN101447027B (en) | Binaryzation method of magnetic code character area and application thereof | |
| CN111882555B (en) | Deep learning-based netting detection method, device, equipment and storage medium | |
| CN103927759A (en) | Automatic cloud detection method of aerial images | |
| CN104217440A (en) | Method for extracting built-up area from remote sensing image | |
| CN112070717A (en) | Power transmission line icing thickness detection method based on image processing | |
| US7620246B2 (en) | Method and apparatus for image processing | |
| JP2009123234A (en) | Object identification method, apparatus and program | |
| CN116958808A (en) | Forest parameter estimation method based on real-time target detection network | |
| CN103871084B (en) | Indigo printing fabric pattern recognition method | |
| JP2004078939A (en) | Object identification method, apparatus and program | |
| CN116129278A (en) | Land utilization classification and identification system based on remote sensing images | |
| Khan et al. | Shadow removal from digital images using multi-channel binarization and shadow matting | |
| CN115187790A (en) | Image contour extraction method based on reference region binarization result |
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 |
Application publication date: 20200703 |
|
| RJ01 | Rejection of invention patent application after publication |