CN111986111B - An Image Segmentation Method - Google Patents

Abstract

The invention discloses an image segmentation method, which comprises the following steps: s1: separating all pixel points of an image to be segmented, and inserting a pixel boundary point between every two adjacent pixel points; s2: scanning pixel boundary points according to a set scanning path in sequence; s3: when each pixel boundary point is scanned, respectively acquiring the pixel value of each pixel point around the pixel boundary point, and comparing the pixel values of two adjacent pixel points around the pixel boundary point; s4: acquiring pixel boundary points around the pixel boundary point, and scanning the pixel boundary points around the pixel boundary point; s5: and connecting all the left coordinates into a line, and segmenting the image according to the line. The invention obtains a point which can be divided by scanning, scans the adjacent periphery of the dividing point to obtain the adjacent dividing point, and finally connects all the dividing points to obtain the image to be divided.

Description

A Method of Image Segmentation

technical field

The invention relates to the field of image processing, in particular to an image segmentation method.

Background technique

Image segmentation is the technology and process of identifying specific and unique regions and proposing objects of interest. It is a key step from image processing to image analysis. In the current image segmentation, whether to segment one image or multiple images, the image to be segmented will be scanned as a whole, and finally the image to be segmented can be obtained, so that each time it is segmented, a large number of The operation makes the image segmentation time longer.

Contents of the invention

The purpose of the present invention is to overcome the problems existing in the above-mentioned prior art, and provide a kind of image segmentation method, obtain a point that can be segmented by scanning, and scan the adjacent surroundings of the segmented point, so that a corresponding image can be obtained. The adjacent segmentation points, and finally connect all the segmentation points to get the image to be segmented.

For this reason, the invention provides a kind of image segmentation method, comprises the steps:

S1: Separate all pixels of the image to be segmented, and insert a pixel boundary point between every two adjacent pixels, so that each pixel boundary point is surrounded by pixels of the image, and obtain all Coordinates of pixel breakpoints.

S2: Scanning pixel boundary points sequentially according to the set scanning path.

S3: When scanning each pixel boundary point, obtain the pixel value of each pixel point around the pixel boundary point, and compare the pixel values of two adjacent pixel points around the pixel boundary point, when adjacent When the difference between the pixel values of two pixel points is larger than the set value, the coordinates of the boundary point of the pixel are saved and step S4 is entered; otherwise, step S2 is entered.

S4: Obtain the pixel boundary points around the pixel boundary point, scan the pixel boundary points around the pixel boundary point, and execute step S3 until the coordinates of the pixel boundary points that are obtained again are the coordinates of the pixel boundary point.

S5: Connect all the retained coordinates into a line, and segment the image according to the line, and finally remove all pixel boundary points in the image.

Further, in step S2, the initial scanning position is the position where any pixel boundary point is located.

Furthermore, the set path is a spiral path.

Further, in step S1, when obtaining the coordinates of the pixel boundary points, coordinates are uniformly assigned to all pixel boundary points. In the coordinates of two adjacent pixel boundary points, the absolute difference between the difference in abscissa or the difference in ordinate The value is 1.

Further, before step S1, denoising processing is performed on the image to be processed.

An image segmentation method provided by the present invention has the following beneficial effects: a point that can be segmented is obtained by scanning, and by scanning around the adjacent segment of the segmented point, the segmented point adjacent to it is obtained, and finally the Connect all the segmentation points to get the image to be segmented.

Description of drawings

Fig. 1 is a schematic block diagram of the overall process of the present invention.

Detailed ways

A specific embodiment of the present invention will be described in detail below in conjunction with the accompanying drawings, but it should be understood that the protection scope of the present invention is not limited by the specific embodiment.

In this application document, unspecified component models and structures are all prior art known to those skilled in the art, and those skilled in the art can set them according to the needs of the actual situation. In the embodiments of this application document No specific restrictions are made.

Specifically, as shown in FIG. 1, an embodiment of the present invention provides an image segmentation method, including the following steps:

S1: Separate all pixels of the image to be segmented, and insert a pixel boundary point between every two adjacent pixels, so that each pixel boundary point is surrounded by pixels of the image, and obtain all Coordinates of pixel breakpoints.

In this step, a pixel boundary point is inserted between two adjacent pixel points using a point insertion algorithm. After traversing each pixel point of the image, a new image can be obtained. In the new image, no two There is a pixel boundary point between adjacent pixel points, so that each pixel point is surrounded by adjacent pixel boundary points, each pixel point is adjacent to a pixel boundary point, and each pixel boundary point is adjacent a pixel.

S2: Scanning pixel boundary points sequentially according to the set scanning path.

In this step, the scanning path may use an S-shaped path or a spiral path. In short, the path needs to traverse to each pixel boundary point. The best way to traverse the pixel division points is to use the S-type path, because the S-type path can traverse each pixel boundary point, and each pixel boundary point is scanned only once.

S3: When scanning each pixel boundary point, obtain the pixel value of each pixel point around the pixel boundary point, and compare the pixel values of two adjacent pixel points around the pixel boundary point, when adjacent When the difference between the pixel values of two pixel points is greater than the set value, the coordinates of the boundary point of the pixel are retained and enter step S4, otherwise, enter step S2.

In this step, judge the value of the pixel points around the pixel boundary point, and judge whether the pixel point boundary point is the image dividing line where the image can be divided according to the difference between the pixel values of the surrounding pixel points. When two adjacent When the difference between the pixel values of the pixel is greater than the set value, it is considered that the boundary point of the pixel is the image dividing line where the image can be divided. If not, proceed to step S2 to scan the next pixel boundary point.

S4: Obtain the pixel boundary points around the pixel boundary point, scan the pixel boundary points around the pixel boundary point, and execute step S3 until the coordinates of the pixel boundary points that are obtained again are the coordinates of the pixel boundary point.

In this step, when a pixel boundary point has been obtained as a qualified pixel boundary point, the adjacent pixel boundary points around it are scanned, so that along one of the pixel boundary points that meets the requirements, quickly obtain Other pixel demarcation points that meet the requirements.

S5: Connect all the retained coordinates into a line, and segment the image according to the line, and finally remove all pixel boundary points in the image.

In this step, the coordinates of all the retained pixel boundary points are connected into a line, which is the image segmentation line, and the segmented image can be obtained by dividing according to the segmentation line. Finally, all the pixel boundary points in the image are all The image after the original image segmentation is obtained.

In this embodiment, in step S2, the initial scanning position is the position where any pixel boundary point is located. This solution is suitable for a spiral path, which can greatly reduce the number of operations, and for an S-shaped path, it is recommended to start scanning from the corner vertex of the image.

Meanwhile, in this embodiment, the set path is a spiral path. When the spiral path is used, it is possible to prevent repeated traversal of the pixel boundary points that have been executed in step S4 when performing step S4, which can save calculation steps during image segmentation processing, thereby improving image segmentation s efficiency.

In this embodiment, in step S1, when obtaining the coordinates of pixel boundary points, coordinates are uniformly assigned to all pixel boundary points, and in the coordinates of two adjacent pixel boundary points, the difference in abscissa or ordinate The absolute value of the difference is 1. In this way, there is no gap or jump in the coordinates between the boundary points of two adjacent pixels. When the coordinates are summarized in the later stage, the obtained image segmentation line will be smoother, and the function of the obtained image segmentation line will be more accurate. precise.

In this embodiment, before step S1, denoising processing is performed on the image to be processed. In the image after denoising processing, the pixel values between adjacent pixel points are smoother, and when the pixel values are compared, the error generated will be smaller.

The above disclosures are only a few specific embodiments of the present invention, however, the embodiments of the present invention are not limited thereto, and any changes conceivable by those skilled in the art shall fall within the protection scope of the present invention.

Claims (4)

1. An image segmentation method, characterized by comprising the steps of:

s1: separating all pixel points of an image to be segmented, inserting a pixel boundary point between every two adjacent pixel points, enabling the periphery of each pixel boundary point to be the pixel point of the image, and acquiring the coordinates of all the pixel boundary points; when the coordinates of the pixel boundary points are obtained, coordinates are uniformly distributed to all the pixel boundary points, and in the coordinates of two adjacent pixel boundary points, the absolute value of the difference of the abscissa or the difference of the ordinate is 1;

s2: scanning pixel boundary points according to a set scanning path in sequence;

s3: when each pixel boundary point is scanned, respectively obtaining the pixel value of each pixel point around the pixel boundary point, comparing the pixel values of two adjacent pixel points around the pixel boundary point, when the difference of the pixel values of two adjacent pixel points is greater than a set value, reserving the coordinate of the pixel boundary point and entering step S4, otherwise, entering step S2;

s4: acquiring a pixel boundary point around the pixel boundary point, scanning the pixel boundary point around the pixel boundary point, and executing the step S3 until the coordinate of the pixel boundary point reserved again is the coordinate of the pixel boundary point;

s5: and connecting all the reserved coordinates into a line, segmenting the image according to the line, and finally removing all pixel boundary points in the image.

2. An image segmentation method as claimed in claim 1, characterized in that in step S2, the initial scanning position is the position of any one pixel boundary point.

3. An image segmentation method as claimed in claim 2, characterized in that the set path is a spiral path.

4. An image segmentation method as claimed in claim 1, characterized in that, before step S1, the image to be processed is denoised.

Images