CN115588018A - Image processing method and device, electronic equipment and storage medium - Google Patents

Abstract

The present application provides an image processing method, device, electronic equipment, and storage medium. The method includes: obtaining the original pixel area, cutting line pixel value, and overlapping area width of the image to be processed; according to the original pixel value, cutting line pixel value, and The width of the overlapping area is used to generate each segmented image pixel area of the image to be processed; and the image to be processed is segmented according to each segmented image pixel area to obtain a segmented image. Each segmented image pixel area of the image to be processed corresponds to a segmented image, and the image to be processed is segmented according to the pixel area of each segmented image to obtain multiple segmented images corresponding to the image to be processed. The segmented images of the to-be-processed image have a certain overlap area width, so as to prevent the objects in the to-be-processed image from being segmented, thereby avoiding false detection and missed detection, and improving the effect of image processing.

Description

Image processing method, device, electronic device and storage medium

technical field

The present application relates to the technical field of image processing, and in particular, to an image processing method, device, electronic equipment, and storage medium.

Background technique

The application field of image processing using the deep neural network model involves all aspects of human life and work. Usually, the image to be processed is input into the model, and the model is used for feature extraction, analysis, detection, etc., and the processing result is obtained. However, if the captured clear image is directly input into the detection model, more information will be lost in the image, which will affect the detection result. The current solution is usually to crop the original image and divide it into small images for input to the network, but the current processing method is prone to false detection and missed detection.

Contents of the invention

The purpose of the embodiment of the present invention is an image processing method, device, electronic equipment and storage medium, by generating each segmented image pixel area of the image to be processed according to the original pixel value, the cutting line pixel value and the width of the overlapping area, each segment The pixel area of the segmented image corresponds to one segmented image, and the image to be processed is segmented according to the pixel area of each segmented image to obtain a plurality of segmented images corresponding to the image to be processed. The segmented images of the to-be-processed image have a certain width of the overlapping area, and the target in the to-be-processed image is prevented from being segmented, thereby avoiding false detection and missed detection, and improving the effect of image processing.

In the first aspect, the embodiment of the present application provides an image processing method, including: obtaining the original pixel area of the image to be processed, the pixel value of the cutting line, and the width of the overlapping area; according to the original pixel value, the pixel value of the cutting line, and the width of the overlapping area, generating each segmented image pixel area of the image to be processed; and segmenting the image to be processed according to each segmented image pixel area to obtain a segmented image.

In the above implementation process, each segmented image pixel area of the image to be processed is generated according to the original pixel value, the cutting line pixel value and the width of the overlapping area, and there is a certain overlap between the segmented image pixel areas on both sides of the cutting line The area width is used to segment the image to be processed according to the pixel area of the segmented image, which can prevent the target in the image to be processed from being segmented and improve the accuracy of image processing.

Optionally, in the embodiment of the present application, obtaining the original pixel area of the image to be processed, the pixel value of the cutting line, and the width of the overlapping area includes: generating the pixel value of the cutting line according to the original pixel area and the subsequent processing pixel area; wherein, the subsequent processing The pixel area represents the pixel area that needs to be conformed to in the subsequent processing of the image to be processed; input the image to be processed into the preset key point detection model to obtain the edge key points of the target to be detected in the image to be processed; according to the pixel value of the cutting line and Edge key points of the target to be detected to obtain the width of the overlapping area.

In the above implementation process, the cutting line pixel value is generated according to the original pixel area and the subsequent processing pixel area, that is, the cutting line pixel value is generated according to the pixel area size relationship between the original pixel area and the subsequent processing pixel area, and the subsequent processing pixel area is Process the pixel area that the image needs to conform to during subsequent processing. After the pixel value of the cutting line is determined, the width of the overlapping area is obtained according to the pixel value of the cutting line and the edge key points of the target to be detected. Make the segmented image on both sides of the pixel value along the cutting line have an overlapping area width, so as to prevent the target located on the cutting line from being segmented, resulting in information loss during subsequent processing.

Optionally, in the embodiment of the present application, after segmenting the image to be processed according to each segmented image pixel area and obtaining the segmented image, the method further includes: obtaining coordinate information of the segmented image; The information is converted to obtain restored coordinate information of the segmented image; wherein, the restored coordinate information represents the coordinate information in the image to be processed obtained by splicing multiple segmented images.

In the above implementation process, after the image to be processed is segmented, the coordinate information of the segmented image is obtained, and the information of the segmented image is converted into the corresponding overlapping stitching of the segmented image to obtain the restored coordinate information in the image to be processed . Through coordinate transformation, the coordinates of each segmented image correspond to the restored coordinates in the image to be processed, and the processing of the image to be processed is completed.

Optionally, in the embodiment of the present application, converting the coordinate information of the segmented image to obtain the restored coordinate information of the segmented image includes: sequentially inputting multiple segmented images into the preset according to the index numbers of the segmented images The detection model is used to obtain the detection result; wherein, the detection result includes the coordinate information of each segmented image, and the index number is obtained according to the positional relationship of the segmented image in the image to be processed; through the coordinate conversion formula, the coordinate information of the segmented image is carried out Conversion, to obtain the restoration coordinate information of the segmented image; the coordinate conversion formula includes: Y=y-hi+(H/N)*i; wherein, Y is the coordinate value of the segmentation direction in the restoration coordinate information, and y is the cutting direction in the coordinate information The coordinate value of the split direction, hi is the width of the overlapping area corresponding to the split image with the index number i, H is the pixel value of the image to be processed in the split direction, N is the number of splits of the image to be processed; i is the split The index number of the image, i∈[0,N-1].

In the above implementation process, the coordinate information of the segmented image is converted through the coordinate transformation formula, the restored coordinate information of the segmented image is obtained, and each segmented image is uniformly converted to facilitate subsequent processing.

Optionally, in the embodiment of the present application, wherein, the original pixel area includes pixel values of the image to be processed in the segmentation direction; the segmented image includes the first segmented image and the second segmented image; according to the original pixel value, The pixel value of the cutting line and the width of the overlapping area are used to generate each segmented image pixel area of the image to be processed, including: according to the pixel value of the cutting line and the width of the overlapping area, each segmented image pixel area of the image to be processed is generated by a division formula; the division formula Including: the pixel area of the first segmented image is: [0, (H/2)+hi]; the pixel area of the second segmented image is: [(H/2)-hi, H]; wherein, H is The pixel value of the image to be processed in the segmentation direction; H/2 is the pixel value of the cutting line; h _i is the width of the overlapping area corresponding to the segmented image.

In the above implementation process, each segmented image pixel area of the image to be processed is generated by the division formula, each segmented image pixel area corresponds to a segmented image, and the pixel area of the segmented image is divided to avoid processing according to the cutting line The image is segmented such that the target is segmented.

Optionally, in the embodiment of the present application, the method further includes: separately obtaining the edge key points of the object to be detected corresponding to both sides of the pixel value of the cutting line; The key point of the edge of the target, and the width of the overlapping area corresponding to each segmented image on both sides of the pixel value of the cutting line is obtained.

In the above implementation process, according to the pixel value of the cutting line and the edge key points of the target to be detected corresponding to the pixel value of the cutting line, the width of the overlapping area corresponding to each segmented image on both sides of the pixel value of the cutting line can be determined respectively , to flexibly determine the overlapping area width of each segmented image, so that each segmented image has an accurate overlapping area width.

Optionally, in the embodiment of the present application, wherein the image to be processed includes a gear tooth surface image; the width of the overlapping area includes a first overlapping area width and a second overlapping area width; the target to be detected includes a gear; the edge key point includes a gear tooth The vertex of each tooth in the surface and the bottom point of each tooth; according to the pixel value of the cutting line and the edge key point of the target to be detected, the width of the overlapping area is obtained, including: according to the pixel value of the cutting line, the value of each tooth in the gear tooth surface The vertex and the bottom point of each tooth respectively obtain the first key point of the gear closest to the pixel value of the cutting line and the second key point of the gear closest to the pixel value of the cutting line; among them, the first key point of the gear and the second key point of the gear The key points are located on both sides of the pixel value of the cutting line; according to the distance from the first key point of the gear to the pixel value of the cutting line, the width of the first overlapping area is obtained; and according to the distance from the second key point of the gear to the pixel value of the cutting line, the second Two overlapping area width.

In the above implementation process, the image to be processed includes a gear tooth surface image; for both sides of a cutting line, there are respectively a first overlapping area width and a second overlapping area width, and the target to be detected includes a gear; according to the gear tooth surface image gear The vertex of each tooth and the bottom point of each tooth in the tooth surface obtain the first key point of the gear closest to the pixel value of the cutting line and the second key point of the gear closest to the pixel value of the cutting line respectively, so as to determine the gear The width of the first overlapping area and the width of the second overlapping area in the tooth surface image ensure that each tooth of the gear is not cut, thereby ensuring the integrity of the information of the gear image when it is subsequently processed.

In the second aspect, the embodiment of the present application also provides an image processing device, including: an acquisition module, used to obtain the original pixel area of the image to be processed, the pixel value of the cutting line, and the width of the overlapping area; The original pixel value, the cutting line pixel value and the width of the overlapping area generate each segmented image pixel area of the image to be processed; and the segmentation module is used to segment the image to be processed according to each segmented image pixel area to obtain the segmented image .

In the third aspect, the embodiment of the present application also provides an electronic device, including: a processor and a memory, the memory stores machine-readable instructions executable by the processor, and when the machine-readable instructions are executed by the processor, they are executed as described above method.

In a fourth aspect, the embodiment of the present application further provides a computer-readable storage medium, where a computer program is stored on the computer-readable storage medium, and when the computer program is run by a processor, the method described above is executed.

Using the image processing method, device, electronic equipment, and storage medium provided by the present application, the image to be processed is segmented according to the pixel area of the segmented image, so that the segmented images of the image to be processed have a certain overlapping area width, and the image to be processed can be avoided. The target in the image is segmented to improve the accuracy of image processing. The coordinate information of the segmented image is converted through the coordinate conversion formula, the restored coordinate information of the segmented image is obtained, and each segmented image is uniformly converted to facilitate subsequent processing.

Description of drawings

In order to more clearly illustrate the technical solutions of the embodiments of the present application, the accompanying drawings that need to be used in the embodiments of the present application will be briefly introduced below. It should be understood that the following drawings only show some embodiments of the present application, so It should not be regarded as a limitation on the scope, and those skilled in the art can also obtain other related drawings according to these drawings without creative work.

FIG. 1 is a schematic flow diagram of an image processing method provided in an embodiment of the present application;

FIG. 2 is a schematic diagram of the width of overlapping regions provided by the embodiment of the present application;

Fig. 3 is a schematic diagram of the width of the overlapping area of the gear tooth surface image provided by the embodiment of the present application;

FIG. 4 is a schematic structural diagram of an image processing device provided in an embodiment of the present application;

FIG. 5 is a schematic structural diagram of an electronic device provided by an embodiment of the present application.

detailed description

Embodiments of the technical solutions of the present application will be described in detail below in conjunction with the accompanying drawings. The following examples are only used to illustrate the technical solution of the present application more clearly, and therefore are only examples, rather than limiting the protection scope of the present application.

Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those skilled in the technical field of the application; the terms used herein are only for the purpose of describing specific embodiments, and are not intended to Limit this application.

In the description of the embodiments of the present application, the technical terms "first", "second" and so on are only used to distinguish different objects, and should not be understood as indicating or implying relative importance or implicitly indicating the number of indicated technical features, A specific order or primary-secondary relationship. In the description of the embodiments of the present application, "plurality" means two or more, unless otherwise specifically defined.

When the image is processed by the model, the model has a fixed input image size, assuming that the fixed input image size of the model is 800*800, that is, the width direction and the height direction are respectively composed of 800 pixel values, assuming that the original image size is 2048*3125, that is, the image The width has 2048 pixel values and the height has 3125 pixel values. If the original image is directly input to the model for processing, the original image is compressed from 2048*3125 to 800*800, the single-pixel accuracy in the width direction will be lost by 2048/800=2.56 times, and the single-pixel accuracy in the high direction will be lost by 3125/800=3.9 times. The loss of single-pixel precision is reflected in the loss of image features on the image, and the features of the image are crucial for model training and reasoning. First crop the original image into two or more segmented images, and then input the segmented images into the model for training and inference. Compared with directly compressing the original image, fewer features will be sacrificed. The more features the image retains, the better the model learns and the more accurate the detection. In addition, if the original image is clipped, if there is a target in the original image located on the cutting line, it is possible to segment the target in the original image, causing any segmented image to fail to recognize the target, resulting in delay The information in the image is lost, resulting in missed detection and false detection in subsequent processing. Therefore, the present application provides an image processing method to solve the above-mentioned problem that the object is clipped.

Please refer to FIG. 1 which is a schematic flowchart of an image processing method provided by an embodiment of the present application.

Step S110: Obtain the original pixel area of the image to be processed, the pixel value of the cutting line, and the width of the overlapping area.

The implementation of the above step S110 includes: obtaining the original pixel area of the image to be processed, the image to be processed may be a complete image including the target, such as a part image captured by a camera; the original pixel area of the image to be processed is an image to be Process the pixel area range of the image to represent the size of the image to be processed. The pixel value of the cutting line is the pixel value of the cutting line of the pre-cut image to be processed, that is, the position of the cutting line is located in the image to be processed by the pixel value of the cutting line. The pixel value of the cutting line can be determined according to the number of segments to be processed and the size of the segmented image, or can be set as a certain value. The width of the overlapping area is the width of the overlapping part between the segmented images in the image to be processed, and the width of the overlapping area can be set to a certain value, or can be calculated and generated according to the target in the segmented image; and the width of the overlapping area of each segmented image can be The same, but can be set separately according to the target in each segmented image.

Step S120: Generate each segmented image pixel area of the image to be processed according to the original pixel value, the cutting line pixel value and the width of the overlapping area.

The implementation of the above step S120 includes: each segmented image pixel region of the image to be processed includes position information and size information of each segmented image in the image to be processed. According to the original pixel value, the pixel value of the cutting line and the width of the overlapping region, generating each segmented image pixel area of the image to be processed can be: first determine the original pixel area of the segmented image according to the original pixel value and the pixel value of the cutting line, and then according to the obtained The width of the overlapping area is to expand the original pixel area of the segmented image, so that after the image to be processed is segmented, the segmented image on both sides of the pixel value of the cutting line has an overlapping area.

A specific example: the size of the image to be processed is W*H, W is the pixel value in the width direction, and H is the pixel value in the height direction. Segment the image in the H direction, set the pixel value of the cutting line as the pixel value at H/2, the original pixel range of the segmented image A is [0, H/2], and the original pixel range of the segmented image B is [( H/2)+1,H]. If the width of the overlapping area is obtained, then the original pixel range of the segmented image A is increased in the H direction by the pixels corresponding to the overlapping area width h, and the pixel area of the segmented image A is obtained; the original pixel range of the segmented image B is H Increase the pixel corresponding to the width of the overlapping area in the direction to obtain the B pixel area of the segmented image.

Step S130: Segment the image to be processed according to each segmented image pixel area to obtain a segmented image.

The implementation of the above-mentioned step S130 includes: after determining the pixel area of each segmented image, that is, determining the pixel position and size of the segmented image in the image to be processed, segmenting the image to be processed according to each segmented image pixel area, and obtaining Segmented image processing image.

In the above implementation process, each segmented image pixel area of the image to be processed is generated according to the original pixel value, the cutting line pixel value and the width of the overlapping area, and there is a certain overlap between the segmented image pixel areas on both sides of the cutting line The area width is used to segment the image to be processed according to the pixel area of the segmented image, which can prevent the target in the image to be processed from being segmented and improve the accuracy of image processing.

Optionally, in the embodiment of the present application, obtaining the original pixel area of the image to be processed, the pixel value of the cutting line, and the width of the overlapping area includes: generating the pixel value of the cutting line according to the original pixel area and the subsequent processing pixel area; wherein, the subsequent processing The pixel area represents the pixel area that needs to be conformed to in the subsequent processing of the image to be processed; input the image to be processed into the preset key point detection model to obtain the edge key points of the target to be detected in the image to be processed; according to the pixel value of the cutting line and Edge key points of the target to be detected to obtain the width of the overlapping area.

The implementation of the above steps is, for example: generating the pixel value of the cutting line according to the original pixel area and the subsequent processing pixel area, and the subsequent processing pixel area represents the pixel area to be conformed to when the image to be processed is subjected to subsequent processing. Subsequent processing can be a model that will input the segmented image, and the pixel area to be matched can be a fixed input image of the model. Subsequent processing can also be other processing methods that have requirements on the size of the segmented image or the pixel area. The pixel value of the cutting line is generated according to the original pixel area and the subsequent processing pixel area. For example, the image to be processed has a width of 700 pixel values and a height of 1400 pixel values. The input image of the subsequent processing model is 800*800, that is, the width is 800 pixel value, up to 800 pixel value. Then, the cutting line can be determined at a height of 700 pixels, so that the original pixel value of the two segmented images segmented according to the cutting line is 700*700, which conforms to the fixed input size of the model. It should be noted that when determining the pixel value of the cutting line, it is necessary to reserve a position for the width of the overlapping area for the segmented image, so that the original pixel value of the segmented image increases the width of the upper overlapping area, which still meets the pixel area requirements for subsequent processing.

Input the image to be processed into the preset key point detection model to obtain the edge key points of the target to be detected in the image to be processed, if the image to be processed includes multiple targets to be detected, then obtain the edge key points of each target to be detected, The edge key points of the target to be detected may be contour points of the target to be detected, and the width of the overlapping region is obtained according to the positional relationship between the key points of the edge of the target to be detected and the pixel values of the cutting line. Specifically, for example, if multiple edge key points are detected, the edge key point closest to the pixel value of the cutting line is obtained, and the vertical distance from the edge key point closest to the pixel value of the cutting line to the pixel value of the cutting line is the width of the overlapping area. It can be understood that the edge key point that is second closest to the pixel value of the cutting line may also be used as the edge key point for determining the width of the overlapping region.

In the above implementation process, the cutting line pixel value is generated according to the original pixel area and the subsequent processing pixel area, that is, the cutting line pixel value is generated according to the pixel area size relationship between the original pixel area and the subsequent processing pixel area, and the subsequent processing pixel area is Process the pixel area that the image needs to conform to during subsequent processing. After the pixel value of the cutting line is determined, the width of the overlapping area is obtained according to the pixel value of the cutting line and the edge key points of the target to be detected. Make the segmented image on both sides of the pixel value along the cutting line have an overlapping area width, so as to avoid the target located on the cutting line from being segmented.

Optionally, in the embodiment of the present application, after segmenting the image to be processed according to each segmented image pixel area and obtaining the segmented image, the method further includes: obtaining coordinate information of the segmented image; The information is converted to obtain restored coordinate information of the segmented image; wherein, the restored coordinate information represents the coordinate information in the image to be processed obtained by splicing multiple segmented images.

The implementation of the above steps is for example: obtain the coordinate information of the segmented image, the coordinate information may be the center coordinate of the segmented image, and the center coordinate may be the detection index of the segmented image obtained after inputting the segmented image into the defect detection model for defect detection. Out-of-frame center coordinates, since the segmented images on both sides of the cutting line have overlapping areas, so if the center coordinates of the segmented images need to be represented in the image to be processed, they need to be converted into restored coordinate information, and the restored coordinate information represents The coordinate information in the image to be processed obtained by splicing multiple segmented images. The image to be processed is an image that has not been segmented, and the coordinate information in the image to be processed may be the center coordinates of the detection frame of the image that has not been segmented. The image to be processed may be obtained by splicing the segmented images by splicing or overlapping splicing of the segmented images according to the pixel areas of the segmented images to obtain a complete image to be processed after splicing.

In the above implementation process, after the image to be processed is segmented, the coordinate information of the segmented image is obtained, and the information of the segmented image is converted into the corresponding overlapping stitching of the segmented image to obtain the restored coordinate information in the image to be processed . Through coordinate transformation, the coordinates of each segmented image correspond to the restored coordinates in the image to be processed, and the processing of the image to be processed is completed.

Optionally, in the embodiment of the present application, converting the coordinate information of the segmented image to obtain the restored coordinate information of the segmented image includes: sequentially inputting multiple segmented images into the preset according to the index numbers of the segmented images The detection model is used to obtain the detection result; wherein, the detection result includes the coordinate information of each segmented image, and the index number is obtained according to the positional relationship of the segmented image in the image to be processed; through the coordinate conversion formula, the coordinate information of the segmented image is carried out Conversion, to obtain the restoration coordinate information of the segmented image; the coordinate conversion formula includes: Y=y-hi+(H/N)*i; wherein, Y is the coordinate value of the segmentation direction in the restoration coordinate information, and y is the cutting direction in the coordinate information The coordinate value of the split direction, hi is the width of the overlapping area corresponding to the split image with the index number i, H is the pixel value of the image to be processed in the split direction, N is the number of splits of the image to be processed; i is the split The index number of the image, i∈[0,N-1].

The implementation of the above steps is for example: the input of the segmented image into the preset detection model has an order, and the segmented image can be input into the preset detection model in sequence according to the index number of the segmented image, and the index number is based on the index number of the segmented image in the image to be processed The positional relationship in is obtained. For example, a group of segmented images is called a batch, and the number of segmented images in a group of segmented images is called batchsize. The model takes a batch of segmented images each time for processing, and outputs the detection results in the order of the segmented images in the batch. If the size of the image to be processed is W*H, W is the pixel value in the width direction, and H is the pixel value in the height direction. The image to be processed is divided into N segmented images, and the batchsize of the model is set to N, that is, the model takes N segmented images each time for processing. The N sliced images have index numbers respectively, and according to the index numbers, the

After the segmented image is input into the preset detection model, the detection result is obtained. The detection result includes coordinate information of each segmented image, where the coordinate information of the segmented image may be the center coordinates of the detection frame of the segmented image obtained after inputting the segmented image into the defect detection model for defect detection. For example, the detection result can be a one-dimensional tensor composed of five elements, namely [x, y, w, h, t], where (x, y) are the center coordinates of the segmented image, (w, h ) is the width and height of the image detection frame in the segmented image, and t is the confidence information.

Through the coordinate conversion formula, the coordinate information of the segmented image is converted to obtain the restored coordinate information of the segmented image; the coordinate conversion formula includes: Y=y-hi+(H/N)*i; wherein, Y is in the restored coordinate information The coordinate value of the segmentation direction, y is the coordinate value of the segmentation direction in the coordinate information, hi is the width of the overlapping area corresponding to the segmentation image with the index number i, H is the pixel value of the image to be processed in the segmentation direction, N is the number of segmented images to be processed; i is the index number of the segmented image, i∈[0,N-1].

Among them, h _i is the width of the overlapping area corresponding to the segmented image with index number i. Specifically, for example, the width of the overlapping area corresponding to the segmented image with index number 1 is h ₁ , and the overlapping area corresponding to the segmented image with index number 2 is The region width is h ₂ .

It should be noted that the embodiment of the present application is based on the example of performing segmentation in the H direction, and the other direction W does not perform segmentation, then the x in the coordinate information of the segmented image is equal to the X in the restored coordinate information, That is, X=x; wherein, X is the coordinate value of the direction in which segmentation is not performed in the restored coordinate information, and x is the coordinate value of the direction in which segmentation is not performed in the coordinate information.

In the above implementation process, the coordinate information of the segmented image is converted through the coordinate transformation formula, the restored coordinate information of the segmented image is obtained, and each segmented image is uniformly converted to facilitate subsequent processing.

Optionally, in the embodiment of the present application, wherein, the original pixel area includes pixel values of the image to be processed in the segmentation direction; the segmented image includes the first segmented image and the second segmented image; according to the original pixel value, The pixel value of the cutting line and the width of the overlapping area are used to generate each segmented image pixel area of the image to be processed, including: according to the pixel value of the cutting line and the width of the overlapping area, each segmented image pixel area of the image to be processed is generated by a division formula; the division formula Including: the pixel area of the first segmented image is: [0, (H/2)+hi]; the pixel area of the second segmented image is: [(H/2)-hi, H]; wherein, H is The pixel value of the image to be processed in the segmentation direction; H/2 is the pixel value of the cutting line; h _i is the width of the overlapping area corresponding to the segmented image.

The implementation of the above steps is for example: the initial pixel area includes pixel values in the segmentation direction of the image to be processed, and if the image to be processed is segmented in the H direction, the pixel values in the segmentation direction are pixel values in the H direction. If the image to be processed includes a cutting line, the image to be processed is segmented into two segmented images, and the segmented images include a first segmented image and a second segmented image.

According to the pixel value of the cutting line and the width of the overlapping area, the pixel area of each segmented image of the image to be processed is generated by a division formula; the division formula includes: the pixel area of the first segmented image is: [0, (H/2)+hi] ; Among them, H is the pixel value of the image to be processed in the segmentation direction; H/2 is the pixel value of the cutting line; h _i is the width of the overlapping area corresponding to the segmentation image. In one embodiment, the first segmented image is located below the image to be processed, the original pixel area of the first segmented image is (0, (H/2)), and the original pixel area is added to the first segmented image corresponding to is the overlapping area width h _i , then the pixel area of the first segmented image is: [0, (H/2)+hi].

The pixel region of the second segmented image is: [(H/2)-hi, H]; Wherein, H is the pixel value of the image to be processed in the segmenting direction; H/2 is the pixel value of the cutting line; h _i is The segmented image corresponds to the overlapping area width. The original pixel area of the first segmented image is ((H/2), H), and the original pixel area plus the second segmented image corresponds to the overlapping area width h _i , since the second segmented image is located in the image to be processed Above, the pixel area of the second segmented image is: [(H/2)-hi,H].

In the above implementation process, each segmented image pixel area of the image to be processed is generated by the division formula, each segmented image pixel area corresponds to a segmented image, and the pixel area of the segmented image is divided to avoid processing according to the cutting line The image is segmented such that the target is segmented.

Please refer to FIG. 2 , which is a schematic diagram showing the width of overlapping regions provided by the embodiment of the present application.

Optionally, in the embodiment of the present application, the method further includes: separately obtaining the edge key points of the object to be detected corresponding to both sides of the pixel value of the cutting line; The key point of the edge of the target, and the width of the overlapping area corresponding to each segmented image on both sides of the pixel value of the cutting line is obtained.

The implementation of the above steps is for example: respectively obtain the edge key points of the object to be detected corresponding to both sides of the pixel value of the cutting line, if the image to be processed has a cutting line, then obtain the corresponding edge points of the object to be detected on both sides of the pixel value of the cutting line Edge key points; if the image to be processed has multiple cutting lines, then obtain the edge key points of the object to be detected corresponding to both sides of the pixel value of each cutting line.

According to the pixel value of the cutting line, and the edge key points of the object to be detected corresponding to the two sides of the pixel value of the cutting line, as shown in Figure 2, the pixel value of the cutting line is at H/2 of the image to be processed, and the cutting line is obtained The edge key points above the pixel value are respectively the first key point and the third key point, wherein the first key point is the key point closest to the pixel value of the cutting line. As an implementation, the first key point, The vertical distance to the pixel value of the cutting line is the width of the first overlapping area, that is, the width of the overlapping area of the segmented image above the pixel value of the cutting line. It is also possible to use the vertical distance from the third key point to the pixel value of the cutting line as the width of the first overlapping area.

And obtain the edge key points below the pixel value of the cutting line, which are respectively the second key point and the fourth key point. The second key point is the key point closest to the pixel value of the cutting line, and the vertical distance from the second key point to the pixel value of the cutting line is the width of the second overlapping area, which is the width of the overlapping area of the segmented image below the pixel value of the cutting line. To achieve the width of the overlapping area corresponding to each segmented image on both sides of the pixel value of the cutting line.

The direction of the cutting line and the position of the pixel value can be set according to the actual situation, which is not limited in the embodiment of the present application. For example, if the direction of the cutting line is horizontal, the corresponding values of the segmented images on the upper and lower sides of the cutting line are respectively obtained. The width of the overlapping area; if the direction of the cutting line is vertical, then obtain the width of the overlapping area corresponding to the segmented images on the left and right sides of the cutting line.

In the above implementation process, according to the pixel value of the cutting line and the edge key points of the target to be detected corresponding to the pixel value of the cutting line, the width of the overlapping area corresponding to each segmented image on both sides of the pixel value of the cutting line can be determined respectively , to flexibly determine the overlapping area width of each segmented image, so that each segmented image has an accurate overlapping area width.

Optionally, in the embodiment of the present application, wherein the image to be processed includes a gear tooth surface image; the width of the overlapping area includes a first overlapping area width and a second overlapping area width; the target to be detected includes a gear; the edge key point includes a gear tooth The vertex of each tooth in the surface and the bottom point of each tooth; according to the pixel value of the cutting line and the edge key point of the target to be detected, the width of the overlapping area is obtained, including: according to the pixel value of the cutting line, the value of each tooth in the gear tooth surface The vertex and the bottom point of each tooth respectively obtain the first key point of the gear closest to the pixel value of the cutting line and the second key point of the gear closest to the pixel value of the cutting line; among them, the first key point of the gear and the second key point of the gear The key points are located on both sides of the pixel value of the cutting line; according to the distance from the first key point of the gear to the pixel value of the cutting line, the width of the first overlapping area is obtained; and according to the distance from the second key point of the gear to the pixel value of the cutting line, the second Two overlapping area width.

Please refer to FIG. 3 , which is a schematic diagram showing the width of overlapping regions of gear tooth surface images provided by the embodiment of the present application.

The implementation of the above steps is for example: the image to be processed includes a gear tooth surface image; the overlapping area width includes a first overlapping area width and a second overlapping area width, wherein the first overlapping area width and the second overlapping area width are cutting line pixels respectively The width of the overlapping region corresponding to the split image on both sides of the value.

As shown in Figure 3, the target to be detected includes the gear, specifically each tooth in the gear; the edge key points include the vertex and the bottom point of each tooth in the gear tooth surface; each tooth in the gear tooth surface image Arranged as slashes, the vertex and bottom point of a slash line are respectively the vertex and bottom point of a tooth. Specifically, for example, the apex of each tooth and the bottom point of each tooth in the gear tooth surface include: the first tooth bottom point, the second tooth bottom point, the third tooth bottom point, the fourth tooth top and bottom point, the fifth tooth bottom point Top and bottom points, sixth tooth apex, and seventh tooth apex.

In this embodiment, the first side of the pixel value of the cutting line is the lower side of the pixel value of the cutting line, and the second side of the pixel value of the cutting line is the upper side of the pixel value of the cutting line. Obtain the first key point of the gear that is closest to the pixel value of the cutting line on the first side, that is, the fourth tooth bottom point; and obtain the second key point of the gear that is the closest to the pixel value of the cutting line on the second side, that is, the first Four-tooth vertex; wherein, the first side and the second side are the two sides of the pixel value of the cutting line, and correspondingly, the first key point of the gear and the second key point of the gear are also located on both sides of the pixel value of the cutting line.

According to the first key point of the gear, that is, the distance from the third tooth bottom point to the pixel value of the cutting line, the width of the first overlapping area is obtained; and according to the second key point of the gear, that is, the distance from the fifth tooth apex to the pixel value of the cutting line, to obtain Second overlap area width. It should be noted that if the first overlapping area width is on the first side of the cutting line pixel value, then the first overlapping area width is the overlapping area width corresponding to the segmented image on the second side of the cutting line pixel value.

In the above implementation process, the image to be processed includes a gear tooth surface image; for both sides of a cutting line, there are respectively a first overlapping area width and a second overlapping area width, and the target to be detected includes a gear; according to the gear tooth surface image gear The vertex of each tooth and the bottom point of each tooth in the tooth surface obtain the first key point of the gear closest to the pixel value of the cutting line and the second key point of the gear closest to the pixel value of the cutting line respectively, so as to determine the gear The width of the first overlapping area and the width of the second overlapping area in the tooth surface image ensure that each tooth of the gear is not cut, thereby ensuring the integrity of the gear image information transmitted to the subsequent processing module.

In a preferred embodiment, the image to be processed can be segmented into a plurality of segmented images according to the pixel values of the plurality of cutting lines, and the segmenting method can be equal or unequal. If the segmentation method is equal, for example, if the segmentation direction is the H direction, and a total of N segmentation images are segmented, then the pixel values of the cutting lines of N-1 cutting lines are required, and the pixel values of the cutting lines are: The pixel value of the first cutting line is H/N, the pixel value of the second cutting line is (2*H)/N...the pixel value of the N-1th cutting line is ((N-1)*H)/N . According to the edge key points closest to the pixel value of the cutting line on both sides of the pixel value of each cutting line, the width of the overlapping area corresponding to the segmented image on both sides of each cutting line is calculated. It should be noted that if the image to be processed is divided into a plurality of segmented images, the segmented images in the middle area respectively have two overlapping area widths in the segmenting direction H, and when determining the pixel area of the segmented image, it is necessary to The corresponding overlapping region widths are added to the original pixel region on both sides of the splitting direction to form a split image pixel region.

Please refer to the schematic structural diagram of the image processing device provided by the embodiment of the present application shown in FIG. 4; the embodiment of the present application provides an image processing device 200, including:

An acquisition module 210, configured to acquire the original pixel area of the image to be processed, the pixel value of the cutting line, and the width of the overlapping area;

Generate pixel area module 220, for generating each segmented image pixel area of the image to be processed according to the original pixel value, the cutting line pixel value and the width of the overlapping area;

The segmentation module 230 is configured to segment the image to be processed according to each pixel region of the segmented image to obtain a segmented image.

Optionally, in the embodiment of the present application, the image processing device 200 and the acquisition module 210 are specifically configured to generate cutting line pixel values according to the original pixel area and the subsequent processing pixel area; wherein, the subsequent processing pixel area indicates that the image to be processed is being processed During subsequent processing, the pixel area that needs to be matched; input the image to be processed into the preset key point detection model to obtain the edge key points of the target to be detected in the image to be processed; according to the pixel value of the cutting line and the edge key point of the target to be detected, Get the overlapping area width.

Optionally, in the embodiment of the present application, the image processing device 200 further includes: a restoration module, configured to obtain coordinate information of the segmented image; convert the coordinate information of the segmented image to obtain restored coordinate information of the segmented image ; Wherein, the restored coordinate information represents the coordinate information in the image to be processed obtained by splicing multiple segmented images.

Optionally, in the embodiment of the present application, the image processing device 200, the restoring module, is specifically configured to sequentially input the multiple segmented images into the preset detection model according to the index numbers of the segmented images to obtain the detection results; wherein, The detection result includes the coordinate information of each segmented image, and the index number is obtained according to the positional relationship of the segmented image in the image to be processed; through the coordinate conversion formula, the coordinate information of the segmented image is converted to obtain the restored coordinate information of the segmented image ;The coordinate transformation formula includes: Y=y-hi+(H/N)*i; wherein, Y is the coordinate value of the segmentation direction in the restored coordinate information, y is the coordinate value of the segmentation direction in the coordinate information, and h _i is an index The width of the overlapping area corresponding to the segmented image numbered i, H is the pixel value of the image to be processed in the segmenting direction, N is the number of segments of the image to be processed; i is the index number of the segmented image, i∈[0 ,N-1].

Optionally, in the embodiment of the present application, the image processing apparatus 200, wherein the original pixel area includes pixel values of the image to be processed in the segmentation direction; the segmented image includes a first segmented image and a second segmented image; Generate the pixel area module 220, specifically for generating each segmented image pixel area of the image to be processed according to the cutting line pixel value and the width of the overlapping area by a division formula; the division formula includes: the pixel area of the first segmented image is: [0 ,(H/2)+hi]; the pixel area of the second segmented image is: [(H/2)-hi, H]; wherein, H is the pixel value of the image to be processed in the segmented direction; H/ 2 is the pixel value of the cutting line; h _i is the width of the overlapping area corresponding to the segmented image.

Optionally, in the embodiment of the present application, the image processing device 200 further includes: an overlapping area acquisition module, configured to respectively obtain edge key points of the target to be detected corresponding to both sides of the pixel value of the cutting line; according to the pixel value of the cutting line, and the edge key points of the object to be detected corresponding to both sides of the pixel value of the cutting line, and obtain the overlapping area width corresponding to each segmented image on both sides of the pixel value of the cutting line.

Optionally, in the embodiment of the present application, the image processing device 200, wherein the image to be processed includes a gear tooth surface image; the width of the overlapping area includes a first overlapping area width and a second overlapping area width; the target to be detected includes a gear; the edge Key points include the apex of each tooth and the bottom point of each tooth in the gear tooth surface; the acquisition module 210 is also used to, according to the pixel value of the cutting line, the apex of each tooth in the gear tooth surface and the bottom point of each tooth, The first key point of the gear closest to the pixel value of the cutting line and the second key point of the gear closest to the pixel value of the cutting line are respectively obtained; wherein, the first key point of the gear and the second key point of the gear are respectively located at the pixel value of the cutting line Both sides; according to the distance from the first key point of the gear to the pixel value of the cutting line, the first overlapping area width is obtained; and according to the distance from the second key point of the gear to the pixel value of the cutting line, the second overlapping area width is obtained.

It should be understood that the device corresponds to the above-mentioned embodiment of the image processing method, and can perform various steps involved in the above-mentioned method embodiment. The specific functions of the device can refer to the description above. To avoid repetition, detailed descriptions are appropriately omitted here. . The device includes at least one software function module that can be stored in a memory in the form of software or firmware (firmware) or solidified in an operating system (operating system, OS) of the device.

Please refer to FIG. 5 , which is a schematic structural diagram of an electronic device provided by an embodiment of the present application. An electronic device 300 provided in an embodiment of the present application includes: a processor 310 and a memory 320. The memory 320 stores machine-readable instructions executable by the processor 310. When the machine-readable instructions are executed by the processor 310, the above methods are executed. .

The embodiment of the present application also provides a storage medium, on which a computer program is stored, and when the computer program is run by a processor, the above method is executed.

Wherein, the storage medium can be realized by any type of volatile or non-volatile storage device or their combination, such as Static Random Access Memory (Static Random Access Memory, referred to as SRAM), Electrically Erasable Programmable Read-Only Memory (Electrically Erasable Programmable Read-Only Memory, referred to as EEPROM), Erasable Programmable Read-Only Memory (Erasable Programmable Read Only Memory, referred to as EPROM), Programmable Read-Only Memory (Programmable Red-Only Memory, referred to as PROM), read-only Memory (Read-Only Memory, ROM for short), magnetic memory, flash memory, magnetic disk or optical disk.

In the several embodiments provided in the embodiments of the present application, it should be understood that the disclosed devices and methods may also be implemented in other ways. The device embodiments described above are only illustrative. For example, the flowcharts and block diagrams in the accompanying drawings show possible implementation architectures of devices, methods, and computer program products according to multiple embodiments of the embodiments of the present application. function and operation. In this regard, each block in a flowchart or block diagram may represent a module, program segment, or portion of code that contains one or more executable instruction. It should also be noted that, in some alternative implementations, the functions noted in the block may occur out of the order noted in the figures. For example, two blocks in succession may, in fact, be executed substantially concurrently, or they may sometimes be executed in the reverse order, depending upon the functionality involved. It should also be noted that each block of the block diagrams and/or flowchart illustrations, and combinations of blocks in the block diagrams and/or flowchart illustrations, can be implemented by a dedicated hardware-based system that performs the specified function or action , or may be implemented by a combination of dedicated hardware and computer instructions.

In addition, each functional module in each embodiment of the embodiment of the present application may be integrated to form an independent part, each module may exist independently, or two or more modules may be integrated to form an independent part.

The above description is only an optional implementation of the embodiment of the present application, but the scope of protection of the embodiment of the present application is not limited thereto. Anyone familiar with the technical field can Changes or substitutions that can easily be thought of should fall within the scope of protection of the embodiments of the present application.

Claims (10)

1. An image processing method, comprising:

obtaining an original pixel area, a cutting line pixel value and an overlapping area width of an image to be processed;

generating each segmentation image pixel area of the image to be processed according to the original pixel value, the cutting line pixel value and the width of the overlapping area; and

and segmenting the image to be processed according to the pixel area of each segmented image to obtain a segmented image.

2. The method of claim 1, wherein obtaining the original pixel area, the cut-line pixel value, and the overlap area width of the image to be processed comprises:

generating the cutting line pixel value according to the original pixel area and the subsequent processing pixel area; the subsequent processing pixel area represents a pixel area which needs to be met when the image to be processed is subjected to subsequent processing;

inputting the image to be processed into a preset key point detection model to obtain edge key points of a target to be detected in the image to be processed;

and obtaining the width of the overlapping area according to the cutting line pixel value and the edge key point of the target to be detected.

3. The method according to claim 1, wherein after the segmenting the image to be processed according to the each segmented image pixel region to obtain the segmented image, the method further comprises:

obtaining coordinate information of the segmented image;

converting the coordinate information of the segmented image to obtain reduced coordinate information of the segmented image; and the reduced coordinate information represents the coordinate information in the image to be processed obtained by splicing the plurality of segmented images.

4. The method according to claim 3, wherein the converting the coordinate information of the sliced image to obtain the restored coordinate information of the sliced image comprises:

sequentially inputting a plurality of the segmentation images into a preset detection model according to the index numbers of the segmentation images to obtain a detection result; the detection result comprises coordinate information of each segmented image, and the index number is obtained according to the position relation of the segmented images in the image to be processed;

converting the coordinate information of the segmented image through a coordinate conversion formula to obtain the reduced coordinate information of the segmented image;

the coordinate conversion formula includes: y = Y-hi + (H/N) × i;

y is a coordinate value of a segmentation direction in the restored coordinate information, Y is a coordinate value of the segmentation direction in the coordinate information, hi is the width of an overlapping area corresponding to the segmented image with the index number i, H is a pixel value of the image to be processed in the segmentation direction, and N is the segmentation quantity of the image to be processed; i is the index number of the segmentation image, and i belongs to [0,N-1].

5. The method according to claim 1, wherein the original pixel region comprises pixel values of the image to be processed in a slicing direction; the segmentation image comprises a first segmentation image and a second segmentation image; generating each segmentation image pixel area of the image to be processed according to the original pixel value, the cutting line pixel value and the width of the overlapping area, wherein the method comprises the following steps:

generating each segmentation image pixel area of the image to be processed through a division formula according to the cutting line pixel value and the width of the overlapping area; the division formula includes:

the pixel area of the first cut image is: [0, (H/2) + hi ];

the pixel area of the second segmentation image is: [ (H/2) -hi, H](ii) a H is a pixel value of the image to be processed in the segmentation direction; h/2 is the cutting line pixel value; h is a total of _i The width of the overlapping area corresponds to the cutting image.

6. The method of claim 2, further comprising:

respectively obtaining edge key points of the target to be detected corresponding to two sides of the pixel value of the cutting line;

and obtaining the width of an overlapping area corresponding to each of the segmentation images at two sides of the cutting line pixel value according to the cutting line pixel value and the edge key points of the target to be detected corresponding to the two sides of the cutting line pixel value.

7. The method of claim 2, wherein the image to be processed comprises a gear tooth surface image; the overlap region width comprises a first overlap region width and a second overlap region width; the target to be detected comprises a gear; the edge key points include a top point of each tooth and a bottom point of each tooth in the gear tooth surface;

the obtaining the width of the overlapping area according to the cutting line pixel value and the edge key point of the target to be detected comprises:

according to the cutting line pixel value, the top point of each tooth in the tooth surface of the gear and the bottom point of each tooth, respectively obtaining a first key point of the gear closest to the cutting line pixel value and a second key point of the gear closest to the cutting line pixel value; the first key point and the second key point of the gear are respectively positioned at two sides of the pixel value of the cutting line;

obtaining the width of the first overlapping area according to the distance from the first key point of the gear to the pixel value of the cutting line; and

and obtaining the width of the second overlapping area according to the distance from the second key point of the gear to the pixel value of the cutting line.

8. An image processing apparatus characterized by comprising:

the acquisition module is used for acquiring an original pixel area, a cutting line pixel value and an overlapping area width of an image to be processed;

a pixel area generating module, configured to generate each split image pixel area of the to-be-processed image according to the original pixel value, the cutting line pixel value, and the width of the overlapping area; and

and the segmentation module is used for segmenting the image to be processed according to each segmented image pixel area to obtain a segmented image.

9. An electronic device, comprising: a processor and a memory, the memory storing machine-readable instructions executable by the processor, the machine-readable instructions, when executed by the processor, performing the method of any of claims 1 to 7.

10. A computer-readable storage medium, having stored thereon a computer program which, when being executed by a processor, is adapted to carry out the method of any one of claims 1 to 7.

Images