CN-116740604-B - Object identification method, device and storage medium

CN116740604BCN 116740604 BCN116740604 BCN 116740604BCN-116740604-B

Abstract

The application provides an object identification method, an object identification device and a storage medium, relates to the technical field of communication, and is used for solving the technical problem that an object which is blocked cannot be accurately identified in the prior art. The object identification method comprises the steps of generating a plurality of first edge images according to a plurality of original images, generating a plurality of to-be-fused images according to the plurality of first edge images, wherein each first edge image comprises at least one target object, fusing the to-be-fused images corresponding to the first edge images according to each first edge image to obtain a plurality of fused images, training according to the plurality of fused images to obtain an object identification model, and the object identification model is used for identifying whether the blocked object in the to-be-detected images is the target object.

Inventors

WU YULIN
DU FUZHI
MENG QINGLU
FU SHUAI
LI JIAHUI

Assignees

中国联合网络通信集团有限公司

Dates

Publication Date: 20260505
Application Date: 20230606

Claims (12)

1. An object recognition method, comprising: Generating a plurality of first edge images according to the plurality of original images, wherein each first edge image in the plurality of first edge images comprises at least one target object; Acquiring the position of a target pixel point with a pixel value being a first preset value in preset target objects, wherein the preset target objects are at least one target object in all target objects contained in the plurality of first edge images; determining the size of a shielding area of the preset target object according to the size of the preset target object; determining the position of the target pixel point as the central position of the shielding region, and determining the position of the shielding region according to the size of the shielding region; according to the position of the shielding area, shielding the original image to which the preset target object belongs to obtain a plurality of images to be fused, wherein the images to be fused comprise the image with the blocked target object and the image without the blocked target object; Fusing the first edge image and the image to be fused corresponding to the first edge image aiming at each first edge image to obtain a plurality of fused images; And training according to the multiple fusion images to obtain an object recognition model, wherein the object recognition model is used for recognizing whether the blocked object in the image to be detected is a target object or not.
2. The method for identifying an object according to claim 1, wherein the performing, according to the position of the occlusion region, occlusion processing on an original image to which the preset target object belongs to obtain the plurality of images to be fused includes: determining the position of the shielding region in the original image to which the preset target object belongs according to the position of the shielding region; and converting the pixel value of each pixel point in a target area into a preset pixel value to obtain the multiple images to be fused, wherein the preset pixel value is the average value of the pixels of each pixel point, and the target area is a shielding area in an original image to which the preset target object belongs.
3. The method of claim 1, wherein generating a plurality of first edge images from a plurality of original images comprises: acquiring a plurality of original images, wherein each original image in the plurality of original images comprises at least one target object; converting the plurality of original images into a plurality of second edge images by an edge detection technique; and converting the pixel values of the pixel points except the target object in each second edge image into a second preset value to obtain a plurality of first edge images.
4. The method according to claim 1, wherein the fusing the first edge image and the image to be fused corresponding to the first edge image for each first edge image to obtain a plurality of fused images includes: Fusing the plurality of first edge images and the images to be fused corresponding to the plurality of first edge images on a channel level aiming at each first edge image to obtain a plurality of fused images; Or for each first edge image, extracting the features of the first edge image and the images to be fused corresponding to the first edge image, fusing the extracted features, and convolving the fused features to obtain the fused images.
5. The object recognition method according to any one of claims 1 to 4, further comprising: Acquiring the image to be detected; and inputting the image to be detected into the object recognition model to obtain a recognition result, wherein the recognition result is whether the blocked object in the image to be detected is the target object or not.
6. An object recognition apparatus includes a processing unit; The processing unit is used for generating a plurality of first edge images according to the plurality of original images, wherein each first edge image in the plurality of first edge images comprises at least one target object; the processing unit is further used for generating a plurality of images to be fused according to the plurality of first edge images, wherein the plurality of images to be fused comprise an image with the target object being blocked and an image without the target object being blocked; The processing unit is also used for acquiring the position of a target pixel point with a pixel value being a first preset value in a preset target object, wherein the preset target object is at least one target object in all target objects contained in the plurality of first edge images, determining the size of a shielding area of the preset target object according to the size of the preset target object, determining the position of the target pixel point as the central position of the shielding area, and determining the position of the shielding area according to the size of the shielding area; The processing unit is further configured to fuse, for each first edge image, the first edge image with an image to be fused corresponding to the first edge image, so as to obtain a plurality of fused images; The processing unit is further used for training to obtain an object recognition model according to the multiple fusion images, and the object recognition model is used for recognizing whether the blocked object in the image to be detected is a target object or not.
7. The object recognition device according to claim 6, wherein the processing unit is specifically configured to: determining the position of the shielding region in the original image to which the preset target object belongs according to the position of the shielding region; and converting the pixel value of each pixel point in a target area into a preset pixel value to obtain the multiple images to be fused, wherein the preset pixel value is the average value of the pixels of each pixel point, and the target area is a shielding area in an original image to which the preset target object belongs.
8. The object recognition device according to claim 6, wherein the processing unit is specifically configured to: acquiring a plurality of original images, wherein each original image in the plurality of original images comprises at least one target object; converting the plurality of original images into a plurality of second edge images by an edge detection technique; and converting the pixel values of the pixel points except the target object in each second edge image into a second preset value to obtain a plurality of first edge images.
9. The object recognition device according to claim 6, wherein the processing unit is specifically configured to: Fusing the plurality of first edge images and the images to be fused corresponding to the plurality of first edge images on a channel level aiming at each first edge image to obtain a plurality of fused images; Or for each first edge image, extracting the features of the first edge image and the images to be fused corresponding to the first edge image, fusing the extracted features, and convolving the fused features to obtain the fused images.
10. The object recognition apparatus according to any one of claims 6 to 9, further comprising an acquisition unit; the acquisition unit is used for acquiring the image to be detected; The processing unit is further configured to input the image to be detected into the object recognition model to obtain a recognition result, where the recognition result is whether the blocked object in the image to be detected is the target object.
11. An object recognition device comprising a memory for storing computer-executable instructions and a processor coupled to the memory via a bus, the processor executing the computer-executable instructions stored in the memory when the object recognition device is in operation to cause the object recognition device to perform the object recognition method of any one of claims 1-5.
12. A computer readable storage medium comprising computer executable instructions which, when run on a computer, cause the computer to perform the object recognition method of any one of claims 1-5.

Description

Object identification method, device and storage medium Technical Field The present application relates to the field of communications technologies, and in particular, to an object identification method, an object identification device, and a storage medium. Background In recent years, with the continuous development of deep learning techniques, deep learning techniques have also been applied to a plurality of scenes. For example, the optical fiber construction site can be automatically detected by a target detection technology in deep learning so as to avoid problems in the construction site. However, due to the complex environment of the construction site, the target object is easily blocked. When the target object is blocked, the existing target detection technology cannot accurately identify the blocked object. Disclosure of Invention The application provides an object identification method, an object identification device and a storage medium, which are used for solving the technical problem that an object which is blocked cannot be accurately identified in the prior art. In order to achieve the above purpose, the application adopts the following technical scheme: The object recognition method comprises the steps of generating a plurality of first edge images according to a plurality of original images, generating a plurality of images to be fused according to the first edge images, wherein each first edge image comprises at least one target object, fusing the first edge images and the images to be fused corresponding to the first edge images according to each first edge image, training according to the fused images to obtain an object recognition model, and recognizing whether the blocked object in the images to be detected is the target object or not. The method comprises the steps of generating a plurality of images to be fused according to a plurality of first edge images, wherein the images to be fused comprise the steps of obtaining positions of target pixel points with pixel values being first preset values in preset target objects, wherein the preset target objects are at least one target object in all target objects contained in the plurality of first edge images, determining the size of a shielding area of the preset target objects according to the size of the preset target objects, determining the positions of the target pixel points as the center positions of the shielding areas, determining the positions of the shielding areas according to the sizes of the shielding areas, and shielding the original images to which the preset target objects belong according to the positions of the shielding areas so as to obtain the images to be fused. Optionally, according to the position of the shielding region, carrying out shielding treatment on an original image to which a preset target object belongs to obtain a plurality of images to be fused, wherein the method comprises the steps of determining the position of the shielding region in the original image to which the preset target object belongs according to the position of the shielding region, converting pixel values of all pixel points in the target region into preset pixel values to obtain a plurality of images to be fused, wherein the preset pixel values are pixel average values of all pixel points, and the target region is the shielding region in the original image to which the preset target object belongs. Optionally, generating a plurality of first edge images according to the plurality of original images includes obtaining a plurality of original images, wherein each original image in the plurality of original images includes at least one target object, converting the plurality of original images into a plurality of second edge images through an edge detection technology, and converting pixel values of pixel points except the target object in each second edge image into a second preset value to obtain a plurality of first edge images. Optionally, for each first edge image, fusing the first edge image and the image to be fused corresponding to the first edge image to obtain a plurality of fused images, including fusing the plurality of first edge images and the image to be fused corresponding to the plurality of first edge images on a channel level to obtain a plurality of fused images for each first edge image, or extracting features of the first edge image and the image to be fused corresponding to the first edge image for each first edge image, fusing the extracted features, and convolving the fused features to obtain a plurality of fused images. Optionally, the method further comprises the steps of obtaining an image to be detected, inputting the image to be detected into an object recognition model to obtain a recognition result, and judging whether the blocked object in the image to be detected is a target object or not. The object recognition device comprises a processing unit, a processing unit and a processing unit, wherein the proc