Patent
Noise detection method based on self-defined features of a single image
Reads0
Chats0
TLDR
In this paper, the authors proposed a self-defined feature extraction method for image noise detection based on transformer substation images, and the method comprises the steps: collecting a transform substation image, and calculating an image x-direction gradient graph and an image y-direction gradients graph; traversing the to-be-detected image and calculating a gradient direction angle of each pixel point according to the gradient map; giving a direction distribution number, calculating the direction distribution value of pixel point in each direction according to gradient value and the direction angle; calculating anintegral graphAbstract:
The invention relates to a noise detection method based on self-defined features of a single image, and the method comprises the steps: collecting a transformer substation image, and calculating an image x-direction gradient graph and an image y-direction gradient graph; traversing the to-be-detected image, and calculating a gradient direction angle of each pixel point according to the gradient map; giving a direction distribution number, calculating a direction distribution value of each pixel point in each direction according to the gradient value and the direction angle, and calculating anintegral graph so as to realize self-defined feature extraction; carrying out statistical analysis on the image gradient information and the self-defined features according to a self-defined feature extraction result, and obtaining an average gradient value and a main direction distribution value under given conditions; calculating an image noise degree, giving a noise degree threshold value, if the noise degree is greater than the given threshold value, considering that the image belongs to a noise image, and outputting the noise degree, otherwise, considering that the image belongs to a normal image, thereby realizing image noise detection. The method solves the problem of noise degree detection of a single image, and has important practical application value.read more
Citations
More filters
Patent
Reference-free spliced image quality evaluation method and system
TL;DR: In this article, a reference-free spliced image quality evaluation method and system is presented, which consists of two steps: first, calculating an adaptive threshold, and then determining the number of pixel points in the neighborhood of each pixel point according to the detail complexity of the splice image, and acquiring the average CLBP window gradient difference value of each point through calculation.