阅读说明：本技术 一种图像特征提取方法及装置 (Image feature extraction method and device ) 是由 王易诚 方晓鑫 但汉光 黄启才 孙艳英 于 2018-05-21 设计创作，主要内容包括：一种图像特征提取方法及装置,涉及图像处理技术领域,用于加快图像特征提取的速度。该方法包括：获取输入图像,该输入图像为灰度图像(S301)；对输入图像进行切片和下采样处理,得到N层金字塔图像,N层金字塔图像中的每层金字塔图像包括多个切片图像(S302)；提取每层金字塔图像中每个切片图像的至少一个特征点,每个特征点为所述每层金字塔图像的角点(S303)；对每层金字塔图像的多个切片图像的特征点进行筛选处理,得到每层金字塔图像的多个目标特征点,每个目标特征点为选择出的预设区域中响应得分值最大的特征点(S304)。(An image feature extraction method and device relate to the technical field of image processing and are used for accelerating the image feature extraction speed. The method comprises the following steps: acquiring an input image, wherein the input image is a gray scale image (S301); slicing and downsampling the input image to obtain N layers of pyramid images, wherein each layer of pyramid image in the N layers of pyramid images comprises a plurality of slice images (S302); extracting at least one feature point of each slice image in each layer of pyramid image, wherein each feature point is the corner point of each layer of pyramid image (S303); and (S304) screening the feature points of the plurality of slice images of each layer of pyramid image to obtain a plurality of target feature points of each layer of pyramid image, wherein each target feature point is the feature point with the largest response score value in the selected preset area (S304).)

An image feature extraction method, characterized in that the method comprises:

acquiring an input image, wherein the input image is a gray image;

slicing and downsampling the input image to obtain N layers of pyramid images, wherein the N layers of pyramid images are N images with different resolutions, each layer of pyramid image in the N layers of pyramid images comprises a plurality of slice images, and N is a positive integer greater than or equal to 2;

extracting at least one feature point of each slice image in each layer of pyramid images to obtain a plurality of feature points of the plurality of slice images of each layer of pyramid images, wherein each feature point is a corner point of each layer of pyramid images, and each feature point is represented by a coordinate and a response score value;

and screening a plurality of feature points of a plurality of slice images of each layer of pyramid images to obtain a plurality of target feature points of each layer of pyramid images, wherein each target feature point is a feature point with the largest response score value in a preset area in each layer of pyramid images.

The method of claim 1, wherein each feature point is a corner point after non-maxima suppression processing.

The method according to claim 1 or 2, wherein the input image is a layer 0 pyramid image; the slicing and downsampling processing of the input image to obtain an N-layer pyramid image includes:

slicing the 0 th layer pyramid image to obtain a plurality of slice images of the 0 th layer pyramid image;

and downsampling a plurality of slice images in the ith layer of pyramid image to obtain an (i +1) th layer of pyramid image, and slicing the (i +1) th layer of pyramid image to obtain a plurality of slice images of the (i +1) th layer of pyramid image, wherein i is more than or equal to 0 and less than or equal to N-2.

The method according to any one of claims 1 to 3, wherein the step of performing a screening process on the plurality of feature points of the plurality of slice images of each layer of pyramid image to obtain a plurality of target feature points of each layer of pyramid image comprises:

acquiring a plurality of feature points of a plurality of slice images of each layer of pyramid image to obtain a feature point set;

and screening a specified number of feature points from the feature point set by using an octree screening mode to obtain a plurality of target feature points.

The method according to any one of claims 1-4, further comprising:

and determining description information of each target characteristic point, wherein the description information comprises a characteristic point direction and a descriptor.

The method according to claim 5, wherein the determining the description information of the target feature point comprises:

carrying out Gaussian blur processing on the slice images in each layer of pyramid images;

and determining the direction of the characteristic point and the descriptor according to the processed slice image and the target characteristic point of the slice image so as to obtain the description information of the target characteristic point.

An image feature extraction device characterized by comprising:

the device comprises an acquisition unit, a processing unit and a display unit, wherein the acquisition unit is used for acquiring an input image which is a gray image;

a processing unit to: slicing and downsampling the input image to obtain N layers of pyramid images, wherein the N layers of pyramid images are N images with different resolutions, each layer of pyramid image in the N layers of pyramid images comprises a plurality of slice images, and N is a positive integer greater than or equal to 2;

extracting at least one feature point of each slice image in each layer of pyramid image to obtain a plurality of feature points of the plurality of slice images of each layer of pyramid image, wherein each feature point is the corner point of each layer of pyramid image, and each feature point is represented by a coordinate and a response score value;

and screening a plurality of feature points of a plurality of slice images of each layer of pyramid images to obtain a plurality of target feature points of each layer of pyramid images, wherein each target feature point is a feature point with the largest response score value in a preset area in each layer of pyramid images.

The apparatus of claim 7, wherein the feature points of each slice image are corner points after non-maxima suppression processing.

The apparatus according to claim 7 or 8, wherein the input image is a layer 0 pyramid image, and wherein the processing unit is further configured to:

slicing the 0 th layer pyramid image to obtain a plurality of slice images of the 0 th layer pyramid image;

and downsampling a plurality of slice images in the ith layer of pyramid image to obtain an (i +1) th layer of pyramid image, and slicing the (i +1) th layer of pyramid image to obtain a plurality of slice images of the (i +1) th layer of pyramid image, wherein i is more than or equal to 0 and less than or equal to N-2.

The apparatus according to any of claims 7-9, wherein the processing unit is further configured to:

acquiring a plurality of feature points of a plurality of slice images of each layer of pyramid image to obtain a feature point set;

and screening a specified number of feature points from the feature point set by using an octree screening mode to obtain a plurality of target feature points.

The apparatus according to any of claims 7-10, wherein the processing unit is further configured to:

and determining description information of each target characteristic point, wherein the description information comprises a characteristic point direction and a descriptor.

The apparatus of claim 11, wherein the processing unit is further configured to:

carrying out Gaussian blur processing on the slice images in each layer of pyramid images;

and determining the direction of the characteristic point and the descriptor according to the processed slice image and the target characteristic point of the slice image to obtain the description information of the target characteristic point.

An image feature extraction device characterized by comprising:

the input interface is used for acquiring an input image, and the input image is a gray image;

a processor configured to process the following operations:

slicing and downsampling the input image to obtain N layers of pyramid images, wherein the N layers of pyramid images are N images with different resolutions, each layer of pyramid image in the N layers of pyramid images comprises a plurality of slice images, and N is a positive integer greater than or equal to 2;

extracting at least one feature point of each slice image in each layer of pyramid image to obtain a plurality of feature points of the plurality of slice images of each layer of pyramid image, wherein each feature point is the corner point of each layer of pyramid image, and each feature point is represented by a coordinate and a response score value;

and screening a plurality of feature points of a plurality of slice images of each layer of pyramid images to obtain a plurality of target feature points of each layer of pyramid images, wherein each target feature point is a feature point with the largest response score value in a preset area in each layer of pyramid images.

The apparatus of claim 13, wherein the processor is a Digital Signal Processor (DSP).

The apparatus according to claim 13 or 14, wherein the feature point of each slice image is a corner point after non-maximum suppression processing.

The apparatus according to any of claims 13-15, wherein the input image is a layer 0 pyramid image, and wherein the processor is further configured to:

slicing the 0 th layer pyramid image to obtain a plurality of slice images of the 0 th layer pyramid image;

and downsampling a plurality of slice images in the ith layer of pyramid image to obtain an (i +1) th layer of pyramid image, and slicing the (i +1) th layer of pyramid image to obtain a plurality of slice images of the (i +1) th layer of pyramid image, wherein i is more than or equal to 0 and less than or equal to N-2.

The apparatus of any of claims 13-16, wherein the processor further performs the following:

acquiring a plurality of feature points of a plurality of slice images of each layer of pyramid image to obtain a feature point set;

and screening a specified number of feature points from the feature point set by using an octree screening mode to obtain a plurality of target feature points.

The apparatus of any of claims 13-17, wherein the processor further performs the following:

and determining description information of each target characteristic point, wherein the description information comprises a characteristic point direction and a descriptor.

The apparatus of claim 18, wherein the processor further performs the following:

carrying out Gaussian blur processing on the slice images in each layer of pyramid images;

and determining the direction of the characteristic point and the descriptor according to the processed slice image and the target characteristic point of the slice image to obtain the description information of the target characteristic point.