Indoor visual positioning method for solving basic matrix on the basis of pixel threshold value
A basic matrix and visual positioning technology, applied in the field of image processing, can solve problems such as large errors, achieve high positioning accuracy, reduce positioning errors, and high robustness
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
Method used
Image
Examples
specific Embodiment approach 1
[0067] Specific implementation mode one: combine figure 1 To describe this embodiment, an indoor visual positioning method based on a pixel threshold to solve the basic matrix provided in this embodiment specifically includes the following steps:
[0068] Step 1. Cutting the size of the user's image to be consistent with the size of the image in the database;
[0069] Step 2, using the SURF algorithm to extract feature points from the user image and the image in the database respectively;
[0070] Step 3: Use the SURF algorithm to match the feature points between the user image and the image in the database, and obtain the Euclidean distance of each matching feature point pair; the image in the database that has the most matching feature point pairs with the user image is the one that matches the user image;
[0071] Step 4, arrange the matching feature point pairs between the user image and the image matching the user in ascending order according to the Euclidean distance; ...
specific Embodiment approach 2
[0077] Specific embodiment 2: The difference between this embodiment and specific embodiment 1 is that the specific process of extracting feature points described in step 2 includes:
[0078] Step 21. Feature point detection:
[0079] The first step of applying the SURF algorithm for feature point extraction is feature point detection, using a box filter to convolve the image, by changing the size of the box filter, using box filters of different sizes in the x, y, Perform convolution in the three directions of z and construct a scale space pyramid to form a multi-scale space function D xx ,D yy ,D xy ; where D xx Represent points on the image with Gaussian second order partial derivatives The result of convolution, where D yy Represent points on the image with Gaussian second order partial derivatives The result of convolution, where D xy Represent points on the image with Gaussian second order partial derivatives The result of convolution; x represents the absciss...
specific Embodiment approach 3
[0086] Specific embodiment three: the difference between this embodiment and specific embodiment two is that the specific process of feature point matching described in step three includes:
[0087] Feature point matching refers to finding the most similar feature vectors in a high-dimensional vector space; the similarity of feature points is measured according to the Euclidean distance between feature vectors.
[0088] Perform the following steps for all feature points in the user image in turn:
[0089] Calculate the Euclidean distance between a feature point in the user image and all the feature points in the image in the database, select the nearest neighbor feature point Euclidean distance Ed_min1 and the second nearest neighbor feature point Euclidean distance Ed_min2, and calculate the ratio ratio of the two, for the ratio ratio A feature point less than or equal to the first threshold T_Ed is considered to be a correctly matched feature point, otherwise it is a wrongly...
PUM
Login to View More Abstract
Description
Claims
Application Information
Login to View More