Local image area characteristic extraction method based on scale prediction
A technology of local area features and extraction methods, applied in computer parts, instruments, characters and pattern recognition, etc., can solve the problem that the SURF feature extraction method does not have affine invariance, etc. Effect
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
Method used
Image
Examples
specific Embodiment approach 1
[0036] Specific implementation mode one: combine figure 1 To describe this embodiment,
[0037] A method for feature extraction of image local regions based on scale prediction, comprising the following steps:
[0038] Step 1: According to the position of the probe relative to the predetermined landing point during the planetary landing process, the attitude of the probe body coordinate system relative to the planetary surface image taken in orbit, the focal length of the camera, the field of view and other information, the currently captured image is The position of the surface of the target celestial body is initially estimated, and the search range in the global feature library is selected;
[0039] Step 2: According to the pose information of the detector and the feature scale of the corresponding feature point in the global feature library, predict the feature scale of the feature in the captured image;
[0040] Step 3: Predict the rotation angle of the feature in the c...
specific Embodiment approach 2
[0044] The specific implementation process of step 2 described in this embodiment is as follows:
[0045] The distance between the detector and the feature is d when the global feature library is established to capture the image 1 , the focal length of the camera used to capture images when establishing a global feature library is f 1 , the angle between the line between the optical center of the camera and the feature point and the optical axis is α when the image is captured in the global feature library 1 , let the feature scale of a feature in the global feature library be σ 1 , then this feature will produce a scaling transformation in the descending image, and its scale transformation is
[0046] σ 2 = d 1 f 2 cosα 1 d ...
specific Embodiment approach 3
[0051] The specific implementation process of step 3 described in this embodiment is as follows:
[0052] A certain feature in the global feature library will generate a rotation transformation in the descending segment image, and the rotation angle is
[0053]
[0054] Other steps and parameters are the same as those in Embodiment 1 or Embodiment 2.
PUM
Login to View More Abstract
Description
Claims
Application Information
Login to View More