This article explains:
1. spatial transform image
2. opencv geometric transformation: movement, rotation, affine transformation and the like.
1. spatial transform image
Spatial transform into geometric and non-geometric transformation transformation, geometric transformation which changes the shape of the main image, instead of changing the main geometric transformation image pixel values.
Geometric transformation: the original image, it needs to be changed in accordance with changes in the size, shape and position of the
Transformation type: two-dimensional geometric transformation geometric transformation of the image, the three-dimensional image, a three-dimensional projection transformation to the two-dimensional plane, etc.
note:
Spatial coordinates
- forward mapping method: calculate the coordinates of the target image from an original image coordinates, image, translating this calculation using the transformation
- Backward mapping method: the original image is calculated from the coordinates of the image coordinate result, rotation, scaling, distortion may be used
Pixel value calculating - Interpolation gradation
- nearest neighbor interpolation method: Select the nearest point of the pixel gray scale value (upper left)
Features: simple, fast, and good gray scale fidelity large errors, poor visual characteristics, the mosaic effect
- bilinear interpolation (first order interpolation)
Features: computing feature adjacent to more fully consider the points having a gray smooth transition characteristics; generally satisfactory results are obtained; a low pass filter characteristic, so that the image blur profile; smoothing degraded image detail, especially in when amplification; discontinuities produce undesirable results
- Best Interpolation
- Advanced Interpolation