SIFT image matching algorithm in view of improved SUSAN operator and affine invariance

Feature points can be used to match images. Candidate feature points are extracted through SIFT firstly. Then feature points are selected from candidate points through singular value decomposing. Distance between feature points sets is computed According to theory of invariability of feature points set, images are matched if the distance is less than a threshold. Experiment showed that this algorithm is available.

Download Full-text

MSURF: A new image matching algorithm which combines homography and SURF algorithm

2015 23rd International Conference on Geoinformatics ◽

10.1109/geoinformatics.2015.7378709 ◽

2015 ◽

Author(s):

Zhong Saishang ◽

Chen Yifu ◽

Zhang Min ◽

Xie Zhong ◽

Liu Can

Keyword(s):

Image Matching ◽

Matching Algorithm

Download Full-text

COMPARISON OF SEMI AUTOMATIC DTM FROM IMAGE MATCHING WITH DTM FROM LIDAR

ISPRS - International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences ◽

10.5194/isprs-archives-xli-b3-373-2016 ◽

2016 ◽

Vol XLI-B3 ◽

pp. 373-380

Author(s):

Aji Rahmayudi ◽

Aldino Rizaldy

Keyword(s):

Image Matching ◽

Point Cloud ◽

Statistical Data ◽

Aerial Images ◽

Mountainous Terrain ◽

Topographic Map ◽

Matching Algorithm ◽

Time Calculation ◽

Accuracy Difference ◽

Map Scale

Nowadays DTM LIDAR was used extensively for generating contour line in Topographic Map. This method is very superior compared to traditionally stereomodel compilation from aerial images that consume large resource of human operator and very time consuming. Since the improvement of computer vision and digital image processing, it is possible to generate point cloud DSM from aerial images using image matching algorithm. It is also possible to classify point cloud DSM to DTM using the same technique with LIDAR classification and producing DTM which is comparable to DTM LIDAR. This research will study the accuracy difference of both DTMs and the result of DTM in several different condition including urban area and forest area, flat terrain and mountainous terrain, also time calculation for mass production Topographic Map. From statistical data, both methods are able to produce 1:5.000 Topographic Map scale.

Download Full-text