Linear target detection method and system based on non-subsampled Contourlet transform

Abstract

The invention discloses a linear target detection method and system based on non-subsampled Contourlet transform, and the method comprises the following steps: carrying out the multi-level and multi-direction NSCT decomposition of an original remote sensing image, and obtaining a plurality of low-frequency subgraphs and high-frequency subgraphs; based on each high-frequency sub-graph and each low-frequency sub-graph which are obtained through decomposition, obtaining a second-sieve high-frequency line segment set sum, performing fusion processing on the second-sieve low-frequency line segmentset, and obtaining a high-frequency fusion result RH and a low-frequency fusion result RL; and carrying out fusion processing again to obtain low-frequency and high-frequency fusion results, and completing linear target detection. The invention provides an identification method and system which are high in confidence coefficient, good in robustness and capable of effectively detecting linear targets such as bridges and airfield runways in complex terrains.

Description

technical field [0001] The invention belongs to the field of aerospace technology, relates to the field of linear target detection in remote sensing images, in particular to a linear target detection method and system based on non-subsampling Contourlet transformation. Background technique [0002] In aerospace research, a large amount of ground observation data and aerial surveillance data can be obtained using sensors installed on aerospace vehicles at different altitudes. In order to make full use of these data and adapt to different application requirements, information fusion technology emerges as the times require. Information fusion technology uses certain rules to analyze, process and comprehensively utilize the collected data information to complete the required decision-making and evaluation tasks. It can effectively improve the utilization rate of data and the reliability of the system's detection and identification of targets. As an important branch of the field...

AI Technical Summary

Problems solved by technology

[0014] Compared with several existing straight line segment detection algorithms, the experimental results show that the existing straight line segment detection algorithms cannot detect linear targets such as roads and bridges in remote sensing images well, and the detection effect of LSD is relatively better, but LSD is easy to detect. affected by noise

Images