Improved Least Absolute Deviation Estimation in the Application of the Puny Target Detection

Xia ZHU

Abstract


Considering the characteristics of weak puny targets in images, a detection method of small targets was proposed. Firstly, a model was built based on improved least absolute deviation. According to the properties of the improved least square estimation, then we used improved particle swarm algorithm to solve problems of extreme values. The predicted residual image was obtained by subtracting the forecasting image from the original image. The puny target was detected by the maximum between-cluster variance threshold segmentation method. The experimental results were compared with the results of puny target detection method based on chaotic PSO and least absolute deviation. The experimental results showed that the proposed method had higher detection probability. The experimental results showed that the method had higher detection probability and better detection results.

Keywords


Puny target detection, Least absolute deviation estimation, Particle swarm optimization, Maximum between-cluster variance, Threshold segmentation

Publication Date


2016-11-17 00:00:00


DOI
10.12783/dtetr/amita2016/3696

Full Text:

PDF

Refbacks

  • There are currently no refbacks.