Remote sensing target detection algorithm based on improved FastMBD

Abstract

The invention discloses a remote sensing target detection algorithm based on improved FastMBD. The method comprises the steps of firstly, dividing an image into a plurality of sub-images according todifferent sizes; calculating a FastMBD saliency map for the sub-images; filtering significant values corresponding to the shadows; combining the sub-images with the same size to form a local saliencymap, then fusing the local saliency maps corresponding to different sizes, performing consistency highlighting on the fused image to obtain an improved FastMBD saliency map, and finally extracting a saliency target through an internal and external saliency ratio to complete a target detection task.

Description

technical field [0001] The invention relates to an improved FastMBD-based target detection algorithm, which belongs to the technical field of digital image processing. Background technique [0002] Target detection can be divided into supervised and unsupervised. Most of the supervised methods are to train the model and then detect the test image, because the training process will take a long time and is restricted by the training samples, even considering the scale, calculation Most of the detection process can be pre-detected through unsupervised detection to reduce the amount of calculation and improve the efficiency of the algorithm. [0003] Salient target detection is often used in the field of target detection due to its low time consumption and unsupervised advantages. Among them, the FastMBD algorithm is a good target detection model with both target detection and high computational efficiency. However, the FastMBD algorithm is based on the assumption that the targ...

AI Technical Summary

Problems solved by technology

However, the FastMBD algorithm is an algorithm based on the assumption that the target is at the center, but in remote sensing images, usually an image contains multiple or even a large number of targets, so the application of the FastMBD algorithm on remote sensing images will lead to a significant reduction in the recall rate.

Images