Visual Target Size Estimation under Noise Jamming

Target detail information missing caused by low resolution or turbulence degraded, makes it extremely difficult to target identification and classification. Meanwhile, the category of the target usually have significant relationship with its size, it is feasible to preliminary identification and classification target by the size of the target. Our paper proposes a target size estimation method under noise jamming. First of all, removing noise of image, Then, get the image edge by detection, And then get the correct edge of the image by singularity removing. Finally, estimating the size of target. Experiments show that the method can estimate target size under a certain accuracy, can provide a preliminary basis for further object identification and classification.