Object Detection in Underwater Image by Detecting Edges using Adaptive Thresholding

Identification of natural and artificial objects in under water images is necessary in order to explore the underwater life because of its various application potentials. Degradation of underwater image because of scattering and absorption in underwater condition makes the objects in under water images diminish their perceivability, the image enhancement is therefore necessary in the detection of the underwater objects to improve the appearance of objects. Our method deals with this problem which enhances the image by stretching the contrast and segments it along with extracting boundaries to find the object using adaptive thresholding with Sobel operator. The experimental results of the proposed method on underwater images has performed better than Canny, Prewitt and Hule et al. [3] edge detection based object detection method both by perceiving visually and by subjective criteria, i.e., PSNR and SSIM.