Images captured in underwater are degraded by the attenuation of light in water. The classification technique of underwater images based on various degradation has not been explored so far. The classification process is quite difficult due to the complex background of the underwater images. Generally multiple features are extracted to improve the classification accuracy. The significant feature selection plays a vital role in the classification process. In this work, the process is carried out in two steps namely, (i) first, the features of the underwater images are extracted, (ii) the extracted features are given as input to the neural network (NN) to classify these images into different classes of degradation. The efficiency of this classification technique is measured based on the accuracy and error percentage. The experimental results imply that NN performs well in case of 5 classes of degradation due to distinguishable features and produces an accuracy of about 100 percent.

水中捕获的图像会因水中的光衰减而退化。到目前为止，尚未探索基于各种退化的水下图像分类技术。由于水下图像的背景复杂，分类过程非常困难。通常，提取多个特征以提高分类精度。重要的特征选择在分类过程中起着至关重要的作用。在这项工作中，该过程分两个步骤进行，即（i）首先，提取水下图像的特征，（ii）将提取的特征作为输入提供给神经网络（NN），以将这些图像分类为不同类别的退化。该分类技术的效率是根据准确性和错误百分比来衡量的。