A Marine Organism Detection Framework Based on the Joint Optimization of Image Enhancement and Object Detection

Abstract

Underwater vision-based detection plays an increasingly important role in underwater security, ocean exploration and other fields. Due to the absorption and scattering effects of water on light, as well as the movement of the carrier, underwater images generally have problems such as noise pollution, color cast and motion blur, which seriously affect the performance of underwater vision-based detection. To address these problems, this study proposes an end-to-end marine organism detection framework that can jointly optimize the image enhancement and object detection. The framework uses a two-stage detection network with dynamic intersection over union (IoU) threshold as the backbone and adds an underwater image enhancement module (UIEM) composed of denoising, color correction and deblurring sub-modules to greatly improve the framework’s ability to deal with severely degraded underwater images. Meanwhile, a self-built dataset is introduced to pre-train the UIEM, so that the training of the entire framework can be performed end-to-end. The experimental results show that compared with the existing end-to-end models applied to marine organism detection, the detection precision of the proposed framework can improve by at least 6%, and the detection speed has not been significantly reduced, so that it can complete the high-precision real-time detection of marine organisms.

Research Challenges, Recent Advances and Benchmark Datasets in Deep-Learning-Based Underwater Marine Object Detection: A Review

“…In article [37], an end-to-end marine creature detection framework is developed to handle the picture quality degradation problem of noise pollution, color cast, and motion blur. In this framework, image enhancement and object detection are jointly optimized, where a backbone based on a two-stage detection network is proposed, with an underwater image enhancement module (UIEM) added to significantly increase the framework's capacity to deal with highly damaged underwater photos.…”

Research Challenges, Recent Advances and Benchmark Datasets in Deep-Learning-Based Underwater Marine Object Detection: A Review

“…In [49], to address the image quality degradation problem of noise pollution, color cast, and motion blur, an end-to-end marine organism detection framework that can jointly optimize the image enhancement and object detection, is proposed. The framework uses a two-stage detection network as the backbone and adds an underwater image enhancement module (UIEM) to greatly improve the framework's ability to deal with severely degraded underwater images.…”

Research Challenges, Recent Advances and Benchmark Datasets in Deep-Learning-Based Underwater Marine Object Detection: A Review

“…Image enhancement as preprocess [26], [27], [28], [29], [30] Image enhancement integrated in detector [22], [31], [32], [33] Small object detection Multi-scale representation [9], [34], [35], [33], [36], [37], [18], [38],…”

“…Without interferences of color distortion, object detection performance can be improved while keeping low computational complexity. Another similar work is [32], where an end-to-end marine organism detection framework was developed to handle underwater image quality degradation problems due to noise pollution, color cast, and motion blurring. As illustrated in Fig.…”