In deep ocean applications, a camera’s viewing range is related to its low-light-level performance. Hence, a high-resolution ultra-low-light-level digital still camera (thereafter L3DSC), designed from the ground up for deep ocean autonomous underwater vehicle (AUV) imaging, is presented. A high-resolution (1024 × 1024) electron-multiplying CCD (EMCCD) is adopted as an image sensor to ensure low-light capability. A thermo electric cooler (TEC) is employed to lower image sensor temperature to promote low-light-level performance. Totem pole circuits that are able to generate 50-V pulse wave at 10 MHz are implemented to drive the EM pin, and power consumption of the circuit is optimized. An EMCCD digital image is buffered in a field-programmable gate array (FPGA) and then transferred through USB interface to a solid-state drive (SSD), which is installed in the image storage unit. The camera is controlled by AUV via Ethernet, and image data stored in the camera could be downloaded via the same interface after AUV retrieved from the sea. The camera module is mounted into a 6000-m depth-rate titanium alloy housing. The diagonal field of view is measured to be 58.4 deg in air. Experiments show that the minimum scene illumination of L3DSC is better than 5 × 10 − 4 Lux; underwater imaging distance is longer than 10 m, and total image data capacity is 200 gigabytes. These results demonstrate the camera’s low-light-level imaging performance and feasibility for AUV applications.
The mobile pipeline-filtering algorithm is a real-time algorithm that performs well in detecting small dim targets, but it is particularly sensitive to interframe vibration of sequence images. When searching for small dim targets at sea based on an infrared imaging system, irregular and random vibration of the airborne imaging platform causes huge interference problems for the mobile pipeline-filtering. This paper puts forward a pipeline-filtering algorithm that has a good performance on self-adaptive anti-vibration. In the block matching method using the normalized cross-correlations coefficient (NCC), the interframe vibration of sequence images is acquired in real time and used to correct coordinates of the single-frame detection results, and then the corrected detection results are used to complete the mobile pipelinefiltering. Experimental results show that the algorithm can overcome the problem of interframe vibration of sequence images, thus realizing accurate detection of small dim maritime targets.