Research Of Low Light Level Image Detection

With further exploration of human beings into this world, image detection technology in low-light environment has been applied in fields such as military affairs, astronomy, deep-sea exploration, security and medical science. Image detection technology, thanks to the progress made in sensor processing and digital image processing, has witnessed great development in itself. However, there are still problems existing, like sensors' low signal to noise ratio(SNR), low contrast, low spatial resolution, under the minimum detectable intensity of illumination.Thus, a method is proposed here for solution, which combines superposition forming signals with the frame accumulation technology to enhance the image sensor's sensitivity and SNR. It then can achieve improvement in sensitivity without changing the sensor's internal structure and material used significantly, reducing cost and complication as a result. Also, this method serves as an effective way for general improvement of image sensor's sensitivity and SNR in detection.Research accomplished in this paper is as follows:Study existing methods in the field as well as techniques involved, such as image sensor's parameter and noise analysis, frame accumulation technology and forming signal; Analyze the feasibility of forming signal if applied in this method; Elaborate several methods to be proposed for applying or removing forming optimal signal;Design a low-light experimental system to test the proposed methods of enhancing sensitivity, which consists of three parts: firstly, to produce forming optimal signals using DDS with FPGA development board and D/A digital analog converter; secondly, to collect video image mainly using NI image capture card and CCD camera; thirdly, to process the image data with upper computer.Use this experimental system to testify two methods: least square estimation and Discrete Fourier transform method. Analyze the results with some objective or subjective image evaluation criteria.