Deep Lightweight Unrolled Network for High Dynamic Range Modulo Imaging

High dynamic image is an important technology of photoelectric information acquisition, providing higher dynamic range and more image details, and it can better reflect the real environment, light and color information. Currently, the method of high dynamic range image synthesis based on different exposure image sequences cannot adapt to the dynamic scene. It fails to overcome the effects of moving targets, resulting in the phenomenon of ghost. Therefore, a new high dynamic range image acquisition method based on multiplex cameras system was proposed. Firstly, different exposure images sequences were captured with the camera array, using the method of derivative optical flow based on color gradient to get the deviation between images, and aligned the images. Then, the high dynamic range image fusion weighting function was established by combination of inverse camera response function and deviation between images, and was applied to generated a high dynamic range image. The experiments show that the proposed method can effectively obtain high dynamic images in dynamic scene, and achieves good results.

The quality assessment of a high dynamic image is a challenging task. The few available no reference image quality methods for high dynamic range images are generally in evaluation stage. The most available image quality assessment methods are designed to assess low dynamic range images. In the paper, we show the assessment of high dynamic range images which are generated by utilizing a virtually flexible fill factor on the sensor images. We present a new method in the assessment process and evaluate the amount of improvement of the generated high dynamic images in comparison to original ones. The results show that the generated images not only have more number of tonal levels in comparison to original ones but also the dynamic range of images have significantly increased due to the measurable improvement values.

Real world scenes contain a large range of light intensities. To adapt to display device, High Dynamic Range (HDR) image should be converted into Low Dynamic Range (LDR) image. A common task of tone mapping algorithms is to reproduce high dynamic range images on low dynamic range display devices. In this paper, a new tone mapping algorithm is proposed for high dynamic range images. Based on the probabilistic model is proposed for high dynamic image's tone reproduction, the proposed method uses a logarithmic normal distribution instead of normal distribution. Therefore, the algorithm can preserve visibility and contrast impression of high dynamic range scenes in the common display devices. Experimental results show the superior performance of the app roach in terms of visual quality.

As spatial light modulators based on Digital Micro-mirror Devices (DMDs) combined with image sensors could detect strong and weak targets in high dynamic scenes and get high dynamic images, this paper researches the application of DMDs to high dynamic scene image detection systems. Firstly, the application principle of the DMD was introduced, and how to control and drive the DMD in the high dynamic scene image detection system was analyzed. Then, based on DMD's working characteristics, a control driver was designed to make optoelectronic image equipment improve its dynamic range by 66 dB. The main modules were described and working parameters in the system were given. Finally, the feasibility of the system was verified experimentally. The result shows that the control driver detects the strong and weak targets simultaneously in the high dynamic scene, and improves the dynamic range of 66 dB. When a 11 bit higher sensitive imaging sensor is used in the system, the dynamic scene will be over 130 dB.

Dynamic imaging technology has a wide range of applications. High dynamic image (HDR) can obtain better realism and has irreplaceable advantages over traditional image. Compared with the hardware implementation, the software algorithm using multiple photos with different exposures has the advantages of simpler and lower cost. However, multiple images are prone to different degrees of displacement, which requires image registration technology for image registration. The main contents of this paper are image registration related technology, image synthesis, camera response function fitting, color synthesis and so on. The core problem is to study image registration technology, including mean binarization (MTB). This is an important link in the process of HDR image synthesis and has important practical significance.

High dynamic range imaging is an emerging technology for generating high quality images. The most common method is to acquire high dynamic range images in a multi-exposure fusion manner. The biggest disadvantage of such algorithms is the artifact phenomenon caused by the target motion, or the time cost of avoiding artifacts for registration. Therefore, Therefore, a high dynamic range imaging algorithm based on the just noticeable difference (JND) and detail enhancement is proposed, which belongs to the generation of high dynamic images from a single image. According to the JND edge of the improved human visual characteristics and local variance matrix, combined with the fuzzy system to obtain the weight matrix describing the quality of different exposure images, so that different exposure images are fused into HDR images. The experimental results show that the algorithm can effectively improve the contrast and clarity of the image, and the generated image is more in line with the subjective visual effect of the human eye.

Traffic light recognition (TLR) detects the traffic light from an image and then estimates the state of the light signal. TLR is important for autonomous vehicles because running against a red light could cause a deadly car accident. For a practical TLR system, computation time, varying illumination conditions, and false positives are three key challenges. In this paper, a novel real-time method is proposed to recognize a traffic light with high dynamic imaging and deep learning. In our approach, traffic light candidates are robustly detected from low exposure/dark frames and accurately classified using a deep neural network in consecutive high exposure/bright frames. This dual-channel mechanism can make full use of undistorted color and shape information in dark frames as well as the rich context in bright frames. In the dark channel, a non-parametric multi-color saliency model is proposed to simultaneously extract lights with different colors. A multiclass classifier with convolutional neural network (CNN) model is then adopted to reduce the number of false positives in the bright channel. The performance is further boosted by incorporating temporal trajectory tracking. In order to speed up the algorithm, a prior detection mask is generated to limit the potential search regions. Intensive experiments on a large dual-channel dataset show that the proposed approach outperforms the state-of-the-art real-time deep learning object detector, which could cause more false positives because it uses bright images only. The algorithm has been integrated into our autonomous vehicle and can work robustly on real roads.

This paper describes about compact approach for high dynamic range imaging in the mobile camera. The dynamic range of real world is much larger than that of the mobile sensor. So it's almost impossible to make high dynamic image with just one shot. This paper focuses on the small number of frames for high dynamic range imaging synthesis and the complexity of algorithm. With only one short and long exposure frames we calculate radiance map in the Bayer domain, process tone mapping with multi-layer decomposition method and compensate for motion artifacts for moving objects.

This work presents a study and analysis of Hallucination in HDR Images it presents a technique for the high dynamic range image in over-exposed and underexposed regions, the hallucination image become blur and hence a bilateral filter algorithm is applied to smooth the image and preserve edges more effectively after applying the tone mapping algorithm to map these high dynamic images to produce a high dynamic range image will contents intact, this mapping technique is known as hallucination technique for High dynamic range images.

High dynamic range images enable more detailed information regarding the content. By performing luminance compression, images can be prepared for the traditional low dynamic range reproduction. There are numerous ways to perform the compression (tone-mapping), but the content may affect the image quality. The reduction of dynamic range may produce unwanted effects (e.g. due to noise) and can be tailored for specific image content. In this paper we analyze several examples related to synthetic aperture radar (SAR) as images with high dynamic range.

High Dynamic Range Image (HDRI) is a type of images with wide luminosity levels in real scenes. It is very important to fuse the multiple photographs into a single HDRI. A method that constructs HDRI through recovering response curve of camera imaging system is proposed. The method can make effect of several important parameters to the synthesis results, and get a synthesis result for images missing the information of exposure time. As for the HDRI tone reproduction, an improved algorithm is presented in last part of the paper for processing HDRI and getting good performance

In order to accomplish the software definition micro-nano satellite demands, which includes that its payload functions and parameters could be reconstructive and controllable by uploading software as needs, we have to break through the design limitations between traditional satellite platform and ordinary optical camera, one new type of optical imaging camera technology is developed based on software definition micro-nano satellite here. we gave full consideration to the possible development joint designing space between the software and the hardware of the payload. Then we analyze the influence of sub-pixel information, satellite platform parameters, optical system parameters, detector parameters, noise and atmosphere on image data processing, especially the super-resolution reconstruction. we establish the physical model and the error model according to the physical mechanism of each factor, as a priori information of the reconstruction method, we apply these prior information constraints in favor of super-resolution to the design of the camera, enabling the images captured by the camera to match the super resolution method very well. This method can simultaneously improve visual resolution and substantial resolution, while maintaining the ability of suppressing noise and may reduce the size and development difficulty of traditional cameras. At the same time, we also carry out high dynamic range imaging technology based on the definition of CMOS software, which assists multi frame image superposition, and fit in the image processing algorithm, breaking through the digital dynamic range enhancement technology, and realizing the ability of high dynamic range clear imaging over high performance CCD. We have developed a general purpose computing optical imaging camera, which integrates the super resolution imaging, dynamic range enhanced imaging, video imaging and other multi intelligent controllable imaging modes. Finally we have completed the related camera integration, testing and experiment.

Fluorescence acquisition and image display over a high dynamic range is highly desirable. However, the limited dynamic range of current photodetectors and imaging CCDs impose a limit on the fluorescence intensities that can be simultaneously captured during a single image acquisition. This is particularly troublesome when imaging biological samples, where protein expression fluctuates considerably. As a result, biological images will often contain regions with signal that is either saturated or hidden within background noise, causing information loss. In this manuscript we summarize recent work from our group and others, to extended conventional to high dynamic range fluorescence imaging. These strategies have many biological applications, such as mapping of neural connections, vascular imaging, bio-distribution studies or pharmacologic imaging at the single cell and organ level.

This paper presents a CMOS image sensor with auto-switched mode readout architecture for high throughput and high dynamic range scientific imaging applications. The proposed readout circuit compares background illumination with a threshold level and automatically adjusts the pixel array to operate in either linear or logarithmic mode by feeding back a mode switching control signal. This novel readout architecture allows simultaneous linear-logarithmic operation of each pixel, resulting in direct readout of image sequence covering a dynamic range of 121 dB. A chip consisting of 16 times 16 8-T pixel array and the proposed readout circuit has been designed and fabricated using AMI 0.5 mum technology. The chip has been verified for imaging clusters of nano particles (CoFe2O4) illuminated using bright green LED. Experiment results confirm high dynamic range performance due to linear-logarithmic auto-switched mode operation.

We investigate the stereo matching techniques for high dynamic range HDR image pairs. It is an emerging topic in computer vision and multimedia applications due to the availability of HDR image capture devices. The disparity computation will eventually take the stereo HDR input. In this work, three state-of-the-art stereo matching algorithms are modified and used to test the advantages of HDR stereo matching. By performing the HDR bit-plane slicing, it is found that only about 16 bits per channel is required for the HDR image format. We propose a 16-bit unsigned integer format to store the HDR image, which allows the available stereo matching algorithms to be adopted for disparity computation. Experiments and performance evaluation are carried out using Middlebury stereo datasets.