Search

Higher dimensional optical imaging has seen remarkable growth in the past decade. As a representative method, multiplexed structured illumination (MSI) has been widely introduced for its ultra-high resolution of high-dimensional information. In this letter, an imaging framework for recovering higher dimensional image data, which terms quadrature multiplexed structured illumination (QMSI), is proposed and demonstrated. QMSI incorporates the illumination pattern spatial phase information as a new parameter into the encoding process. This strategy can double the frequency spectrum utilization without any overhead compared to MSI, which means it can double the image sequence depth or image spatial resolution. As a proof of concept, we demonstrate the performance of QMSI by experiments on multispectral imaging.

An integrated physical diffractive optical neural network (DONN) is proposed based on a standard silicon-on-insulator (SOI) substrate. This DONN has compact structure and can realize the function of machine learning with whole-passive fully-optical manners. The DONN structure is designed by the spatial domain electromagnetic propagation model, and the approximate process of the neuron value mapping is optimized well to guarantee the consistence between the pre-trained neuron value and the SOI integration implementation. This model can better ensure the manufacturability and the scale of the on-chip neural network, which can be used to guide the design and manufacturing of the real chip. The performance of our DONN is numerically demonstrated on the prototypical machine learning task of prediction of coronary heart disease from the UCI Heart Disease Dataset, and accuracy comparable to the state-of-the-art is achieved.

A high-speed wavelength-dependent speckle generator with the refreshing rate up to 100MHz is proposed. This can be used for compressive video sensing for recovering high-speed moving scenes from one blurred image.

The current three-dimensional (3D) imaging technique of dynamic scene has some problems, such as a large amount of data and high requirements for the hardware equipment. To remedy these defects, we propose a novel 3D dynamic compressive imaging system that combines structured-light 3D surface imaging with temporal ghost imaging (TGI). The system aims to acquire a 3D video through single exposure imaging, and then restore the 3D dynamic scene through the post-processing algorithm. Since a slow camera can be used to capture 3D video, the system has the advantages of low detector bandwidth and low data volume. Specifically, a fringe pattern with a frequency component is projected onto a moving object to obtain 3D information, and then image the object onto the digital metal micro-mirror device (DMD) for pixel-by-pixel optical encoding to achieve video compression, finally, the camera captures an image in a single exposure. To decode the image, the optical intensity correlation is calculated to obtain the temporal frames, which retain the deformed fringe patterns modulated by the object, then the 3D surface profile of the frame can be reconstructed through Fourier fringe analysis and phase-height mapping. The experiment verifies two types of motion scenes, translation and rotation, 15-frame 3D video is compressed in one image, that is, the frame rate of the camera is increased by 15 times, and the equivalent compression ratio is 6.67%.

Single-shot ultrafast photography is a powerful tool for science research and industry applications. In this Letter, a novel strategy for ultrafast imaging of two-dimensional complex (amplitude and phase) objects, which is termed single-shot ultrafast phase retrieval photography (SUP), is proposed and demonstrated. The key component of SUP is a silicon photonic integrated chip, which not only has the function of multi-angle illumination, but also provides ultra-short delays for each illumination source. Combined with an ultra-short pulse source and coherent diffraction imaging, SUP can realize ultrafast single-shot imaging. As a proof of concept, the self-developed multiplexed time delay illumination chip was used for experiments, and we demonstrated the reconstructing of a static complex-valued object with a frame sequence depth of 16 frames from single-shot ptychographic data.

The novel camera architecture facilitates the development of machine vision. Instead of capturing frame sequences in the temporal domain as traditional video cameras, FourierCam directly measures the pixel-wise temporal spectrum of the video in a single shot through optical coding. Compared to the classic video cameras and time-frequency transformation pipeline, this programmable frequency-domain sampling strategy has an attractive combination of characteristics for low detection bandwidth, low computational burden, and low data volume. Based on the various temporal filter kernel designed by FourierCam, we demonstrated a series of exciting machine vision functions, such as video compression, background subtraction, object extraction, and trajectory tracking.

An opto-electronic neural network is designed for video object detection from a long-exposure blurred image. This network combines an optical encoder, convolutional neural network decoder, and object detection module, which are jointly optimized end-to-end. The joint loss is adopted for updating the network according to the physical constraints of hardware via back-propagation. A high-speed refreshed spatial light modulator is used as the encoder part of the network to generate coded sub-images, and then, a single blurred image is obtained after a common camera. The rest of the network is used for video object detection. Both simulations and experiments demonstrate that our framework can successfully retrieve object labels and bounding boxes at different moments in the long exposure. To the best of our knowledge, this is the first work investigating video object detection from a single motion-degraded image.

Time is crucial for survivals. According to the last signal of the plane, we set up four models to find survivals or find the black box: the possible falling area model, calculate initial searching area, rescue model, enlarge searching area model. Based on these models, we use actual map to calculate rescue time. Then we iterate 5 times to ensure to locate the target. Finally, SearchVessel.cpp and SatelliteSearch.cpp are given to test the model. Restatement of the searching task After the air. Rash of MH370 in March 8th 2014,people all over the world have been worrying about 227 passengers and 12 crew members of MH370.How to avoid air crash and search for the black boxes is still a worldwide problem.Relatives of the victims are in deep sorrow. The Aircraft of Communications Addressing and Reporting System, as ACARS can still receiving the signals of the transponder, despite the human factor such as Captain Mr.H and military radar screen .Satellite engineers painstakingly made a big breakthrough saying that the plane may not fall in the South China Sea. Assumption The Atlantic Ocean, the Pacific Ocean, the Indian Ocean, the Southern Ocean, the Arctic Ocean are the five mean oceans. Only 1.2% of the ocean in the world is deeper than 6000 meters, so we give priority to the epicontinental sea. The living condition of the Arctic Ocean and there is extreme difficult, so once there is an accident in these places, it is impossible to have any survivals. We only need to find the black box. We assume that we use bluefin to search survivals as rescue tools. Regardless of drag coefficient, we model the possible falling area. The position indicator listen to the signals from the black box, rescue team operate on the control screen to mark the strong signal point. We ignore the signal interference such as stock of fish, the wreck of old ships. The TPL-25 System meets the Navy requirement for locating emergency relocation pingers on downed Navy and commercial aircraft down to a maximum depth of 20,000 feet anywhere in the world. The current TPL-25 replaces all previous systems; TPL-20, 30, & 40. If plane blast in the air, the condition is much more difficult to discuss. Justification of our approach. The system consists of the tow fish, tow cable, winch, hydraulic power unit, generator, and topside control console, although not all of these components are required on every mission. Navigation is accomplished by using algorithms incorporating the amount of cable in the water, the depth indication from the pressure sensor and other parameters. The generator provides electrical power for the system or power from the support platform can be used if it is compatible. The received acoustic signal of the pinger is transmitted up the cable and is presented audibly, and can be output to either an Oscilloscope, or Signal Processing Computer. The operator monitors the greatest signal strength and records the navigation coordinates. This procedure is repeated on multiple track lines until the final position is triangulated. There are five difficult problems we need to deal with. 2nd International Conference on Education Technology, Management and Humanities Science (ETMHS 2016) © 2016. The authors Published by Atlantis Press 114 (1)Large search area.There is usually current in the ocean. The air crash happened two days ago .The place where we find the plane wreckage is usually not the location where the plane falls to the ocean. So we should increase the search area. (2)The climate is always changing.The climate is always changing in the ocean.Sometimes there is catastrophe like hurricane. It is difficult to search for the plane under the weather like this. (3)The black can only signal for thirty days.The black can signal for only thirty days.But it took eight days for French navy to go to the incident area. (4)The depth of the ocean.The incident area is of the average depth of about 2000m to 3000m and the deepest area is of the depth of 7000m according to the data from Brazil.The French rescue boat carries a mini deep sea submarine.When the submarine dives to the deep sea,it can discharge a deep-sea robot.But the robot can only word within the deep of 6000m under the ocean. (5)Small signal range.From a technical point of view, the signal of the black box can be detected in the range of one square kilometer. If the depth of the black box under the ocean is beyond 3000m the effective signal range is about zero. So we can seldom find the black box.

Based on the Nagel-Schrekenbercellular automaton model of traffic flow, the article analyzed the influence that driving on the right side takes to the traffic flow (practical capacity) in the condition of light and heavy traffic. With the combination of fluid dynamics and vehicle dynamics, we established the Cellular Automata model (referred to as C) for mixed speed two-lane traffic flow on the rule driving on the right side. We also got the relationship among speed, density and traffic though the traffic simulation platform and MatLab numerical simulation. Then we used the method to find out the large bus share rate, the number of vehicles changing lanes and the influence law of safety factors on traffic flow. And we came to a conclusion that the relationship between traffic flow and load are inverted U shape changes in the low load and high load conditions,The conclusion, which provides a theoretical basis for the reasonable organization of expressway traffic management.