Computational Optical Imaging: Pushing the limits of traditional optical imaging
Release:
Release:
outline
Computational optical imaging is an emerging multidisciplinary field. It takes the specific application task as the criterion, obtains or encodes light field information (such as angle, polarization, phase, etc.) through multiple dimensions, and designs a new paradigm of perception for sensors far beyond that of the human eye. At the same time, combined with mathematical and signal processing knowledge, the light field information is deeply excavated to break through the limits of traditional optical imaging. At present, computational optical imaging is in a stage of rapid development, and many exciting research results have been achieved, and large-scale applications have begun in mobile phone cameras, medical treatment, unmanned driving and other fields. In the future, computational optical imaging is expected to further subvert the traditional imaging system and bring more creative and imaginative applications, such as lensless imaging, non-field imaging, etc.
Trend interpretation
Traditional optical imaging is based on geometric optics, drawing on the principle of "what you see is what you get" in human vision, while ignoring many optical high-dimensional information. At present, traditional optical imaging is close to the physical limit in terms of hardware function and imaging performance, and can no longer meet the application needs in many fields. For example, in the field of mobile phone photography, it is impossible to reduce the weight and volume of the device while ensuring the imaging effect, and the criticism of "front bangs" and "rear bath bombs" appears; In the field of microscopic imaging, it is impossible to meet the needs of wide field of view and high resolution at the same time; In the field of surveillance remote sensing, it is difficult to obtain clear images in complex environments with low light and low visibility.
With the continuous evolution of new generation information technologies such as sensors, cloud computing, and artificial intelligence, new solutions are gradually emerging: computational optical imaging. Computational optical imaging takes the specific application task as the criterion, obtains or encodes light field information (such as angle, polarization, phase, etc.) through multiple dimensions, and designs a new paradigm of perception for sensors far beyond that of the human eye. At the same time, combined with mathematical and signal processing knowledge, the light field information is deeply mined to break through the limits of traditional optical imaging.
Computational optical imaging is an emerging multidisciplinary field, with early concepts only taking shape in the mid-70s. With the vigorous development of information technology, computational optical imaging has become an international research hotspot. Due to the wide coverage of computational optical imaging research, there is no clear classification method. According to the application problems solved by computational imaging technology, it can be roughly divided into the following three categories:
(1)function improvement: imaging or measuring optical information that cannot be obtained by traditional methods, such as light field, polarization, coherence, etc.;
(2)Performance improvement: that is, improve the performance indicators of existing imaging technologies, such as spatial resolution, temporal resolution, depth of field, robustness of complex environments, etc.;
(3) Simplification and intelligence: Simplify imaging systems through specific technologies such as single pixel and lensless, or achieve specific AI tasks at the speed of light.
Computational optical imaging is an emerging multidisciplinary field. It takes the specific application task as the criterion, obtains or encodes light field information (such as angle, polarization, phase, etc.) through multiple dimensions, and designs a new paradigm of perception for sensors far beyond that of the human eye. At the same time, combined with mathematical and signal processing knowledge, the light field information is deeply excavated to break through the limits of traditional optical imaging. At present, computational optical imaging is in a stage of rapid development, and many exciting research results have been achieved, and large-scale applications have begun in mobile phone cameras, medical treatment, unmanned driving and other fields. In the future, computational optical imaging is expected to further subvert the traditional imaging system and bring more creative and imaginative applications, such as lensless imaging, non-field imaging, etc.
Trend interpretation
Traditional optical imaging is based on geometric optics, drawing on the principle of "what you see is what you get" in human vision, while ignoring many optical high-dimensional information. At present, traditional optical imaging is close to the physical limit in terms of hardware function and imaging performance, and can no longer meet the application needs in many fields. For example, in the field of mobile phone photography, it is impossible to reduce the weight and volume of the device while ensuring the imaging effect, and the criticism of "front bangs" and "rear bath bombs" appears; In the field of microscopic imaging, it is impossible to meet the needs of wide field of view and high resolution at the same time; In the field of surveillance remote sensing, it is difficult to obtain clear images in complex environments with low light and low visibility.
With the continuous evolution of new generation information technologies such as sensors, cloud computing, and artificial intelligence, new solutions are gradually emerging: computational optical imaging. Computational optical imaging takes the specific application task as the criterion, obtains or encodes light field information (such as angle, polarization, phase, etc.) through multiple dimensions, and designs a new paradigm of perception for sensors far beyond that of the human eye. At the same time, combined with mathematical and signal processing knowledge, the light field information is deeply mined to break through the limits of traditional optical imaging.
Computational optical imaging is an emerging multidisciplinary field, with early concepts only taking shape in the mid-70s. With the vigorous development of information technology, computational optical imaging has become an international research hotspot. Due to the wide coverage of computational optical imaging research, there is no clear classification method. According to the application problems solved by computational imaging technology, it can be roughly divided into the following three categories:
(1)function improvement: imaging or measuring optical information that cannot be obtained by traditional methods, such as light field, polarization, coherence, etc.;
(2)Performance improvement: that is, improve the performance indicators of existing imaging technologies, such as spatial resolution, temporal resolution, depth of field, robustness of complex environments, etc.;
(3) Simplification and intelligence: Simplify imaging systems through specific technologies such as single pixel and lensless, or achieve specific AI tasks at the speed of light.