3606 results about "Spectral image" patented technology

A hyperspectral imaging system having an optical path. The system including an illumination source adapted to output a light beam, the light beam illuminating a target, a dispersing element arranged in the optical path and adapted to separate the light beam into a plurality of wavelengths, a digital micromirror array adapted to tune the plurality of wavelengths into a spectrum, an optical device having a detector and adapted to collect the spectrum reflected from the target and arranged in the optical path and a processor operatively connected to and adapted to control at least one of: the illumination source; the dispersing element; the digital micromirror array; the optical device; and, the detector, the processor further adapted to output a hyperspectral image of the target. The dispersing element is arranged between the illumination source and the digital micromirror array, the digital micromirror array is arranged to transmit the spectrum to the target and the optical device is arranged in the optical path after the target.

An imaging apparatus for obtaining spectral image data from an object that includes: a) a light source; b) an input optics section; c) a programmable spectral filter that conditions the multispectral image bearing light according to a predetermined spectral transmission function; d) a detector array in the path of the conditioned multispectral image bearing light and providing a corresponding output signal; and, e) a control logic processor in communication with the spatial light modulator for modulating addressable areas of the spatial light modulator to provide the spectral transmission function thereby.

The present invention relates to the production of instantaneous or non-instantaneous multi-band images, to be transformed into multi- or hyperspectral images, comprising light collecting means (11), an image sensor (12) with at least one two dimensional sensor array (121), and an instantaneous colour separating means (123), positioned before the image sensor array (121) in the optical path (OP) of the arrangement (1), and first uniform spectral filters (13) in the optical path (OP), with the purpose of restricting imaging to certain parts of the electromagnetic spectrum. The present invention specifically teaches that a filter unit (FU) comprising colour or spectral filter mosaics and / or uniform colour or spectral filters mounted on filter wheels (114) or displayed by transmissive displays (115), is either permanently or interchangeably positioned before the colour separating means (123) in the optical path (OP) in, or close to, converged light (B). Each colour or spectral filter mosaic consists of a multitude of homogeneous filtering regions. The transmission curves (TC) of the filtering regions of a colour or spectral filter mosaic can be partly overlapping, in addition to overlap between these transmission curves and those belonging to the filtering regions of the colour separating means (123). The transmission curves (TC) of the colour or spectral filter mosaics and the colour separating means (123) are suitably spread out in the intervals of a spectrum to be studied. The combination of the colour separating means (123) and the spectral or colour or spectral filter mosaics produces different sets of linearly independent transmission curves (TC). The multiple-filter image captured by the image sensor (12) is demosaicked by identifying and segmenting the image regions that are affected by the regions of the multiple filter mosaic, and after an optional interpolation step, a multi-band image is obtained. The resulting multi-band image is transformed into a multi- or hyperspectral image.

The present invention relates to a hyperspectral image classification method based on spectral-spatial cooperation of a deep convolutional neural network, which leads the conventional deep convolutional neural network applied to a two-dimensional image into the three-dimensional hyperspectral image classification problem. Firstly, the convolutional neural network is trained by using a small volume of label data, and a spectral-spatial feature of a hyperspectral image is autonomously extracted by using the network without carrying out any compression and dimensionality reduction processing; then, a support vector machine (SVM) classifier is trained by using the extracted spectral-spatial feature so as to classify an image; and finally, the trained neural network is combined with the trained classifier, the neural network extracts a spectral-spatial feature of a to-be-classified target and the classifier determines a specific category of the extracted spectral-spatial feature so as to acquire a structure (DCNN-SVM) that can autonomously extract the spectral-spatial feature of the hyperspectral image and carry out classification to the spectral-spatial feature, thereby forming a set of hyperspectral image classification method.

A scanning endoscope, amenable to both rigid and flexible forms, scans a beam of light across a field-of-view, collects light scattered from the scanned beam, detects the scattered light, and produces an image. The endoscope may comprise one or more bodies housing a controller, light sources, and detectors; and a separable tip housing the scanning mechanism. The light sources may include laser emitters that combine their outputs into a polychromatic beam. Light may be emitted in ultraviolet or infrared wavelengths to produce a hyperspectral image. The detectors may be housed distally or at a proximal location with gathered light being transmitted thereto via optical fibers. A plurality of scanning elements may be combined to produce a stereoscopic image or other imaging modalities. The endoscope may include a lubricant delivery system to ease passage through body cavities and reduce trauma to the patient. The imaging components are especially compact, being comprised in some embodiments of a MEMS scanner and optical fibers, lending themselves to interstitial placement between other tip features such as working channels, irrigation ports, etc.

Tissue information of a desired deep portion of a biological tissue based on a spectral image obtained from signal processing is adjusted to image information in a color tone suitable for observation. Outputs of a matrix computing section 436 are respectively connected to integrating sections 438a to 438c, and after integrating computation is performed for them, color conversion computation is performed for respective spectral image signals ΣF1 to ΣF3 in a color adjusting section 440, spectral color channel image signals Rch, Gch and Bch are created from the spectral image signals ΣF1 to ΣF3, and images of the spectral color channel images Rch, Gch and Bch are sent to a display monitor 106 via a switching section 439.

A two-dimensional focal plane array (FPA) is divided into sub-arrays of rows and columns of pixels, each sub-array being responsive to light energy from a target object which has been separated by a spectral filter or other spectrum dividing element into a predetermined number of spectral bands. There is preferably one sub-array on the FPA for each predetermined spectral band. Each sub-array has its own read out channel to allow parallel and simultaneous readout of all sub-arrays of the array. The scene is scanned onto the array for simultaneous imaging of the terrain in many spectral bands. Time Delay and Integrate (TDI) techniques are used as a clocking mechanism within the sub-arrays to increase the signal to noise ratio (SNR) of the detected image. Additionally, the TDI length (i.e., number of rows of integration during the exposure) within each sub-array is adjustable to optimize and normalize the response of the photosensitive substrate to each spectral band. The array provides for parallel and simultaneous readout of each sub-array to increase the collection rate of the spectral imagery. All of these features serve to provide a substantial improvement in the area coverage of a hyperspectral imaging system while at the same time increasing the SNR of the detected spectral image.

The present invention provides an apparatus for photodynamic therapy and fluorescence detection, in which a combined light source is provided to illuminate an object body and a multispectral fluorescence-reflectance image is provided to reproduce various and complex spectral images for an object tissue, thus performing effective photodynamic therapy for various diseases both outside and inside of the body.For this purpose, the present invention provides an apparatus for photodynamic therapy and photodetection, which provides illumination with light of various wavelengths and multispectral images, the apparatus including: an optical imaging system producing an image of an object tissue and transmitting the image to a naked eye or an imaging device; a combined light source including a plurality of coherent and non-coherent light sources and a light guide guiding incident light emitted from the light sources; a multispectral imaging system including at least one image sensor; and a computer system outputting an image of the object tissue to the outside. Thus, the apparatus for photodynamic therapy and photodetection of the present invention can effectively perform the photodynamic therapy and photodetection by means of the combined light source capable of irradiating light having various spectral components to an object tissue and the multispectral imaging system capable of obtaining images from several spectral portions for these various spectral ranges at the same time, thus improving the accuracy of diagnosis and efficiency of the photodynamic therapy.