A dual-channel fusion system of visual and infrared images based on color transfer The increasing availability and deployment of imaging sensors operating in multiple spectrums has led to a large research effort in image fusion, resulting in a plethora of pixel-level image fusion algorithms. However, most of these algorithms have gray or false color fusion results which are not adapt to human vision. Transfer color from a day-time reference image to get natural color fusion result is an effective way to solve this problem, but the computation cost of color transfer is expensive and can’t meet the request of real-time image processing. We developed a dual-channel infrared and visual images fusion system based on TMS320DM642 digital signal processing chip. The system is divided into image acquisition and registration unit, image fusion processing unit, system control unit and image fusion result out-put unit. The image registration of dual-channel images is realized by combining hardware and software methods in the system. False color image fusion algorithm in RGB color space is used to get R-G fused image, then the system chooses a reference image to transfer color to the fusion result. A color lookup table based on statistical properties of images is proposed to solve the complexity computation problem in color transfer. The mapping calculation between the standard lookup table and the improved color lookup table is simple and only once for a fixed scene. The real-time fusion and natural colorization of infrared and visual images are realized by this system. The experimental result shows that the color-transferred images have a natural color perception to human eyes, and can highlight the targets effectively with clear background details. Human observers with this system will be able to interpret the image better and faster, thereby improving situational awareness and reducing target detection time.
Holographic display is a true three-dimensional (3D) display technology presenting all depth cues without using any
special glasses. With this technology we have developed a system of city photonics map. In the system there are three
key steps: gaining raw 3D point clouds data, Computer-generated holograms, and digital holographic display of 3D data.
In the first step, the theory of structure from motion (SfM) is applied. Then an algorithm of phase CGH with inverse
diffraction is proposed and implemented. Difference from ordinary CGH is that the diffraction fringes of 3D
reconstructed terrain data are obtained from the diffraction results by the reverse strike means. At last, 3D hologram is
displayed on the developing experimental platform and printed on the holographic silver halide plate.
A system of binocular stereo vision based on the near-infrared imaging characteristics of ordinary black-and-white CCD
(B/W CCD) camera is proposed to measure the three-dimensional profile of reverse engineering. In this system the
infrared LED is used as the marker for the active probe tool, and the near-infrared filter is fixed on the black and white
camera to simplify the background and enhance the signal-to-noise ratio of image, the compution of segmentation for the
follow-up goal can be greatly reduced.
The Paper details the principle and hardware components of the binocular vision system, then analyses and discusses the
critical technologies of 3-D profile measurement such as camera calibration, feature extraction and three-dimensional
The experimental results show that the method has many advantages, such as the great contrast images consisted of the
markers and background , processing image easily, adapting the change of illumination and better precision positioning.