In the modern world, the use of visual information is ubiquitous. Small-sized data analysis devices have found application in almost all areas, including security systems, industrial production, gaming, car control, etc. To increase the accuracy, fault tolerance, and decision-making speed, combined information data fields are used as input data. The article developed an approach to combining data obtained in different ranges into a single image. To improve the alignment accuracy, the paper proposes using a group of methods for the primary processing of image enhancement. The article proposes finding anchor points for a series of images used for their subsequent merging. An algorithm for calibrating a system of several cameras is presented, a description of the influence of parameters on the result of combining is given. To increase the processing speed and the selection of the processing methods' primary parameters, it is proposed to change the bitness of the image by simplifying them. The operation includes absorbing small objects and averaging the ranges of histograms of color gradients. As test data, a series of images of one object obtained in: visible (camera with a resolution of 1920 x 1080 pixels, color depth of 8 bits); far-infrared (thermal imaging images with a resolution of 320x240 pixels, grayscale); 3D model (construction is carried out based on a stereo pair with a resolution of 1024x768 pixels).
|