Görüntü işleme ile piksel çakıştırmaya dayalı mesafe tespit yöntemi

Today, image sensors (cameras) are widely used for image analysis (classification, segmentation, etc.) and synthesis (object detection, tracking, distance detection, etc.). In the study, it was aimed to lay the theoretical foundations for the development of an image processing-based measuring device (image meter), which can be used for laser-meter, lidar meter, radar and similar industrial purposes. For this purpose, image processing-based distance detection methods in the literature were examined. In these methods, the main factors that negatively affect the performance were determined, and a new method that was not affected by these factors was proposed. First, the theoretical foundations of the proposed method were laid. These theoretical foundations were created from hardware and software components. In the study, these theoretical foundations were given and the designs of hardware and software components were made. Thus, an electromechanical system with hardware and software components was developed. In the developed system, one fixed and one moving camera were used. A linear motor with 0.01mm/step sensitivity and an angular motor operating with 526.375 steps/revolution sensitivity were used for the motion camera. In the study, measurements were carried out for targets with a distance of 1-800 m. Target points with mm precision with Total Station were accepted as reference. A laser meter was used as a second measuring device. Both theoretical calculations and experimental measurements were made with the image meter for the determined target distances. Using the theoretical and experimental measurement results, the measurement errors of the device were determined. When the experimental measurement results were compared with the theoretical calculations for 19 different distances, the average error of the experimental results was 1.2%. In addition, the experimental measurement results of the device were compared with the laser meter measurement results. As a result of the measurements of the image meter, the success rate was below 2%. The mean error was found to be 1.244% for 19 separate measurement points at certain distances with the image meter. For the same distances, this rate was 1.425% in measurements with laser meters.