Stere calculation of trees laden vehicles by image processing methods

In this study, it was proposed to use image processing techniques to perform stere calculations of tree-laden vehicles by using a computer-aided system, instead of by manual calculations, made by the weighbridge staff. Images of loaded trucks from the Kastamonu Entegre Company, and were obtained at different times of day, in different weather conditions. For our operating environment, we chose an MDF production plant weighbridge that had heavy traffic, receiving tree-laden vehicles daily. A data communication infrastructure was created for a high-resolution camera, and with the help of this camera, instant images of trucks were obtained and information about the truck was stored on a computer. Benefiting from a variety of image processing techniques, real-time software was developed. During the development of the software, a variety of pre-processing techniques were used, such as Watershed Segmentation (Beucher, 2014), Hough Line Transform, Harris Corner Detection and other image processing techniques, from the OpenCV, image-processing library, and that system was tested on previously saved truck images, and the results saved to the computer again. While making calculations, the pixel/cm calibration was designated as 521/470. The data recorded on the computer consists of the average height value (as calculated manually by the weighbridge staff (v1)) and other information, such as the average height value (as calculated by the system), stere values that were, in turn, calculated using these values. The results of measurements taken both manually and using the computer system were compared. These were also referenced with the values given in the truck shipping documents (waybill). After this analysis and comparison, it was found that a variance in values might exist between those calculated in the conventional way (manual weighbridge) and those of the computer but that the deviation in these values was limited to approximately 10% or within the accepted deviance in value. The experimental results of this study showed that, using image processing techniques, approximate stere values of tree-laden vehicles can be calculated with an acceptable deviation value.