Methodology enhancement for program determination of roundwood batch volume to improve the accuracy of its application

The problem of improving the accuracy and automation of round timber accounting is still quite urgent. In domestic and foreign practice various manual piece-by-piece methods of volume measurement are used, both with and without taking bark into account. These include the method of end sections, the middle section method, and the truncated cone method. At the same time automated methods, mainly sectional methods, are finding wide application. . Group methods are also being developed. In addition to the manual stacking method, accounting practice uses weight methods, construction of 2-D and 3-D models on the basis of photographic and laser surveying. However, quite often there is no normative regulation of timber accounting issues: which method and when to use, what measurement errors are acceptable and how to deal with discrepancies. To date, application software packages based on machine vision have been developed to significantly reduce the labor intensity of operations to measure the volume of round timber in a dense stack or on a timber truck. The article shows that in software-based volume determination of coniferous and hardwood timber the direct estimates are overestimated as compared with the control ones. The distribution of relative deviations of estimates from the reference values does not adhere to the normal distribution law. As a result of analyzing a sample of hardwood timber batches, a regression model was obtained to adjust the software estimates of the batch volume with regard to the volume, average diameter, and number of logs. This model allows improving the accuracy of program estimation compared to the control one. The work was performed within the framework of the scientific school «Innovative developments in the field of logging industry and forestry» of the Arctic State Agrotechnological University.