Our words: Forest Engineering student Ana Clara da Costa compared the accuracy and speed of Timbeter to manual measurements on sample piles of eucalyptus for academic research purposes. Her results showed that Timbeter provides a significantly faster and more accurate alternative for roundwood measurements, as well as an AI based solution to bring digital transformation to the sector.
Study by Ana Clara da Costa, University of São Paulo
A study with Timbeter was developed at Klabin S/A – Brazil, in 2021, under the guidance of Forest Engineering student Ana Clara da Costa from the University of São Paulo, Campus “Luiz de Queiroz” – Brazil, and with the contribution of other collaborators, in order to validate the accuracy of the application as an alternative to automating the measurements of log diameter and volume of eucalyptus piles.
In this work, the accuracy of the application was studied through a triple factorial experiment aimed at analyzing the influence of (i) two different devices with 8 and 48 megapixels of resolution respectively, (ii) two different alignments of the logs in the piles (end of logs with level aligned and misaligned by more than 50 centimeters) and (iii) the application and non-application of manual adjustments on the detection of diameters by Timbeter since the adjustment is a functionality offered during the measurement process with the application.
Fifty logs were sampled in each pile, six piles with the end of the logs aligned and six piles with the end of the logs misaligned. In addition, the piles sampled had logs with small diameters (greater than or equal to 4 cm), medium (greater than or equal to 12 cm) and thick (greater than or equal to 25 cm) diameters.
To compare the diameters and volumes measured with the Timbeter, the diameters of the control were measured by the conventional method using the measuring tape and the target length of the pile to calculate the cylindrical volume. The measurements with Timbeter were standardized at 5 meters away from the pile, using the diameter measurement mode and selecting the cylindrical volume. In addition, a rod was used that standardized all devices at a height of 1.25 meters and a physical reference of 2 meters in length allocated at an intermediate level at the end of the sampled logs.
An Analysis of Variance with 10% of significance was performed on the volume data and on the data of each diameter class considered, both for piles with aligned and misaligned logs.
With no significant interaction between the factors tested, the results showed that, regardless of the alignment of the logs in the piles, Timbeter can be used in devices with 8 and 48 megapixels and without the need to make manual adjustments to the detection of the application when measuring the cylindrical volume of the piles and the diameter of medium and thick logs. However, it is necessary to perform manual adjustment on the identification and measurement of the diameter of thin logs regardless of the device tested and the alignment of the logs in the piles.
The results showed Timbeter accuracy in wood measurement, compared to the conventional method, were that the functionality offered by the technology to make adjustments on the detection is a viable and intelligent alternative to achieve more assertive results when measuring the diameter of thinner logs, compared to the traditional manual method. In addition, the alternative measurement method made it possible to measure 6.4 times more logs compared to the conventional method every hour, allowing greater sampling with field measurements in a shorter time.
It is concluded, therefore, that Timbeter can be recommended to automate measurements of eucalyptus logs and piles considering the influencing factors tested in this study.