The use of new technologies is already a reality, and the correlation of drone overflight data with quality control data is an important tool to aid in decision making and represents an evolution in the conventional quality system.

The birth of quality control dates back to the 17th century, when ways were sought to ensure the standards of manufactured products. At that time, the focus was on finding defective products that did not meet acceptable standards, acting in a corrective and not preventive way as we know it. Thus, the action taken was the removal of defective products. This implies loss of time and productive resources, financial impact, since it was spent to produce such products, and, of course, reduced efficiency, and, most of the time, rework occurred to replace the problematic products.

With the increase in mass production, statistical control begins to emerge, with concepts and analysis by sampling, gaining strength the idea that quality control should have managerial contours and be used as a differentiation for gains in competitiveness, being used as a strategy differentiation between companies.

But what does the forest area have to do with it? Now, as mentioned earlier, with mass production, there is a need to control the quality of the products and operations, and, in this sense, the forest area fits perfectly: in 9 years, we almost doubled the planted area; to be more precise, we increased it by 45%.



According to the Brazilian Tree Industry, the number of trees planted per day reaches 1 million. If we consider the beginning of tax incentives in the 1970s, going from a rudimentary and low- tech activity, the planted forest sector reaches the position of an important activity in the Brazilian trade balance, launching the country as one of the main players in the sector, reaching records productivity and consolidating Brazil as one of the most attractive countries to plant forests, with climate, soil and conditions favorable to growth.

With the evolution of technology and the digital world increasingly present in everyday life, the forestry area had to adapt and insert, in its routine, new means of handling the mass of information generated from operations. As a result, the term Forest 4.0, as well as the concept of digitizing processes and automating a large part of activities, has become a reality.

When talking about mechanization, the range of possibilities and tools available on the market is immense, however, the great success in this world, or the real gain, is how to transform this technology into a competitive differential. Translating this universe of variables and data into information that adds value and brings meaning is the great challenge in the forest sector.

In this sense, traditional quality control, with the release of parcels and treatment of deviations, gains help with the aid of drones and imaging. All of this is important, but it is of no use if there is no return with well-defined action plans related to the strategy and perceived as an important decision-making tool by the operational area.

At this point, it is worth mentioning the pine industry, most of which is located in sloping regions and, consequently, still involves many manual activities, especially in silviculture (soil preparation, planting, silvicultural treatments), which, for itself, increases variability in operations. When the activity is mechanized, it is possible to implement sensors and monitor several parameters, but when the activity is manual, this task becomes more complex and, in most cases, the human factor significantly influences the result achieved.

Still related to pine cultivation, the production cycle is longer. Whereas, in eucalyptus cultivation, the average is around 7 years, in pine, we are talking about 15 to 30 years, and quality control, as a rule, evaluates a window of 30 to 45 days post-planting. This assessment is necessary for scheduling replanting, for example. However, when it comes to such long cycles, this evaluation window, which is the same applied to eucalyptus, may seem premature, or even precocious.



In this context, image analysis becomes an important ally in the management of forest assets, moving from a view in portions to a view of the entire field, allowing the manager to identify points that are below the indicators and take corrective actions or correct the training plan for future rotations. Imaging provides important information about the quality of the stand, such as percentage of survival, vigor (square meters per crown), weed infestation, competition, and they have a direct correlation with the data collected in conventional quality control.

Another point that is being studied, and the images and data are providing subsidy, is in relation to late mortality, that is, after the period of 45 days recommended for replanting, it is possible to identify how much continued to die, or how the population tendency until the end of the cycle, preserving the phytosanitary conditions of plagues and fires.

This result is obtained, on average, 1.5 years after planting and serves as confirmation for the silvicultural practices used, or shows what should be improved or corrected in future cycles. In cases where it is possible to identify failures, an entire investigation process is initiated, consulting the activities carried out in the field and serving for adjustments in future deployments.

This post-planting time is interesting, as it is a compromise between conventional quality assessments (30 days) and the total crop cycle time, providing a qualitative idea of how the stand is, post-implanted. A practical example obtained with this information is crossing with the management data in the overflyed plots is in the application of herbicide, in which it was identified that, with three post-planting applications, the best survival result was achieved, and this should be a practice pursued during the culture cycle.

Thinking about the maximum use of technology, the ideal would be the total replacement of conventional quality control by imaging, but caution is needed, the algorithms for identifying survival in pines can only identify seedlings after 1.5 years of age; this is a sensitive point, as there has been great progress in identifying eucalyptus seedlings, making it possible, even after 30 days, to carry out a survival survey. However, in pine, this is still not possible, due to the slower growth of the crop. Algorithms still have difficulty identifying young seedlings, often mistaking them for local vegetation.

Another point that makes total replacement impossible is the fact that there are still quality parameters that imaging cannot capture, such as herbicide drift, subsoiling depth, pitting quality, which are dependent on in locu evaluation per person empowered. Spatial analysis, or imaging , still presents points for improvement and necessary advances, but it already constitutes an important tool that, correlated to the traditional quality survey, provides important subsidies for decision making, helping the manager and giving a more concrete notion about the population quality.