https://www.alice.cnptia.embrapa.br/alice/handle/item/386 Artigo em anais de congresso (CNPUV) 2026-08-04T09:44:43Z https://www.alice.cnptia.embrapa.br/alice/handle/doc/1188861 Título: Automated detection of european canker (Neonectria ditissima) in apple trees via multispectral sensors and computer vision. Autoria: ALVES, S. A. M.; SILVA NETO, E. C. D.; MARCHIORETTO, L. de R.; GEBLER, L. Conteúdo: European canker, caused by the fungus Neonectria ditissima, represents one of the major economic challenges for Brazilian pomiculture, severely affecting ‘Gala’ and ‘Fuji’ cultivars. The disease manifests primarily in woody tissues, such as trunks and branches, although it can also cause fruit rot during the pre-harvest stage. Infection occurs obligatorily through wounds, whether natural (leaf scars) or resulting from management practices (pruning and harvesting). Effective disease management depends on early detection and the immediate elimination of inoculum sources. However, conventional manual inspection is time-consuming, costly, and limited in identifying incipient lesions or those hidden by foliage. In this context, this study aims to develop faster and more efficient detection methods through an automated system based on image processing. To enable this system, an image database of apple tree branches was established under field conditions, monitoring the stages of disease development associated with pruning wounds. The experimental design consisted of: an Inoculated Group, comprising 50 branches (across ten plants) subjected to pruning cuts and inoculation with a N. ditissima fungal suspension; and a Control Group, with 50 branches subjected only to cutting and treated with distilled water. Image captures were performed at approximately seven-day intervals using three distinct sensors: conventional RGB, multispectral RedEdge (725 nm), and RG-NIR. A sample of 132 images was annotated with segmentation masks and used to train a YOLOv8nseg model to assess the feasibility in finding European canker on branches. The database obtained from this experiment comprises more than 4,000 images. Preliminary results indicate that physiological changes resulting from infection generate detectable contrasts in the images, especially at wavelengths above 700 nm. This spectral signature highlights the potential of this range for the early identification of diseased branches. The generated database will allow for detailed spectral analysis and the training of convolutional neural networks (CNNs) for the automated recognition of infection patterns. The preliminary training showed promising results, achieving mAP50-95(M) of 0.63, Precision(M) of 0.89, and recall(M) of 0.91, even with a relatively low number of annotated images. Future stages will involve testing different segmentation models, the development of equipment for the detection, quantification, and geospatial localization of diseased branches in the orchard. Such tools are fundamental for the consolidation of precision fruit farming, enabling localized interventions, rapid diagnostics, and a reduction in operational costs for growers. 2026-01-01T00:00:00Z https://www.alice.cnptia.embrapa.br/alice/handle/doc/1188868 Título: AgDataBox-Map: web application for spatial analysis and management zone delineation for dynamic fruit harvest in precision agriculture. Autoria: FARIA, C. B. de; MARCHIORETTO, L. de R.; GEBLER, L.; BAZZI, C. L. Conteúdo: Yield maps are fundamental tools in precision agriculture, as they allow for the evaluation of whether corrective actions taken before and during the season achieved the intended effects by visualizing production heterogeneity within the cultivation area. However, in fruit farming, the manual nature of harvesting has historically made the generation of thematic maps problematic. In Brazil, initiatives developed by the Federal University of Technology – Paraná (UTFPR) in partnership with Embrapa resulted in national fruit harvest mapping systems, contributing to the adaptation of this technology to local conditions and expanding its application across different production chains. Nevertheless, early implementations required a partially static workflow, relying on fixed collection points within the orchard. This did not accommodate practices used in large-scale operations, where bins for unloading harvest bags are itinerant, moved by tractors and trailers. Given this context, the objective of this experiment was to evaluate methods for processing harvest data and generating thematic maps using non-fixed unloading points. The advancement toward an itinerant harvest data collection system enabled the broader application of yield maps to different fruit harvesting systems. Furthermore, the algorithm created for generating thematic maps of itinerant fruit harvesting can be applied in automated precision agriculture tools such as AgDataBox-Map (ADB-Map)—a web-based solution focused on the analysis and visualization of agricultural spatial data, with an emphasis on using yield maps for management support. The algorithm was applied in an experiment conducted in partnership with the company RASIP in apple orchards, utilizing an innovative data collection system during harvest with itinerant unloading points. Harvest workers operated along the rows carrying bags equipped with sensors capable of recording bag identification and GPS position every six seconds, among other relevant information. These bags, which were previously unloaded into fixed bins positioned within the orchard, are now unloaded into bins transported continuously by tractor-pulled trailers (itinerant bins). This shift from a static to an itinerant system required the recalculation of bin location centroids during map generation. The data obtained were processed via an algorithm developed in Python, to be implemented in AgDataBox-Map in the future, enabling the generation of detailed yield maps and their use in delineating result-based management zones. The results demonstrate that the algorithm is a practical and efficient tool for transforming harvest data into useful spatial information, contributing to more precise and sustainable decisions in fruit farming. 2026-01-01T00:00:00Z https://www.alice.cnptia.embrapa.br/alice/handle/doc/1188863 Título: Spatiotemporal variability of apple tree vegetative vigor using proximal sensing. Autoria: GEBLER, L.; RUFATO, A. de R.; ABREU, J. T. DE; SPERANZA, E. A.; SESSI, A.; MARCHIORETTO, L. de R. Conteúdo: Apple production in southern Brazil often suffers from excessive vegetative growth due to the region's subtropical climate. Currently, farmers rely on empirical, subjective monitoring to make crucial management decisions regarding pruning, thinning, and the application of growth regulators. Coupled with increasing labor shortages and costs, this empirical approach often leads to operational inefficiencies and compromised fruit quality. To address this, there is a critical need for quantitative, data-driven methods to monitor vegetative vigor and generate precise spatial maps for targeted orchard management. This study investigates the use of multispectral proximal sensors to measure the Normalized Difference Vegetation Index (NDVI) as an objective indicator of plant biomass, replacing traditional empirical assessments. Linear and non-linear regression models were used to correlate NDVI with bud growth rates in two apple orchards in Vacaria-RS across six growing seasons. The analysis incorporated various time windows (ranging from 5 to 11 days) alongside meteorological variables such as accumulated precipitation, solar radiation, and thermal sum. The results indicate that the optimal periods for data collection to accurately identify vegetative vigor variability are windows of 7 to 11 days, specifically occurring between 55 and 67 days after dormancy break. Identifying these crucial periods enables the generation of accurate temporal vigor maps and the delineation of specific management zones. Ultimately, this methodology allows producers to objectively adjust the intensity of vigor control practices according to specific zones, optimizing resource allocation, reducing labor dependency, and ensuring optimal fruit yield and quality. 2026-01-01T00:00:00Z https://www.alice.cnptia.embrapa.br/alice/handle/doc/1188858 Título: Profitability mapping in precision agriculture: na economic-spatial model applied to soybean farming in northwestern Mato Grosso. Autoria: MACEDO, W. N.; GEBLER, L. Conteúdo: Precision agriculture generates large volumes of spatial data; however, their integration into economic decision models and financial planning still presents a significant gap. This paper presents and validates a parameterized economic simulation tool built in an Excel spreadsheet to support decision-making by rural producers in Precision Agriculture (PA) systems. The tool integrates historical production cost data from IMEA/Senar-MT (crop years 2021/22 to 2025/26) with parameters from productive stability zones (ZAE, ZME, ZBE, and Variable/Unstable), calculating Gross Margin, Break-Even Point, and Safety Margin by zone and by crop year, under both Fixed Rate (FR) and Variable Rate (VR) management. The difference between margins constitutes the Spatial Information Value (SIV), an economic indicator of the return from sitespecific management. The structure comprises seven interdependent analytical modules: IMEA historical series, zone analysis under FR, zone analysis under VR, weighted FR vs. VR comparison, simulation of eight risk scenarios per zone, scenario summary matrix, and twodimensional price-productivity sensitivity analysis. Results show that total operating cost (COT) grew 38.3% over the period, the Safety Margin declined from 30.71 sc/ha to 8.07 sc/ha, and the average SIV was R$ 75.07/ha (R$ 325,313.65/crop year on the farm). The tool constitutes a practical, renewable, and low-cost instrument to support decision-making on farms adopting PA. 2026-01-01T00:00:00Z