The coffee leaf miner (Leucoptera coffeella) is a primary pest for coffee plants. The attack of this pest reduces the photosynthetic area of the leaves due to necrosis, causing premature leaf falling, decreasing the yield and the lifespan of the plant. Therefore, this study aims to analyze vegetation indices (VI) from images of healthy coffee leaves and those infested by coffee leaf miner, obtained using a multispectral camera, mainly to differentiate and detect infested areas. The study was conducted in two distinct locations: At a farm, where the camera was coupled to a remotely piloted aircraft (RPA) flying at a 3 m altitude from the soil surface; and the second location, in a greenhouse, where the images were obtained manually at a 0.5 m altitude from the support of the plant vessels, in which only healthy plants were located. For the image processing, arithmetic operations with the spectral bands were calculated using the “Raster Calculator” obtaining the indices NormNIR, Normalized Difference Vegetation Index (NDVI), Green-Red NDVI (GRNDVI), and Green NDVI (GNDVI), the values of which on average for healthy leaves were: 0.66; 0.64; 0.32, and 0.55 and for infested leaves: 0.53; 0.41; 0.06, and 0.37 respectively. The analysis concluded that healthy leaves presented higher values of VIs when compared to infested leaves. The index GRNDVI was the one that better differentiated infested leaves from the healthy ones.

Remotely piloted aircraft systems (RPASs) have gained prominence in recent decades primarily due to their versatility of application in various sectors of the economy. In the agricultural sector, they stand out for optimizing processes, contributing to improved sampling, measurements, and operational efficiency, ultimately leading to increased profitability in crop production. This technology is becoming a reality in coffee farming, an essential commodity in the global economic balance, mainly due to academic attention and applicability. This study presents a bibliometric analysis focused on using RPASs in coffee farming to structure the existing academic literature and reveal trends and insights into the research topic. For this purpose, searches were conducted over the last 20 years (2002 to 2022) in the Web of Science and Scopus scientific databases. Subsequently, bibliometric analysis was applied using Biblioshiny for Bibliometrix software in R (version 2022.07.1), with emphasis on the temporal evolution of research on the topic, performance analysis highlighting key publications, journals, researchers, institutions, countries, and the scientific mapping of co-authorship, keywords, and future trends/possibilities. The results revealed 42 publications on the topic, with the pioneering studies being the most cited. Brazilian researchers and institutions (Federal University of Lavras) have a strong presence in publications on the subject and in journals focusing on technological applications. As future trends and possibilities, the employment of technology optimizes the productivity and profitability studies of coffee farming for the timely and efficient application of aerial imaging.

Studies related to the mechanization of coffee transplanting combined with precision agriculture techniques demonstrate a need to verify the quality of the operation to optimize future production processes in the field and reduce costs. The objective of this study was to analyze the operational costs arising from a semi-mechanized coffee transplanting process using an autopilot. The study was conducted on a rural property in the municipality of Santo Antônio do Amparo—MG—where 7458 coffee seedlings were transplanted using a planting platform pulled by a tractor equipped with an autopilot and a GNSS antenna, over a period of 4 days. The date and time data of the operation, recorded every second by the autopilot, were collected and recorded in spreadsheets to assist in calculating operational costs. Two semi-mechanized transplanting scenarios were compared: one using autopilot and the other using conventional semi-mechanized transplanting. The results indicated that the hourly cost of operation with autopilot was US$2130.42 h⁻¹, while the conventional system presented US$326.03 h⁻¹. The effective operational cost was US$3975.61 ha⁻¹ for the system with autopilot and US$442.31 ha⁻¹ for the conventional system in 2020. After monetary updating to 2025, the operational costs increased to US$1845.19 ha⁻¹ and US$116.28 ha⁻¹, respectively. The investment analysis indicated an Internal Rate of Return of 89.7%, highlighting the potential return on investment in the study. Therefore, it was emphasized that even with a high investment cost, the application of autopilot is viable for improving tractor steering during operation and ensuring uniformity in the positioning of coffee seedlings.

Coffee production has become increasingly technified in order to optimize the use of inputs and the sustainable use of natural resources. In this context, one way that farmers are investing in their coffee plantations is in the use of precision agriculture techniques, termed precision coffee farming. Over the last few years, research has been conducted to facilitate the application of this technology, and sampling grids with two points per hectare have been recommended by several studies. These georeferenced demarcations in a plot are generally shaped as equidistant squares or rectangles, and the sampling points are located at the centers of these areas. Coffee farmers typically plant their crops following the level line, which greatly hinders the navigation of equidistant points within the field. Thus, the objective of this study was to develop an optimized walking route method to reduce the distance for sampling soil, leaf, and yield attributes. The experimental plots were established in 2000 at Samambaia Farm, located in Santo Antônio do Amparo, Minas Gerais, Brazil, with coffee the cultivar Acaia IAC 479-19, totaling 56.65 ha. The 111 sampling points were distributed in the land following the new method proposed in this study, and, after walking simulations using Farm Works Mapping Software, the new method was compared with the conventional method using the mean displacement between points. The new optimized walking routes method reduced the mean distance traveled to sample the points by 50.1%.

Early yield estimation is important for improving decision-making and management practices in coffee production systems. However, approaches that use aerial imagery to characterize flowering and relate it to subsequent yield remain limited. This study evaluated the potential of RGB imagery acquired by remotely piloted aircraft (RPA) combined with Deep Learning techniques to classify coffee plants according to three levels of a flowering proxy and to investigate its relationship with yield. A dataset comprising 1593 segmented RGB images was classified into Low, Medium, and High flowering levels using a custom convolutional neural network and three transfer learning architectures: VGG16, ResNet50V2, and MobileNetV2. ResNet50V2 was selected for final evaluation based on its validation performance and achieved 92.12% accuracy, 92.40% macro-precision, 91.81% macro-recall, and 92.06% macro-F1 on the test dataset. The flowering proxy showed a positive association with field-observed yield, with Spearman’s ρ = 0.701 (p < 0.001) and R² = 0.528 for the combined dataset of Arara and Mundo Novo. In consolidated out-of-sample predictions, the model achieved R² = 0.489, MAE = 1.03 kg plant⁻¹, and RMSE = 1.18 kg plant⁻¹. Performance differed between cultivars, with consolidated R² values of 0.567 for Arara and 0.329 for Mundo Novo. Spatial analysis also showed correspondence between the distribution of flowering levels derived from RGB imagery and the subsequent spatial distribution of yield. These findings indicate that RPA-based RGB imagery and Deep Learning can provide spatialized information related to coffee flowering and its productive performance, with potential applications in precision agriculture. Further evaluation across production cycles, cultivars, and environmental conditions is needed to assess the robustness and generalization of the approach.