Field-deployable coffee yield estimation from mobile phone images using branch segmentation and occlusion correction

While coffee beans have been studied for many years, researchers are showing a growing interest in coffee leaves and by-products, but little information is currently available on coffee species other than Coffea arabica and Coffea canephora. The aim of this work was to perform a targeted and untargeted metabolomics study on Coffea arabica, Coffea canephora and Coffea anthonyi. The application of the recent high-resolution mass spectrometry-based metabolomics tools allowed us to gain a clear overview of the main differences among the coffee species. The results showed that the leaves and fruits of Coffea anthonyi had a different metabolite profile when compared to the two other species. In Coffea anthonyi, caffeine levels were found in lower concentrations while caffeoylquinic acid and mangiferin-related compounds were found in higher concentrations. A large number of specialized metabolites can be found in Coffea anthonyi tissues, making this species a valid candidate for innovative healthcare products made with coffee extracts.

Coffee is a beverage widely consumed in the world. The coffee species most commercialized worldwide are Arabica (Coffea arabica) and Robusta (Coffea canephora). Roasted coffee beans are the most used, but coffee leaves are also consumed as infusion in several countries for traditional medicinal purposes. They contain several interesting phenolic antioxidant compounds mainly belonging to chlorogenic acids (CGAs). In the present work, a liquid chromatography-electrochemical detection (LC-EC) method was developed for the determination of three main chlorogenic acid isomers, namely 3-, 4-, and 5-caffeoylquinic acids (CQA), in coffee leaves aqueous extracts. Samples from eight coffee species, namely; Coffea arabica, Coffea canephora, Coffea liberica, Coffea humilis, Coffea mannii, Coffea charrieriana, Coffea anthonyi, and Coffea liberica var. liberica, were grown and collected in tropical greenhouses. Linearity of the calibration graphs was observed in the range from the limit of quantification to 1.0 × 10−5 M, with R2 equal to 99.9% in all cases. High sensitivity was achieved with a limit of detection of 1.0 × 10−8 M for 3-CQA and 5-CQA (i.e., 3.5 µg/L) and 2.0 × 10−8 M for 4-CQA (i.e., 7.1 µg/L). The chromatographic profile of the samples harvested for each Coffea species was studied comparatively. Obtained raw data were pretreated for baseline variations and shifts in retention times between the chromatographic profiles. Principal Component Analysis (PCA) was applied to the pretreated data. According to the results, three clusters of Coffea species were found. In the water sample extracts, 5-CQA appeared to be the major isomer, and some species contained a very low amount of CQAs. Fluctuations were observed depending on the Coffea species and harvesting period. Significant differences between January and July were noticed regarding CQAs content. The species with the best CQAs/caffeine ratio was identified. The LC-EC data were validated by liquid chromatography-high resolution mass spectrometry (LC-HRMS).

Brazilian coffee production relies on the cultivation of Coffea arabica and Coffea canephora. Climate change has been responsible for the decreasing yield of the crops in the country yet the associated microbial community can mitigate these effects by improving plant growth and defense. Although some studies have tried to describe the microorganisms associated with these Coffea species, a study that compares the microbiome on a wider spatial scale is needed for a better understanding of the terroir of each coffee planting region. Therefore, our aim was to evaluate the microbial communities harbored in soils and fruits of these Coffea species in four Brazilian floristic domains (Amazon, Atlantic Forest Caatinga, and Cerrado). One hundred and eight samples (90 of soil and 90 of fruits) were used in the extraction and sequencing of the fungal and bacterial DNA. We detected more than 1000 and 500 bacterial and fungal genera, respectively. Some soil microbial taxa were more closely related to one coffee species than the other species. Bacillus bataviensis tends to occur more in arid soils from the Caatinga, while the fungus Saitozyma sp. was more related to soils cultivated with C. arabica. Thus, the species and the planting region (floristic domain) of coffee affect the microbial composition associated with this crop. This study is the first to report microbial communities associated with coffee produced in four floristic domains that include sites in eight Brazilian states. Data generated by DNA sequencing provides new insights into microbial roles and their potential for the developing more sustainable coffee management, such as the production of biofertilizers and starter culture for fermentation of coffee cherries.

Rice floret number per unit area as one of the key yield structure parameters is directly related to the final yield of rice. Previous studies paid little attention to the effect of the variations in vegetation indices (VIs) caused by rice flowering on rice yield estimation. Unmanned aerial vehicles (UAV) equipped with hyperspectral cameras can provide high spatial and temporal resolution remote sensing data about the rice canopy, providing possibilities for flowering monitoring. In this study, two consecutive years of rice field experiments were conducted to explore the performance of florescence spectral information in improving the accuracy of VIs-based models for yield estimates. First, the florescence ratio reflectance and florescence difference reflectance, as well as their first derivative reflectance, were defined and then their correlations with rice yield were evaluated. It was found that the florescence spectral information at the seventh day of rice flowering showed the highest correlation with the yield. The sensitive bands to yield were centered at 590 nm, 690 nm and 736 nm–748 nm, 760 nm–768 nm for the first derivative florescence difference reflectance, and 704 nm–760 nm for the first derivative florescence ratio reflectance. The florescence ratio index (FRI) and florescence difference index (FDI) were developed and their abilities to improve the estimation accuracy of models basing on vegetation indices at single-, two- and three-growth stages were tested. With the introduction of florescence spectral information, the single-growth VI-based model produced the most obvious improvement in estimation accuracy, with the coefficient of determination (R2) increasing from 0.748 to 0.799, and the mean absolute percentage error (MAPE) and the root mean squared error (RMSE) decreasing by 11.8% and 10.7%, respectively. Optimized by flowering information, the two-growth stage VIs-based model gave the best performance (R2 = 0.869, MAPE = 3.98%, RMSE = 396.02 kg/ha). These results showed that introducing florescence spectral information at the flowering stage into conventional VIs-based yield estimation models is helpful in improving rice yield estimation accuracy. The usefulness of florescence spectral information for yield estimation provides a new idea for the further development and improvement of the crop yield estimation method.

This study developed a rapid rice yield estimation workflow and customized yield prediction model by integrating remote sensing and meteorological data with machine learning (ML). Several issues need to be addressed while developing a crop yield estimation model, including data quality issues, data processing issues, selecting a suitable machine learning model that can learn from few available time-series data, and understanding the non-linear relationship between historical crop yield and remote sensing and meteorological factors. This study applied a series of data processing techniques and a customized ML model to improve the accuracy of crop yield estimation at the district level in Nepal. It was found that remote sensing-derived NDVI product alone was not sufficient for accurate estimation of crop yield. After incorporating other meteorological variables into the ML models, estimation accuracy improved dramatically. Along with NDVI, the meteorological variables of rainfall, soil moisture, and evapotranspiration also exhibited a strong association with rice yield. This study also found that stacking multiple tree-based regression models together could achieve better accuracy than benchmark linear regression or standalone ML models. Due to the unique and distinct physio-geographical setting of each district, a variation in estimation accuracy from district to district could be observed. Our data processing and ML model workflow achieved an average of 92% accuracy of yield estimation with RMSE 328.06 kg/ha and MAE 317.21 kg/ha. This methodological workflow can be replicated in other study areas and the results can help the local authorities and stakeholders understand the factors affecting crop yields as well as estimating crop yield before harvesting season to ensure food security and sustainability.

The coffee berry borer (Hypothenemus hampei) is the most damaging insect pest of global coffee production. Despite its importance, our knowledge on the insect’s natural habitat, range, and wild host species remains poorly known. Using archival sources (mainly herbaria but also other museum collections), we surveyed 18,667 predominantly wild-collected herbarium specimens mostly from Africa, Madagascar, and Asia for coffee berry borer occurrence. A total of 72 incidences were confirmed for presence of the coffee berry borer, with identifications assisted by micro-CT for SEM. Of the 72 positive infestations, all were from tropical African coffee (Coffea) species, of which 32 were from wild (non-cultivated) plants. Of the 32 wild occurrences, 30 were found in C. canephora (robusta coffee), 1 in C. liberica (Liberica coffee), and 1 in C. arabica (Arabica coffee). Our herbarium survey confirms literature and anecdotal reports that the coffee berry borer is indigenous to tropical Africa, and that coffee species, and particularly robusta coffee, are important hosts. We identify the wetter type of Guineo-Congolian forest as either the preferred or exclusive native habitat of the coffee berry borer. Other than coffee, we find no evidence of other naturally occurring hosts. Characters of infestation (e.g., hole position on coffee fruits) infers a certain degree of specificity between the coffee berry borer and its host.

A coffee crop may consist of up to 1/5 of defective beans and finding a suitable destination for this material is economically interesting. Many coffee industries collect the selections - material containing coffee defects - and blend them with non-defective coffee fruits in specific proportions to obtain a marketable product. Studies on the composition of selections are scarce. Hydro- and liposoluble bioactive compounds were determined in five types of roasted and ground selections of coffee Arabica and in healthy Arabica and Robusta coffee species throughout an optimized HPLC-UV/Vis-MS-based method. Nicotinic acid and 4-CQA were not detected. Black and sour beans seem to increase the level of caffeine (variation from 1.3 to 2.4 g 100 g-1 of sample) in the selections. The occurrence of defects decreases the level of 5-CQA, the main representative chlorogenic acid. Trigonelline content is high in Arabica coffee, and the presence of defects does not promote a clear variation in its amount. Kahweol concentration (~74.6 to 76.9 mg g-1 of oil) was practicallythe same up the Arabica sample set; this diterpene was absent in Robusta coffee. Cafestol (variation between 12.4 and 16.4 mg g-1 of oil) is a good quality indicator. Kahweol and 16-O-methyl cafestol are species indicators, and caffeine can point out the species of coffee. PCA revealed that sour beans were associated with the presence of kahweol, while cafestol and trigonelline were correlated to the occurrence of coffee skin. The higher the proportion of black beans, the more balanced the contents of water-soluble and liposoluble compounds. Key words: PVA; Coffee harnessing; HPLC-based method; Selection of coffee; Water- and fat-soluble compounds.

In this paper, we compared five crack detection algorithms using terrestrial laser scanner (TLS) point clouds. The methods are developed based on common point cloud processing knowledge in along- and across-track profiles, surface fitting or local pointwise features, with or without machine learning. The crack area and volume were calculated from the crack points detected by the algorithms. The completeness, correctness, and F1 score of each algorithm were computed against manually collected references. Ten 1-m-by-3.5-m plots containing 75 distresses of six distress types (depression, disintegration, pothole, longitudinal, transverse, and alligator cracks) were selected to explain variability of distresses from a 3-km-long-road. For crack detection at plot level, the best algorithm achieved a completeness of up to 0.844, a correctness of up to 0.853, and an F1 score of up to 0.849. The best algorithm’s overall (ten plots combined) completeness, correctness, and F1 score were 0.642, 0.735, and 0.685 respectively. For the crack area estimation, the overall mean absolute percentage errors (MAPE) of the two best algorithms were 19.8% and 20.3%. In the crack volume estimation, the two best algorithms resulted in 19.3% and 14.5% MAPE. When the plots were grouped based on crack detection complexity, in the ‘easy’ category, the best algorithm reached a crack area estimation MAPE of 8.9%, while for crack volume estimation, the MAPE obtained from the best algorithm was 0.7%.

Accurately estimating regional‐scale crop yields is substantial in determining current agricultural production performance and effective agricultural land management. The Yuncheng Basin is an important grain‐producing area in the Shanxi Province. This paper used Sentinel 2A with a spatial resolution of 10 m and MODIS with a temporal resolution of 1 d in 2020. The spatial and temporal nonlocal filter‐based fusion model (STNLFFM) was used to obtain fused data with a spatial resolution of 10 m and a temporal resolution of 1 d, combined with the Carnegie–Ames–Stanford Approach (CASA) and light‐use efficiency model to achieve summer maize (Zea mays L.) yield estimation. The results showed that the fused normalized difference vegetation index (NDVI) could inherit the spatial Sentinel‐2A NDVI details and express the spatial differences between smaller features more effectively. The STNLFFM NDVI curve was consistent with the actual summer maize growth condition, which accurately reflects the NDVI trend and local abrupt change information during the summer maize growth period. Moreover, the fused NDVI was influenced by topographic differences and artificial irrigation factors, whereas the summer maize yield in mountainous and plateau areas of the Yuncheng Basin was <5,000 kg ha−1 and those in the alluvial plain of the Sushui River reached 8,000 kg ha−1. The accuracy of the yield estimation model constructed based on STNLFFM NDVI (mean absolute percentage error [MAPE] = 5.47%, −13.74% ≤ relative error [RE] ≤0.12%) was significantly higher than that of the model based on MODIS NDVI (MAPE = 15.65%, −19.67% ≤ RE ≤ 20.88%), indicating that the use of spatio‐temporal fusion technology can effectively improve the summer maize yield estimation accuracy.

Influence of roasting levels on chemical composition and sensory quality of Arabica and Robusta coffee: A comparative study

Many plant species possess structures on their leaves that often harbour predatory or fungivorous mites. These so‐called domatia are thought to mediate a mutualistic interaction; the plant gains protection because mites decimate plant pathogenic fungi or herbivores, whereas the mites find shelter in the domatia. We tested this hypothesis using two species of coffee (Coffea spp.) plants that posses domatia consisting of small cavities at the underside of the leaves, and which often harbour mites. We assessed densities of domatia, of the predatory mite Iphiseiodes zuluagai Denmark and Muma (Acari: Phytoseiidae) and of herbivorous mites Oligonychus ilicis (McGregor) (Acari: Tetranychidae) and Brevipalpus phoenicis (Geijskes) (Acari: Tenuipalpidae) on Coffea arabica L. (Rubiaceae) and Coffea canephora Pierre in the field. Over a period of 50 days, C. arabica harboured on average 7.5 times more predatory mites and 0.4–0.66 fewer prey mites than C. canephora. Hence, the higher density of predatory mites on C. arabica could not be explained by higher densities of prey. However, the density of domatia on C. arabica was on average 1.65 times higher than on C. canephora, and within each species, leaves with higher densities of domatia also harboured more predators. This suggests a positive effect of domatia on predatory mites. In the laboratory, survival of adult female predatory mites on leaves of C. arabica with open domatia was indeed significantly higher than on leaves with closed domatia. Hence, predatory mites benefited from the domatia. However, plants with higher densities of domatia did not harbour fewer herbivores. Taken together, our study only provides partial evidence for a mutualistic interaction between coffee plants and predatory mites, mediated by domatia.

Estimating maize and cotton yield in southeastern Turkey with integrated use of satellite images, meteorological data and digital photographs

The main characteristics of the large number of coffee species are differences in aroma and caffeine content. Labeled blends of Coffea arabica (C. arabica) and Coffea canephora (C. canephora) are common to broaden the flavor profile or enhance the stimulating effect of the beverage. New emerging species such as Coffea liberica (C. liberica) further increase the variability in blends. However, significant price differences between coffee species increase the risk of unlabeled blends and thus influence food quality and safety for consumers. In this study, a prototypic hyphenation of trapped headspace-gas chromatography-ion mobility spectrometry-quadrupole mass spectrometry (THS-GC-IMS-QMS) was used for the detection of characteristic compounds of C. arabica, C. canephora, and C. liberica in green and roasted coffee samples. For the discrimination of coffee species with IMS data, multivariate resolution with multivariate curve resolution–alternating least squares (MCR-ALS) prior to partial least squares–discriminant analysis (PLS-DA) was evaluated. With this approach, the classification accuracy, as well as sensitivity and specificity, of the PLS-DA model was significantly improved from an overall accuracy of 87% without prior feature selection to 92%. As MCR-ALS preserves the physical and chemical properties of the original data, characteristic features were determined for subsequent substance identification. The simultaneously generated QMS data allowed for partial annotation of the characteristic volatile organic compounds (VOC) of roasted coffee.

It is well known that coffee is one of the major exports of the East African countries. Uganda chiefly produces the robusta strains, but Kenya and Tanzania sell mostly arabica coffee, the quality of that grown in Kenya being comparable to the best from Jamaica and Costa Rica. The homeland of the arabica species would appear to be the highlands of Ethiopia, and varieties of it, with possibly other species of coffee as well, may have been growing wild in certain regions of East Africa from the earliest times. Commercial cultivation in Kenya and Tanzania, however, began relatively recently, in the late nineteenth and early twentieth centuries. The origins of the arabica coffee industry of today date back no earlier than the agricultural experiments of a group of French Roman Catholic missionaries.

Arabica and Robusta coffee are the two main coffee tree species that are cultivated around the world. Premium Arabica coffee is rich in flavour and aroma. Meanwhile, the cheaper Robusta coffee has a bitter taste, a muddy odour, and a weaker aroma and flavour. This study aimed to improve the quality of Robusta coffee through cellulase and protease assisted germination. Germination results in producing flavour precursors and can be accelerated by degrading the endosperm of coffee beans through the activity of cellulase and protease. Quality parameters were based on morphological study (scanning electron microscope [SEM]), flavour profile analysis (gas chromatography-mass spectrometry [GC-MS]) and sensory evaluation. Results indicated that sugar-and enzyme-added germinated Robusta (R5) has similar morphological characteristics and pH values to Arabica. Flavour profile analysis indicated that 2-Furanmethanol of treated Robusta, which contributes to bitter taste, was decreased compared to untreated Robusta (control). In sensory evaluation, the final score of sensory properties of germinated Robusta (R2), enzyme-added Robusta (R3), and sugar-and enzyme-added germinated (R5) was improved, which indicated that there was no significant difference (p>0.05) compared to Arabica. These results showed that the proposed additional processing method of green coffee beans in this study was able to improve the quality of cheaper Robusta coffee which has some similar properties to premium Arabica.