Template-Type: ReDIF-Article 1.0 Author-Name: Raphael Berdin Author-Workplace-Name: Department of Electrical Engineering, University of the Philippines Los Baños, Laguna, Philippines Author-Name: Rob Christian Caduyac Author-Workplace-Name: Department of Electrical Engineering, University of the Philippines Los Baños, Laguna, Philippines Title: Towards interpretability: Assessment of residual networks for tomato leaf disease classification Abstract: The tomato occupies a prominent place in the Philippines' agricultural economy. However, tomato leaf diseases are challenges in tomato crop production leading to economic losses. Among the tomato leaf diseases, early blight and Septoria leaf spot are prevalent in the Philippines due to the climate. Thus, the accurate identification of diseases affecting tomato leaves is essential. Currently, a visual inspection is the primary method for diagnosing tomato leaf diseases which is time-consuming and inefficient. This study aims to develop a quantized Residual Network with convolutional 50 layer (ResNet-50) based model to classify tomato leaves as healthy or affected by Septoria leaf spot or early blight. Furthermore, to enhance the reliability of the models' classification, gradient-weighted class activation mapping (Grad-CAM) was implemented. In contrast with the visual inspection, a programmed system does not get tired and can provide consistent performance results. As a result, the original 32-bit floating point model attained an accuracy rate of 91.22%. The quantized 16-bit floating point model demonstrated comparable performance with 90.10% accuracy with a 50% reduction in the model size and inference time of 0.3942 seconds. The minimal accuracy loss of the 16-bit model relative to the 32-bit model is due to the post-training quantization. The reduction to 16-bit precision is significant for the future deployment of edge devices where resources are limited. Keywords: tomato leaf disease classification, Grad-CAM, quantization, ResNet50 Journal: Research in Agricultural Engineering Pages: 1-9 Volume: 71 Issue: 1 Year: 2025 DOI: 10.17221/67/2024-RAE File-URL: http://rae.agriculturejournals.cz/doi/10.17221/67/2024-RAE.html File-Format: text/html X-File-Ref: http://agriculturejournals.cz/RePEc/caa/references/rae-202501-0001.txt Handle: RePEc:caa:jnlrae:v:71:y:2025:i:1:id:67-2024-RAE Template-Type: ReDIF-Article 1.0 Author-Name: Andriy Anta Kacaribu Author-Workplace-Name: Doctoral Program of Agricultural Science, Postgraduate School, Universitas Syiah Kuala, Banda Aceh, Indonesia Author-Name: Darwin Darwin Author-Workplace-Name: Department of Agricultural Engineering, Faculty of Agriculture, Universitas Syiah Kuala, Banda Aceh, Indonesia Author-Name: Yuliani Aisyah Author-Workplace-Name: Department of Agricultural Product Technology, Faculty of Agriculture, Universitas Syiah Kuala, Banda Aceh, Indonesia Author-Name: Febriani Author-Workplace-Name: Department of Chemistry, Faculty of Mathematics and Natural Science, Universitas Syiah Kuala, Banda Aceh, Indonesia Title: Anaerobic bio-processing of agricultural waste for the biotechnological production of lactic acid and volatile fatty acid by landfill soil inoculums Abstract: With the increase in the world population and the ensuing surge in organic waste, effective management strategies are crucial to prevent environmental pollution. This study aims to address this challenge by utilising organic waste (OW) as the substrate for the production of lactic acid (LA) and volatile fatty acids (VFAs) through anaerobic bioprocessing. The substrates used, included grass, starch, and fruit wastes inoculated with non-sterile inoculum landfill soil (LS). The anaerobic bioconversion was performed by varying the substrate to the inoculum. The results unveil that a digester loaded with 150 g.L-1 of fruit waste, exhibits the highest concentration of LA, reaching a significance of 25 mmol.L-1. A digester fed with 100 g.L-1 starch, also manifests significant LA production (18.50 mmol.L-1). A digester, supplied with 150 g.L-1 starch waste, showcases the highest VFA (92.5 mmol.L-1). Intriguingly, the anaerobic bioprocessing of the grass substrate did not produce LA at all, yet al. the substrates showcased VFA production, albeit with fluctuating and lower concentrations. This study highlights the potential of incorporating simple sugar for enhanced LA production and starch-based substrates for increased VFA production when utilising LS as the inoculum. The anaerobic bioprocessing shows promising outcomes for the future development in sustainable waste utilisation. Keywords: bioproducts, waste conversion, biochemicals, organic acids Journal: Research in Agricultural Engineering Pages: 10-21 Volume: 71 Issue: 1 Year: 2025 DOI: 10.17221/52/2024-RAE File-URL: http://rae.agriculturejournals.cz/doi/10.17221/52/2024-RAE.html File-Format: text/html X-File-Ref: http://agriculturejournals.cz/RePEc/caa/references/rae-202501-0002.txt Handle: RePEc:caa:jnlrae:v:71:y:2025:i:1:id:52-2024-RAE Template-Type: ReDIF-Article 1.0 Author-Name: Saravanan Karuppanan Author-Workplace-Name: Dhanalakshmi Srinivasan College of Engineering and Technology, Mamallapuram, Chennai, India Author-Name: Saravanan Ramasamy Author-Workplace-Name: Centre for Water Resources, Anna University, Chennai, India Author-Name: Balaji Lakshminarayanan Author-Workplace-Name: Centre for Water Resources, Anna University, Chennai, India Author-Name: Sreemanthrarupini Nariangadu Anuthaman Author-Workplace-Name: Centre for Water Resources, Anna University, Chennai, India Title: An effective machine learning model for the estimation of reference evapotranspiration under data-limited conditions Abstract: Reference crop evapotranspiration (ETo) is a vital hydrological component influenced by various climate variables that impact the water and energy balances. It plays a crucial role in determining crop water requirements and irrigation scheduling. Despite the availability of numerous approaches for estimation, accurate and reliable ETo estimation is essential for effective irrigation water management. Therefore, this study aimed to identify the most suitable machine learning model for assessing ETo using observed daily values of limited input parameters in tropical savannah climate regions. Three machine learning models - a long short-term memory (LSTM) neural network, an artificial neural network (ANN), and support vector regression (SVM) - were developed with four different input combinations, and their performances were compared with those of locally calibrated empirical equations. The models were evaluated using statistical indicators such as the root mean square error (RMSE), coefficient of determination (R2), and the Nash-Sutcliffe efficiency (NSE). The results showed that the LSTM model, using the combination of temperature and wind speed, provided more reliable predictions with R2 values greater than 0.75 and RMSEs less than 0.63 mm.day-1 across all the considered weather stations. This study concludes that, especially under limited data conditions, the developed deep learning model improves the ETo estimation more accurately than empirical models for tropical climatic regions. Keywords: artificial neural networks, empirical equations, long short-term memory neural networks, machine learning, reference crop evapotranspiration, support vector machines Journal: Research in Agricultural Engineering Pages: 22-37 Volume: 71 Issue: 1 Year: 2025 DOI: 10.17221/101/2023-RAE File-URL: http://rae.agriculturejournals.cz/doi/10.17221/101/2023-RAE.html File-Format: text/html X-File-Ref: http://agriculturejournals.cz/RePEc/caa/references/rae-202501-0003.txt Handle: RePEc:caa:jnlrae:v:71:y:2025:i:1:id:101-2023-RAE Template-Type: ReDIF-Article 1.0 Author-Name: Sharmin Akter Author-Workplace-Name: Department of Agricultural Engineering, Gazipur Agricultural University, Gazipur, Bangladesh Author-Name: Md Mostafizar Rahman Author-Workplace-Name: Department of Agricultural Engineering, Gazipur Agricultural University, Gazipur, Bangladesh Author-Name: Rafatul Zannat Author-Workplace-Name: Department of Agricultural Engineering, Gazipur Agricultural University, Gazipur, Bangladesh Author-Name: Md Masud Rana Author-Workplace-Name: Department of Agricultural Engineering, Gazipur Agricultural University, Gazipur, Bangladesh Author-Name: Md Moinul Hossain Oliver Author-Workplace-Name: Department of Agricultural Engineering, Gazipur Agricultural University, Gazipur, Bangladesh Author-Name: Md Aslam Ali Author-Workplace-Name: Department of Agro-Processing, Gazipur Agricultural University, Gazipur, Bangladesh Title: Development of smart micro-irrigation system using Arduino Uno for okra cultivation in Bangladesh Abstract: Conventional irrigation practices result in a substantial amount of water loss with okra cultivation. Although micro-irrigation can address this issue by delivering water directly near the rootzone, it requires manual operation. These issues, however, can be resolved with the introduction of a smart micro-irrigation system. This study aims to develop a smart micro-irrigation system for okra, in conjunction with the sub-components of drip irrigation, a microcontroller, and a soil sensor. The experiment was laid out with a randomised complete block design (RCBD) having three treatments: (i) control irrigation (T1), (ii) drip irrigation (T2), and (iii) smart micro-irrigation (T3). The experimental field was irrigated based on soil moisture regimes in the crop rootzone. The plant growth, yield, and water use efficiency were assessed to evaluate the system. The results showed no significant differences among these treatments (at P < 0.05). The best water usage efficiency (15.98 kg.m-3) was observed in the T3 treatment, which also provided about 13.10% water savings compared to the conventional irrigation. This study indicates that a smart micro-irrigation system could be a promising technology for water-efficient okra cultivation. Keywords: drip irrigation, soil moisture sensor, solenoid valve, yield parameters, water use efficiency Journal: Research in Agricultural Engineering Pages: 38-49 Volume: 71 Issue: 1 Year: 2025 DOI: 10.17221/18/2024-RAE File-URL: http://rae.agriculturejournals.cz/doi/10.17221/18/2024-RAE.html File-Format: text/html X-File-Ref: http://agriculturejournals.cz/RePEc/caa/references/rae-202501-0004.txt Handle: RePEc:caa:jnlrae:v:71:y:2025:i:1:id:18-2024-RAE Template-Type: ReDIF-Article 1.0 Author-Name: Adrienn Bablena Author-Workplace-Name: Institute of Technology, Hungarian University of Agriculture and Life Sciences, Gödöllő, Hungary Author-Name: János Beke Author-Workplace-Name: Institute of Technology, Hungarian University of Agriculture and Life Sciences, Gödöllő, Hungary Author-Name: István Keppler Author-Workplace-Name: Institute of Technology, Hungarian University of Agriculture and Life Sciences, Gödöllő, Hungary Title: Particle motion in mixed flow dryers: The effect of the wall inclination angle and friction Abstract: In Europe, the weather patterns require harvested grain crops to be dried before storage to prevent significant quality loss. The uneven movement of grains inside the drying equipment is a key issue affecting the drying process, causing under- or over drying the harvested crops and thus leading to quality degradation and ultimately to financial losses. To characterise the unevenness of material flow, we introduced a dimensionless displacement ratio. This dimensionless parameter was suitable for comparing the uniformity of the material movement processes within the dryer. Using experimental investigations and numerical simulations, we determined the effect of the lamella inclination angle, the friction between the grain-wall and grain-grain on the uniformity of the flow. The linear functions approximating the quantitative relationships were determined in all the cases. Our findings indicate a significant variation in the displacement ratio ξ corresponding to different lamella inclination angles and friction values demonstrating that the discrete element modelling approach provides further opportunities for determining the optimal operating parameters of mixed flow dryers. Keywords: agricultural particulate materials, drying, DEM, particle motion, optimisation Journal: Research in Agricultural Engineering Pages: 50-59 Volume: 71 Issue: 1 Year: 2025 DOI: 10.17221/51/2024-RAE File-URL: http://rae.agriculturejournals.cz/doi/10.17221/51/2024-RAE.html File-Format: text/html X-File-Ref: http://agriculturejournals.cz/RePEc/caa/references/rae-202501-0005.txt Handle: RePEc:caa:jnlrae:v:71:y:2025:i:1:id:51-2024-RAE Template-Type: ReDIF-Article 1.0 Author-Name: Jiří Brožovský Author-Workplace-Name: Department of Vehicles and Ground Transport, Faculty of Engineering, Czech University of Life Science Prague, Prague, Czech Republic Author-Name: Veronika Hartová Author-Workplace-Name: Department of Vehicles and Ground Transport, Faculty of Engineering, Czech University of Life Science Prague, Prague, Czech Republic Author-Name: Martin Kotek Author-Workplace-Name: Department of Vehicles and Ground Transport, Faculty of Engineering, Czech University of Life Science Prague, Prague, Czech Republic Author-Name: Jan Hart Author-Workplace-Name: Department of Vehicles and Ground Transport, Faculty of Engineering, Czech University of Life Science Prague, Prague, Czech Republic Author-Name: Jitka Kumhálová Author-Workplace-Name: Department of Vehicles and Ground Transport, Faculty of Engineering, Czech University of Life Science Prague, Prague, Czech Republic Title: Use of thermal imaging camera for wild animal detection along roads Abstract: Vehicle collisions with wild animals are a common problem on roads, having a significant impact on road safety and wildlife populations. Collisions with wild animals are one of the most frequent road accidents. According to police statistics, there were nearly 16 000 road accidents caused by collisions with animals in the Czech Republic in 2019. Collisions with deer are the most common. There are several technologies and measures that can help reduce the risk of a vehicle colliding with a wild animal. One of the technologies used is a night vision system based on infrared spectrum sensing. This technology is slowly becoming part of the equipment of, in particular, premium car brands dueto its high cost. This paper tested a low-cost solution using a commercially available thermal imaging camera and found a substantial reduction in the time to detect wild animals along the road, namely in the order of seconds. Keywords: night vision, active vehicle safety, systems for preventing traffic accidents, traffic accident Journal: Research in Agricultural Engineering Pages: 60-67 Volume: 71 Issue: 1 Year: 2025 DOI: 10.17221/88/2024-RAE File-URL: http://rae.agriculturejournals.cz/doi/10.17221/88/2024-RAE.html File-Format: text/html X-File-Ref: http://agriculturejournals.cz/RePEc/caa/references/rae-202501-0006.txt Handle: RePEc:caa:jnlrae:v:71:y:2025:i:1:id:88-2024-RAE