Apples are one of the most consumed fruits, and they require efficient harvesting procedures to remains in optimal states for a longer period, especially during transportation. Therefore, automation has been adopted by many orchard operators to help in the harvesting process, which includes apple localization on the trees. The de facto sensor that is currently used for this task is the standard camera, which can capture wide view information of various apple trees from a reasonable distance. Therefore, this paper aims to produce the output mask of the apple locations on the tree automatically by using a deep semantic segmentation network. The network must be robust enough to overcome all challenges of shadow, surrounding illumination, size variations, and occlusion to produce accurate pixel-wise localization of the apples. A high-resolution deep architecture is embedded with an optimized design of group and shuffle operators (GSO) to produce the best apple segmentation network. GSO allows the network to reduce the dependency on a few sets of dominant convolutional filters by forcing each smaller group to contribute effectively to the task of extracting optimal apple features. The experimental results show that the proposed network, GSHR-Net, with two sets of group convolution applied to all layers produced the best mean intersection over union of 0.8045. The performance has been benchmarked with 11 other state-of-the-art deep semantic segmentation networks. For future work, the network performance can be increased by integrating synthetic augmented data to further optimize the training phase. Moreover, spatial and channel-based attention mechanisms can also be explored by emphasizing some strategic locations of the apples, which makes the recognition more accurate.
Deforestation remains one of the key concerning activities around the world due to commodity-driven extraction, agricultural land expansion, and urbanization. The effective and efficient monitoring of national forests using remote sensing technology is important for the early detection and mitigation of deforestation activities. Deep learning techniques have been vastly researched and applied to various remote sensing tasks, whereby fully convolutional neural networks have been commonly studied with various input band combinations for satellite imagery applications, but very little research has focused on deep networks with high-resolution representations, such as HRNet. In this study, an optimal semantic segmentation architecture based on high-resolution feature maps and an attention mechanism is proposed to label each pixel of the satellite imagery input for forest identification. The selected study areas are located in Malaysian rainforests, sampled from 2016, 2018, and 2020, downloaded using Google Earth Pro. Only a two-class problem is considered for this study, which is to classify each pixel either as forest or non-forest. HRNet is chosen as the baseline architecture, in which the hyperparameters are optimized before being embedded with an attention mechanism to help the model to focus on more critical features that are related to the forest. Several variants of the proposed methods are validated on 6120 sliced images, whereby the best performance reaches 85.58% for the mean intersection over union and 92.24% for accuracy. The benchmarking analysis also reveals that the attention-embedded high-resolution architecture outperforms U-Net, SegNet, and FC-DenseNet for both performance metrics. A qualitative analysis between the baseline and attention-based models also shows that fewer false classifications and cleaner prediction outputs can be observed in identifying the forest areas.
1 April 2020 in Journal of Physics: Conference Series
The freshness of the chicken meat will be degraded due to microbiological and chemical processes and will affect the quality of the chicken’s meat. Measurements of freshness were done by a laboratory test that usually destructively and takes a long time. In this study, a VNIR imaging system was built with a wavelength range of 400-1000 nm to determine the freshness of broiler chicken meat. The freshness of the chicken meat was analyzed by using the organoleptic and pH measurement approach. Classification using Random Forest (RF) modeling has been developed to predict the freshness of chicken meat. The freshness of chicken meat was evaluated by using the correction value of 85.5%. The Partial Least Square Regression (PLSR) algorithm was successfully used to determine the pH. The pH measurement system for fresh chicken meat was evaluated using a correlation coefficient of 0.80 and RMSE 0.16. Meanwhile, for the spoiled chicken meat, pH was measured using a correlation coefficient of 0.84 and RMSE of 0.18. Both classification and regression methods indicate that this measurement system is adequate for predicting the quality of chicken meat.
Introduction: As a tropical archipelago, Indonesia is exceptionally susceptible to climate change impacts. Since mitigation requires accurate regional climate data, a reliable model assessment is essential to address the biases and uncertainties of Global Climate Model’s (GCMs). This study evaluates and ranks 35 NASA NEX-GDDP-CMIP6 models, including their multi-model mean ensemble (ENSMEAN), on their capacity to simulate monthly precipitation climatology over Indonesia. Methods: The methodology employed MSWEP dataset as the observational reference, utilizing statistical metrics including Correlation Coefficient (CC), Normalized Standard Deviation (NSTD), Root Mean Square Deviation (RMSD) and Mean Bias (MB). A dual-scale evaluation framework was adopted, assessing the model’s spatio-temporal performance. Taylor Diagrams used to visualize model distribution, while Min-Max normalization and the Summation of Rank (SR) applied to ensure fair comparison and identify the best-performing models. Discussion: The findings demonstrate that NEX-GDDP-CMIP6 models generally capture Indonesia’s seasonal precipitation patterns in close alignment with MSWEP observations. Notably, five models that consistently identified as high performers across spatio-temporal dimensions were ACCESS-CM2, CMCC-ESM2, TaiESM1, MRI-ESM2-0 and CESM2-WACCM. Specifically, ACCESS-CM2 showed the highest temporal accuracy, while TaiESM1 demonstrated the strongest spatial accuracy. ENSMEAN ranked seventh across spatio-temporal dimensions, proving its capability of reducing errors and enhancing simulation reliability. Despite the model’s overall accuracy, systematic biases persists, such as a “February dip” that underestimates peak wet-season precipitation and a tendency to overestimate precipitation during the dry season. These discrepancies suggests that simulating precipitation interactions among monsoon dynamics, topography and land-sea contrast remain challenging in Indonesia Maritime Continent. This study offers a benchmark for GCMs selection and underscores the need for improved regional models to support climate adaptation and hydrological policymaking.
19 November 2024 in Journal of Radioanalytical and Nuclear Chemistry