The Academic Resources Recommender System (ARReS) includes collection of Theses and Dissertations available in the Camarines Sur Polytechnic Colleges Library.
Nowadays, learning about a location’s accessibility is very crucial for everyone, especially those who intend to travel, transport goods, and open or expand their business. Hence, road accessibility opens up new market possibilities. Thus, this study, "Road Accessibility Using Satellite Image Segmentation with Deep Learning", provided results that can address road accessibility for business location viability. Quantitative research method was used for this study for it required data collection wherein researchers gathered images as their primary source in building dataset and added some images from the internet as their secondary source then directly proceed to data labeling. This research design fits this study as its specific objectives suggest the quantity of data such as the level of performance of the U-net Algorithm that is used before and after the optimization in terms of data analysis, regression, factor analysis, classification, and clustering Google Earth Pro geolocation images were collected and labeled accordingly using Labelbox as the primary dataset. The labeled dataset was used to train the Unet architecture model to classify road and building regions in the images. The said model has been evaluated using mean intersection over union to check the performance of the CNN algorithm and it has achieved 62% accuracy. Moreover, the web application prototype was also built incorporating the trained model to further test in the said model in determining the location’s business viability in terms of road accessibility. The overall result shows that the developed application was reliable in determining road accessibility using satellite image segmentation.
Janet B. Sasaluya; Caren F. Vega; Crystallene R. Lagatic; Louis Carlo SF. Fornoles
2022 Computer ScienceHarvesting the rice plant too early resulted in a higher percentage of unfilled or immature grains, lowering production and increasing grain breakage during milling. However, harvesting rice too late resulted in severe losses and breakage. This problem pushed the researchers to create a model that h...
Namirah L. Balindong; Arnold S. Sayat Jr.; Mat Andrew V. Tantiado; Rose Ann B. Sabilla; Mary Rose T. Lazaro
2022 Computer ScienceThe convolutional neural network is one of the most promising applications in computer vision. Deep learning-based classification has recently enabled the recognition of chickens from images. However, no research has been done on using the Convolutional Neural Network (CNN) Algorithm in image proces...
Hamadie Seed D. Alswaidi; Anthony B. Amilano; Felix Jose I. Awa; Marian Mildred D. Bedural
2022 Computer ScienceCoconut diseases cause production problem. In connection with this, the study aimed to gather and collect datasets of healthy coconut leaves and with coconut diseases particularly Graphiola Leaf Spot, Stigmina Leaf Spot, and Whiteflies; create a CNN model for detecting the specified coconut diseases...
Rose Ann R. Navales; Judy Ann Soliman; Marlou B. Toledo; Filbert O. Tucio
2022 Computer ScienceThe research, “Rice Paddy Moisture Content Classification using Convolutional Neural Network,” was created to assist rice farmers and dealers in classifying rice paddy to determine moisture content. Also, the researcher used an experimental research design to perform research in an objective and con...