Soil texture analysis is vital in agricultural management due to its influence on crop growth and yield. Defined by the proportions of clay, sand, and silt particles, soil texture affects properties like aeration, water-holding capacity, and nutrient retention, all crucial for plant development. The OBJECTIVES: This study aims to design a Genetic-Based Neural Network (GBNN) for accurate soil texture analysis, particularly for soils with similar structures but different compositions. It also aims to collect environmental impact data through soil sensors to enhance the understanding of soil texture.METHODS: The methodology involves developing a GBNN, leveraging genetic algorithms to group homogeneous particles, thus improving texture classification. This approach addresses the shortcomings of previous deep learning models. Additionally, soil sensor data will be collected to study environmental factors affecting soil texture.RESULTS: The GBNN showed improved accuracy in texture classification compared to previous models. Genetic algorithms effectively grouped similar particles, and soil sensor data provided insights into environmental impacts on soil texture.CONCLUSION: The GBNN for soil texture analysis overcame previous models' challenges, improving classification accuracy. The integration of soil sensor data provided valuable environmental insights, aiding farmers in optimizing crop selection, fertilizer application, and soil management for better yields and sustainability.
Read full abstract