Getting Started & Hardware Configuration
Configure your app settings to adapt to your hardware profile and fieldwork demands.
Step-by-Step Instructions
- First Launch Selection When launching the application for the very first time, you are prompted to select an AI Model Tier (Premium, Standard, or Legacy) which dictates how deep learning models are executed. The Premium tier uses INT4 quantized neural networks optimized for neuromorphic processing units (NPU) found in modern high-end SoCs (like Snapdragon or MediaTek APU). The Standard tier uses INT8 quantization targeting the GPU via WebGL/Vulkan runtimes, offering balanced performance. The Legacy tier relies on standard INT8 execution on the CPU using XNNPACK fallback kernels, ensuring total compatibility for older hardware. Selecting the right tier ensures optimal inference speed while keeping your device's memory footprint under check and preventing unexpected app closures during heavy operations. Furthermore, the application tests your device's available RAM and processor threads during this phase to suggest the best model type. It establishes standard default pipelines so that the neural engine runs smoothly without lagging your primary user interface. This hardware setup process is essential to guarantee that real-time edge calculations will not crash when you are deep in remote regions with no internet connection.
- Storage Preferences Navigate to the Settings panel and select the desired database storage strategy. By default, the application runs on a local-first architecture using a highly optimized SQLite database. This database can be encrypted using SQLCipher to ensure the confidentiality of your proprietary geological observations and coordinates in the field. You can choose to store the database on the internal storage or configure a secondary path on a removable SD card to safeguard your data. If you have active network coverage, you can also establish a secure connection to the GeoStratum Cloud Sync endpoint via HTTPS RESTful calls, enabling real-time remote backups and collaborative data sharing among field campaign members. This storage setup also allows you to configure automated local backup cycles that dump compressed SQL files to your selected directory at specified time intervals. By setting a customized storage directory, you make sure that geographic information, structural measurements, and high-resolution outcrop photographs are stored in a structural manner that prevents overwrite conflicts when importing data into external GIS software.
- Assistant Integration Enable the Android App Functions integration within the general configuration parameters. This exposes the app's internal capabilities and structural geological schemas directly to system-level virtual assistants (such as Google Gemini or Assistant). By authorizing this bridge, the assistant can execute voice commands and query your field data hands-free. You can ask queries like 'Show my last limestone outcrop' or 'Add a bedding measurement of 45 degrees dip' while your hands are busy using a geological hammer or holding onto a steep rock face, significantly increasing safety and operational efficiency during intense field exploration campaigns. The integration also maps specific voice prompts to actions like starting a new observation card, capturing a photo, or triggering the structural compass. The assistant parses your spoken descriptions using natural language processing to extract key details (like color, grain size, or mineral composition) and automatically fills the corresponding form fields, reducing physical typing in cold or rainy conditions.
When planning prolonged field surveys in remote outcrops, toggle on the 'Low Power Screen Wake' option in the settings. This specialized mode uses the device's built-in accelerometer to detect when the phone is held horizontally or placed in a pocket. It automatically suspends display rendering and touch digitizers while keeping the core GPS logging threads active in the background. This reduces overall battery drainage by up to 45%, ensuring your smartphone survives the entire day without losing track of your geographic coordinates and structural measurements. Additionally, you can configure the sensor polling interval to strike the perfect balance between spatial precision and energy consumption. We also recommend disabling unnecessary wireless search services (such as Wi-Fi scanning and Bluetooth beacons) when mapping remote structures, as constant signal searching in areas without network coverage consumes significant power without providing any geological utility.