NETxTEN Core

NETxTEN Core is a C++ SDK for OEMs, camera manufacturers, and system integrators. It packages the same patented detection algorithm that powers the NETxTEN App for direct integration into existing products and platforms. It provides detection in real-time and supports alarming. NETxTEN Core is fast, tiny, does not need GPU or accelerated processing and can run on-camera hardware, on hardware boards with small processors, in existing OEM software, and in the cloud.

Connect with us on LinkedIn
for your TRIAL

Products
|

NETxTEN Core

Specifications

Specification Details
Size 20 - 130 MB depending on platform
Type C++ SDK
Algorithm Patented physics-based gas detection; fully automated, no human intervention required (USPTO US-12469250-B2)
Input Raw 16-bit radiometric video frames
Output (per frame) Processed gas signal frame; original frame passthrough; false color overlay (see API Overview section below)
Configuration Via NETxTEN's PresetManager (vendor + camera type + level) or manual parameter struct (see Configuration section below)
FFI Support C#, Dart; compatible with any language supporting C FFI bindings (see Language Support section below)
SDK Deliverables Platform-specific headers, compiled libraries, and binaries (see SDK Deliverables section below)
Platforms Windows, macOS, Linux, Android, iOS, and embedded/on-camera ARM (see Platform Support section below)

API Overview

NETxTEN Core processes video frame-by-frame through a stateful FrameProcessor that maintains temporal state across frames. The integration flow is:

  1. Create a frame grabber (file-based or camera)
  2. Create a FrameProcessor
  3. Configure parameters via PresetManager or manual Params struct
  4. (Optional) Build an ROI mask via RoiMaskBuilder
  5. Call newRecordingLoaded(fps, frameSize) to initialize
  6. Loop: supply raw frame → processFrame(frame, params) → consume result

Per-frame output (processFrame() returns):

Output Description
result.processed_frame Extracted gas signal as a processed frame (16-bit)
result.original_frame Original raw frame passthrough
False color overlay Generated separately via FalseColorUtils::preprocessFrame() — overlays detected gas regions in the selected color on a grayscale base 
// Process a frame
auto result = frame_processor.processFrame(frame, params);

// Generate false color overlay
auto overlay = FalseColorUtils::preprocessFrame(
    result.processed_frame, result.original_frame,
    params.camera_type, OverlayColor::RED);

Configuration

Parameters can be set two ways — both can be changed frame-by-frame mid-stream:

Option 1 — PresetManager (recommended) Load a preset by camera vendor, camera type, and sensitivity level. Presets are FPS-aware and automatically calculate temporal parameters.

auto params = preset_mgr.loadPresetAsParams(
    CameraVendor::FLIR, CameraType::COOLED,
    3,          // preset level
    frame_rate
);

Option 2 — Manual Params struct Configure every field directly without PresetManager.

Parameter Description
camera_type COOLED or UNCOOLED
std_count Temporal window size (typically fps / 3)
iterations Noise elimination iterations
hpf_removal_factor Highpass filter strength
alignment_strategy Camera motion compensation — AUTO_DECIDE, ALWAYS_ALIGN, or NEVER_ALIGN
run_hpf Enable / disable highpass filter
hist_equal Enable / disable histogram equalization
overlay_color 0 RED · 1 GREEN · 2 BLUE · 3 YELLOW · 4 MAGENTA
roi_mask Region of interest mask (uint8 pointer)
filter_blobs_size Minimum detectable plume size
filter_blobs_porosity Blob porosity threshold
pixel_difference_threshold Per-pixel change threshold
enable_frame_resize Enable / disable frame downscaling
temporal_subtraction_params.enabled Enable / disable temporal subtraction
fast_image_segmentation_params.enabled Enable / disable fast image segmentation
gradient_noise_processor_params.enabled, .std_noise_factor Enable / disable gradient noise processing; noise threshold factor
motion_magnitude_extractor_params.enabled Enable / disable motion magnitude extraction
temporal_canny_edge_tracker_params.enabled Enable / disable Canny edge tracking

Region of Interest (ROI)

NETxTEN Core supports inclusion and exclusion of regions in the field of view to focus detection on specific areas of the frame.

Method Description
RoiMaskBuilder OOP shape API — define rectangle (and other) include/exclude regions in normalized coordinates
Raw pointer Supply a pre-built binary mask as a uint8_t* directly to params.roi_mask (255 = keep, 0 = exclude)

SDK Deliverables

Deliverable Platform Notes
C++ header files All Public API — exposes FrameProcessor, Params, PresetManager, RoiMaskBuilder, and FalseColorUtils
Static libraries (.lib / .a) All For static linking
Dynamic library — .dll Windows For dynamic linking
Dynamic library — .so Linux, Android Shared object
Dynamic library — .dylib / .framework macOS, iOS Framework bundle available for Apple platforms

Language Support

Language Integration Method Status
C++ Native ✓ Explicitly supported
C# P/Invoke (FFI) ✓ Explicitly supported
Dart dart:ffi ✓ Explicitly supported
Python, Swift, Kotlin, Java, Rust, and others C FFI bindings Compatible in principle — any language with a C FFI mechanism can bind to the SDK

Platform Support

Platform Architecture Notes
Windows x64, ARM64
macOS x64, ARM64 Universal binary available
Linux x64, ARM64
Android ARM (all ABIs)
iOS ARM64
Embedded / On-Camera ARM Suitable for deployment directly on camera hardware

Additional platform and architecture support — including 32-bit ARM and other embedded targets — available on inquiry.

Target Integrations

Integration Type Description
Camera Manufacturers To embed advanced gas detection into camera firmware or companion software; run on-camera on ARM hardware
Drone & Robotics Platforms To add real-time advanced detection in airborne or ground-based inspection systems
System Integrators To add advanced detection capability onto remote continuous emissions monitoring solutions
Inspection Software Vendors To extend existing OGI or LDAR software with advanced plume detection
Solution Providers To build complete end-to-end LDAR solutions on top of NETxTEN Core

NETxTEN — Automated methane detection for optical gas imaging Aquanta Vision · Patent: USPTO US-12469250-B2

Connect with us on LinkedIn
for your TRIAL