**2.4. Data processing and decision making AI logic system**

All data processing in intelligent MCM operations is carried through a systematic data fusion approach. AI machine learning and perception, feature extraction via real-time image or intelligent navigation and mine-clearing operation sequences are integrated with mission management processes. Main control center will support and coordinate automatic classifica‐ tion of mines and other ordnance, as well as intelligence obstacle avoidance navigations using data from navigation sensors. The image and pattern identification-processing techniques, which have been adapted from video image and sensor signal, focus on classification of mines by the recovery of explicit feature-based descriptions of man-made objects [12]. Silhouette descriptions and classification are used to support the dynamic view positioning and auto‐ matic object classification.

The algorithmic partitioning in a front-end digital signal processing (DSP) dedicates image acquisition, dynamic range equalization, image segmentation and region-of-interest identifi‐ cation [13]. The parallel processing engine supporting applications of a statistical evaluation of linearity or regular curvature have gradient extraction, edge thinning and threshold algorithms [14]. All the data from traditional navy data base, mine warfare environmental decision aid library (MEDAL) systems on MIW, MCM and tactical oceanography also could be accessed in main processing unit in the MCM UUV system and fed into the identification and classification processors [15]. A typical MCM data-processing flow of MCM UUV systems is shown in **Figure 1**.

**Figure 1.** Mine warfare environmental data flow.

Reacquisition and relocalization of predesignated mine or mine fields need huge amount of signal, communications and data packages from various sensor systems. Data sets for reconstructing three-dimensional (3D) and two-dimensional (2D) modeling are very big and very difficult to transfer through current acoustic carrier in underwater environments [5]. Some other information comes from distance oriented and directional angle of illumination of light source, which gives some incentive in the reconstruction of 3D or 2D model of mines.

Identification, reconfirmation of mines and mine-like objects (MLO) classifications are critical factors for mine disposal operations. Efficient and different bandwidth characteristics of communication careers are critically needed at the main control center of MCM operations to gain access to high-quality mine detection sensor data from a remote area, due to the lack of computational capabilities of the existing sensor data-processing systems [16, 17].
