Intelligent Road Inspection Vehicle with AI Inference Industrial PC

Major cities across the world boast an extensive network of tens of thousands of roads. This intricate system of roads plays a crucial role in the infrastructure of any major urban center, influencing both its economy and the daily lives of its inhabitants. An efficiently designed road network takes into account factors such as traffic lights, stop signs, traffic flow patterns, and population density to facilitate smooth traffic circulation. However, given that roads are human-made structures, they are susceptible to significant wear and tear caused by factors like vehicle traffic, intense sunlight, and heavy rainfall. As a result, defects and damages frequently emerge, posing potential hazards to safety.

In terms of classification, there are a total of 13 distinct categories for defects and damages.

• Alligator cracking
• Block cracking
• Longitudinal (linear) cracking
• Transverse cracking
• Edge cracking
• Joint reflection cracking
• Slippage cracking
• Potholes
• Depressions

• Rutting
• Shoving
• Upheaval
• Ravelling
• Other issues requiring treatment before maintenance
• Oil spots
• Grass
• Mud, tree sap, etc.

Network (or system) level

It addresses the entire pavement system and primarily focuses on strategic decisions related to network-wide planning, policies, and budget allocation.

Project Level

It addresses individual segments within the network and typically pertains to more detailed decisions related to factors like condition assessment, maintenance, reconstruction, and rehabilitation (MR & R) assignment, as well as unit costs.

Nonetheless, for the integration of AI-based defect detection into a vehicle, it must overcome challenges related to environmental conditions, the time required for software development to identify defects, and the accuracy of detection.

Solution

Intelligent road inspection eliminates human-related factors like inconsistency and fatigue by incorporating an AI inference industrial PC. This AI inference industrial PC identifies and categorizes defects using the same knowledge database, ensuring greater consistency in results. When it identifies defects requiring immediate attention, these can be relayed to a central control hub through wireless 5G/4G communication. The capacity to deploy multiple vehicles simultaneously enables broader coverage of roads and larger areas.

Our client opted to integrate the Neousys GC series rugged AI inference platform into their intelligent inspection vehicle. This system offers genuine wide-temperature functionality and provides various connectivity options, including support for PoE, USB, or GigE industrial cameras, which are essential for detecting road cracks and irregularities. The inclusion of a robust 120W NVIDIA GTX graphics card empowers the system with AI inference capabilities to identify the nature and extent of damage. Furthermore, with built-in ignition control and MIL-STD-810G shock and vibration certification, this system is purpose-built to withstand the demanding conditions of in-vehicle deployment.

The benefits of adopting an intelligent road inspection vehicle include enhanced consistency in inspections and improved recognition of pavement damages. The industrial PC draws upon an expanding database of pavement images gathered from other intelligent road inspection vehicles and continuously learns from it. This approach also reduces the time required for manual inspections and allows for early detection, addressing minor issues before they escalate into significant problems.

Outcomes:

> 95%

Precise anticipation of road conditions in advance.

> 80%

Reduce the time needed for manual inspections

> 50%

Reduce maintenance costs

In addition to addressing in-vehicle and environmental conditions, the Neousys GC series industrial PCs offer expansion slots for accommodating 5G/4G wireless communication and GPS modules. The 5G/4G communication module enables data upload, live image feedback, and general communications, while the GPS module allows tracking of the intelligent road inspection vehicle's location, even in remote areas if necessary.

With the integration of an inference accelerator, efficient thermal dissipation, resistance to shock and vibration, extensive connectivity for cameras, Lidars, sensors, and the capacity for expanding functionality, it's evident why our clients have chosen a Neousys GC series industrial PC as the central component for their intelligent road inspection vehicle.

Advantages offered by Neousys Rugged Embedded GC Series Industrial PCs

Neousys' rugged embedded GC series platforms provide the following benefits:

Environment

• True -25°C up to 60°C wide-temperature operations for harsh, volatile environments
• Patented damping bracket to withstand shock and vibration conditions

Connectivity/ expandability

• PoE+ connectivity for GigE cameras
• 2 Gen1/ Gen2 connectivity for USB cameras
• Connection ports with screw-lock for rugged connectivity
• WiFi 6/ WiFi 5/ 5G/ 4G wireless communication, expansion via mini-PCIe/ M.2 modules

Inference processing power

• Supports an NVIDIA® graphics card up to 130W
• Supports an NVIDIA® Tesla/ Quadro/ L4 inference accelerators

Electrical

• Patented SuperCAP UPS to counteract unforeseen power interruptions
• Wide-range DC input

Expansion

• MezIO ignition control
• MezIO digital input/ output, COM, Ethernet, USB

To explore more innovative industrial PC and machine vision products, visit our website for buildable modular solutions for your equipment needs.