Applications

Automated mine-car filling with machine vision.

In continuous mining operations, the QDEEP SDK uses video analysis to monitor mine-car filling and send commands to the machinery, reducing the need for operators at the chute.

Mine cart filled with material in an underground tunnel

The challenge

Continuous operation in dusty conditions.

Mine-car filling has traditionally required operators to monitor and adjust the process manually. The work may run 24 hours a day in dust and noise, with each fill requiring close observation.

Automation is intended to improve filling efficiency, reduce manual workload in harsh conditions, and limit exposure to safety risks.

Material moving along a conveyor in an underground mine

The solution

Detect the car, fill material, and car number.

The QDEEP SDK provides mine-car detection, mineral-sand segmentation, and car-number recognition. The system records video for later playback and review.

Detection results can be passed to PLC controls for the mine car and mineral hopper. A production filling loop requires site-specific validation, PLC safety interlocks, operator override procedures, and testing under the mine's lighting and dust conditions.

Mineral material pouring into a mine cart

The payoff

Controlled fills with less operator exposure.

Automated filling is intended to reduce manual work at the chute and control the fill amount. Actual throughput, material waste, and operating cost depend on the process, PLC logic, model performance, and site conditions.

The workflow can keep video and filling records for operator review. Staffing and exposure changes should be assessed through the site's own safety and operations plan.

Row of loaded mine carts in an underground tunnel

The platform

Capture hardware that tolerates the factory floor.

The same industrial-vision architecture connects multi-channel SDI and GMSL2 capture to NVIDIA Jetson-class edge platforms, including the VPP6N0 series and the Jetson Thor-based EDG6N0-S T5X.

YUAN develops the hardware, drivers, firmware, FPGA logic, and SDK used in these production-line systems.