Mask R-CNN vs The Ranch: How drones, AI, and instance segmentation count cattle for banks

Insider's Drone Guide

The discussion covers various topics pertinent to commercial drone operations, including an attendee’s recent drone incident involving a guidewire and troubleshooting a laptop issue during overnight data uploads. A significant portion of the conversation focuses on thermal training and certification, with participants discussing the pros and cons of in-person versus online UAS thermal courses and the impracticality of using DJI goggles for low-altitude inspections. Furthermore, the meeting includes a demonstration of the UGCS mission planning software, highlighting its simulation capabilities and advantages over DJI Flight Hub, alongside an active debate about hardware purchases, potential Black Friday deals, and the looming threat of a DJI drone ban.

Professional Drone Pilots Battle Ban and Automation

The discussion covers various topics pertinent to commercial drone operations, including an attendee’s recent drone incident involving a guidewire and troubleshooting a laptop issue during overnight data uploads. A significant portion of the conversation focuses on thermal training and certification, with participants discussing the pros and cons of in-person versus online UAS thermal courses and the impracticality of using DJI goggles for low-altitude inspections. Furthermore, the meeting includes a demonstration of the UGCS mission planning software, highlighting its simulation capabilities and advantages over DJI Flight Hub, alongside an active debate about hardware purchases, potential Black Friday deals, and the looming threat of a DJI drone ban.

A Drone Market on the Brink

This meeting summary details a discussion among drone industry professionals concerning market strategy, specific equipment, and project challenges, notably addressing the impending DJI ban. A key strategic decision was the group's agreement to acquire existing DJI equipment as a temporary measure until competitive US-made alternatives become available. The participants, including Michael Lilley and Jeff Musser, reviewed the new DJI M4E Enterprise drone's capabilities and praised the utility of Air Data software for flight management. Furthermore, the meeting addressed project-specific issues, such as Juan's need for entry-level drone recommendations for virtual tours and Peter Morssink’s scaling challenges in the Raising Canes project, which necessitate the use of ground control points (GCPs).

Drone Tech, Regulations, and Business Strategy

The source material is a transcript and summary of a weekly meeting focused on drone strategy, technology, and business operations among several participants, including Michael Lilley and Jeff Musser (SkyBorne Drone Solutions). A central topic of discussion is the impending DJI ban, with participants agreeing to a strategy of acquiring current DJI equipment, such as the new M4E Enterprise drone, as a temporary measure until viable US-made alternatives become available. The group also covered software recommendations, specifically praising Air Data for flight and maintenance tracking, and addressed practical business issues like scaling problems in construction projects that necessitate the use of Ground Control Points (GCPs). Finally, participants shared updates on new business opportunities, personal financial priorities, and concerns about the long-term financial viability of their drone businesses given the rising cost of non-DJI equipment.

Mask R-CNN vs The Ranch: How drones, AI, and instance segmentation count cattle for banks

Mask R-CNN, a state-of-the-art deep learning algorithm, for the automated counting of cattle captured in aerial imagery from quadcopters. The research addresses the challenge of reliably counting livestock in both extensive pastures and high-density feedlots to enhance intelligent livestock management. The study establishes that the Mask R-CNN framework performs reliably, achieving high counting accuracies and outperforming existing object detection algorithms, particularly in challenging scenarios involving occlusion and overlapping of animals. Ultimately, this work provides a promising step toward integrating artificial intelligence into quadcopter vision systems for precise and effective animal monitoring.