The development of a SoftSeparatorTM for a lameness diagnostic system

Lameness in dairy herds is a critical economic factor and a crucial animal welfare issue. The authors, along with other scientists, developed an automatic lameness detection system named reaction force detection (RFD) system. The RFD performs lameness analysis based on measurements of ground reaction forces (GRF) as the animal walks freely through the system. This technology-based lameness analysis system is in its advanced commercialization stages, and it may prove to be a valid alternative to current animal scoring schemes that are observer dependent. The main outputs of the RFD are GRF and limb/floor contact locations of individual limbs. These GRF and limb/floor location records were used in the past to distinguish lame from sound cows. This paper further demonstrates that the very same records can be used to separate the RFD outputs, generated when a group of n cows walks through, into n files of a single animal in each. This file separation is a software-based algorithm that is based on induction. The induction-based animal separation process enables one to separate the records of any number, n, of animals that walk through as a single group. RFD outputs of a group of two and four animals are rewritten into multiple files, each one containing the records of a single animal. This animal separation algorithm could eliminate the need for gates that result in slowing down animal traffic.