Hoof lesion detection with manual and automatic locomotion scores in dairy cattle

Andres Schlageter Tello, Tom van Hertem, S. Viazzi, Eddie Bokkers, P.W.G. Groot Koerkamp, Machteld Steensels, C.E.B. Romanini, C. Bahr, I. Halachmi, D. Berckmans, Kees Lokhorst

Research output: Chapter in Book/Report/Conference proceedingChapter

Abstract

The detection of hoof lesions is an important management practice in dairy farms. Under farm conditions, manual locomotion scoring is often used to detect hoof lesions. Recently, different automatic locomotion scoring systems have been developed. The objective of this study was to determine the capability of a manual (MLS) and an automatic (ALS) locomotion score for hoof lesion detection. The experiment was performed at a dairy farm with 250 milking cows. The presence and severity of hoof lesions were assessed while cows were hoof trimmed. Manual locomotion scoring was performed before hoof trimming. Automatic locomotion scoring was performed with a system based on a 3D camera, positioned in top-down perspective. Both manual and automatic locomotion scoring were performed using a 5-level scale and later transformed into a lame/non-lame classification (lame ≥3). The lame/non-lame classification from MLS and ALS was used to calculate the sensitivity and specificity using as reference hoof lesions and severe hoof lesions. The percentage of cows which had hoof lesions at each level of the 5-level MLS was 72% at level 1, 86% at level 2, 89% at level 3, 96% at level 4, and 50% at level 5. The percentage of cows with severe hoof lesions at each level of the MLS was 34% at level 1, 52% at level 2, 62% at level 3, 82% at level 4 and 50% at level 5. The percentage of cows which had hoof lesions at each level of the 5-level ALS was 89% at level 1, 75% at level 2, 78% at level 3, 82% at level 4 and 100% at level 5. The percentage of cows with severe hoof lesions at each level of the ALS was 37% at level 1, 39% at level 2, 59% at level 3, 45% at level 4 and 100% at level 5. When transformed into a lame/nonlame classification MLS showed a sensitivity of 36% and specificity of 81% when hoof lesions were used as reference and a sensitivity of 43% and specificity of 78% when severe lesions were used as reference. For the ALS, sensitivity for hoof lesions was 47% and for severe hoof lesions was 58%. In conclusion, both manual and automatic locomotion scores demonstrated a poor to moderate capability to detect hoof lesions and severe hoof lesions.
Original languageEnglish
Title of host publicationPrecision livestock farming applications
Subtitle of host publicationMaking sense of sensors to support farm management
EditorsIlan Halachmi
PublisherWageningen Academic Publishers
Chapter2.4
Pages65-70
Number of pages6
ISBN (Electronic)9789086868155
ISBN (Print)9789086862689
DOIs
Publication statusPublished - 15 Jun 2015

Fingerprint Dive into the research topics of 'Hoof lesion detection with manual and automatic locomotion scores in dairy cattle'. Together they form a unique fingerprint.

Cite this