IR-camera method to determine urine puddle area in dairy cow houses

Dairy cow houses are a major contributor to ammonia (NH3) emission in many European countries. A better insight in the emission process is required to develop mitigation measures to control environmental pollution. To understand and predict NH3 emissions from cubicle dairy cow houses Monteny (1998) developed a mechanistic model. With this model a sensitivity analysis was performed to assess the contribution of various inputs to the variation in NH3 emission. It appeared that NH3 emission was most sensitive for five input variables: four relating to the puddle (area, depth, pH, temperature), plus initial urea concentration. Unfortunately, cow house data of these variables are scarce due to the lack of proper measurement methods that can be applied. In this study we focus on the urine puddle area. Our objective was to develop a method for measuring the urine puddle area on the floor in commercial dairy cow houses. To do so, we explored measurement principles that can both be used in cow houses and as reference methods. In this study we performed a preliminary experiment and assessed the results in order to define a measurement method for commercial dairy cow houses. We measured 35 fresh and warm urine puddles at two commercial dairy farms and 30 simulated puddles at an experimental setup. Measurements were carried out with a measurement grid and a thermal infrared camera (IR-camera). The IR-images were analysed with vision software. With the IR-camera we obtained IR-images of complete urine puddles on the floor in commercial dairy cow houses. We took an IR-image directly after urination, since puddle temperature drops quickly to floor temperature. The centre of a puddle in the image was warm and the temperature gradually decreased to floor temperature at the edges. It was difficult to define a temperature threshold that precisely distinguished the boundary of the puddle from the floor. With subtraction of the IR-image of the location background, the temperature values caused by a urination became more clear-cut compared to the original and the values differed less among puddles. This made it easier to assign each pixel to be part of the puddle or not. However, errors in pixel assignment can still occur easily, resulting in large variation in the determined absolute puddle area. For example a puddle was 1.001 m2 at a threshold of 1.0 °C and was 0.809 m2 at 2.0 °C. On the other hand, when using one and the same threshold for each IR-image, a precise comparison of areas among puddles could be made. We conclude that it is possible to use an IR-camera to compare puddle areas of urine puddles in dairy cow houses. To determine the absolute puddle area additional measurements are required to correctly appoint a pixel to be puddle or not.