Rapid and Non-Destructive Monitoring of Moisture Content in Livestock Feed Using a Global Hyperspectral Model

SIMPLE SUMMARY: Moisture content is an important parameter for monitoring the quality of feed and feed materials as its established ranges serve as markers for safe storage, mixing, and feeding animals. The moisture content of feed materials changes very rapidly and necessitates rapid measurement. Current moisture content measurement methods are time-consuming, destructive, and require specialized skills. This often causes reduced and/or delayed testing, which results in the spoilage of feed and feed materials. Additionally, the improper balance of dry matter intake which is inversely proportional to moisture content often causes metabolic diseases for animals consuming the diet. To solve these, we have developed a rapid and non-destructive global hyperspectral model that could quantify moisture content in feed materials. Our results show that the developed model is robust, could provide a method to measure the distribution of moisture in feed, and has potential for implementation in a commercial setting. ABSTRACT: The dry matter (DM) content of feed is vital in cattle nutrition and is inversely correlated with moisture content. The established ranges of moisture content serve as a marker for factors such as safe storage limit and DM intake. Rapid changes in moisture content necessitate rapid measurements. A rapid and non-destructive global model for the measurement of moisture content in total mixed ration feed and feed materials was developed. To achieve this, we varied and measured the moisture content in the feed and feed materials using standard methods and captured their images using a hyperspectral imaging (HSI) system in the spectral range of 1000–2500 nm. The spectral data from the samples were extracted and preprocessed using seven techniques and were used to develop a global model using partial least squares regression (PLSR) analysis. The range preprocessing technique had the best prediction accuracy (R(2) = 0.98) and standard error of prediction (2.59%). Furthermore, the visual assessment of distribution in moisture content made possible by the generated PLSR-based moisture content mapped images could facilitate precise formulation. These applications of HSI, when used in commercial feed production, could help prevent feed spoilage and resultant health complications as well as underperformance of the animals from improper DM intake.