Computer‐based image analysis to quantify the number of micro‐arthropods in a sample

Stored grains, household dust, litter, and soil habitats are dominated by micro-arthropods. Pest control and ecological studies concerning these habitats are inherently confronted with the problem of enumerating huge quantities of individuals. Stored grain may contain up to several thousands of individuals per 100 g sample (Athanassiou et al., 2003; Kučerová et al., 2003; Stejskal et al., 2003; Hubert et al., 2006; Palyvos et al., 2008). The small size and large number of individuals make accurate counting and identification difficult and time-consuming. Microarthropods are usually extracted from infested grain, dust, or soil samples in the laboratory using a Tullgren apparatus (Krizkova-Kudlikova et al., 2007; Stejskal et al., 2008) or a heat-escape method (Bischoff et al., 1992). The extracted biological material is usually counted using a stereomicroscope. Direct visual search of micro-arthropods is the subject of a voluminous literature (Wolfe, 2005), which provides evidence that counting is error-prone and individually biased. Solomon (1945) attempted to facilitate the counting of micro-arthropods (mites) by introducing disks divided into segments. Others have used photography (Asquish, 1965; Sircom, 2000), but this was not less timeconsuming. The present-day technique of digital photography (Spring, 2000; Weyda, 2002) coupled with digital image analysis (DIA) (Erickson et al., 2001; Richardson et al., 2001) provides a new method for identifying and counting micro-arthropods. Automated computer imaging systems have been developed for the identification of midges (Weeks et al., 1999), collembolans (Krogh et al., 1998), nematodes (Brieri et al., 1987), and fungal biomass (Morgan et al., 1991). Despite these recent applications of DIA to agricultural pest control and stored-product pest research and practice (e.g., Panneton & Drummond, 1991; Lukáš & Stejskal, 2004; Fornal et al., 2007), many important areas remain unexplored. Here, we describe DIA for the estimation of population density in laboratory-analysed samples infested with stored-product pest micro-arthropods. We compare this approach with traditional direct visual counting using Acarus siro L. (Acari: Acaridae) and Liposcelis bostrychophila Badonnel (Psocoptera: Liposcelididae).