Acoustic surveys are effective methods to quantify the distribution and abundance of many aquatic fauna, and are particularly useful for the assessment of pelagic fishes. To convert data gathered from an acoustic survey of living resources into estimates of abundance, it is essential to know the target strength (TS, in dB) for the species and size of interest. TS is a logarithmic description of the quantity of acoustic backscattered energy from an individual and can be measured in situ using split-beam echosounders. Whereas in ideal conditions these may be the most representative estimates available, there is no comprehensive approach on how to account for the potential problem of multiple target detections and other potential sources of error. This paper describes the current state of the art and best methods to obtain standardized in situ TS measurements. We present recommendations on: equipment settings; selection of appropriate data; single-target detection criteria; removal of dense-echo regions to reduce the acceptance of multiple targets; post-processing; and validation methods. Most methods tested here were implemented using Echoview post-processing software. This is the first systematic and comprehensive description of ways to deal with the multitude of issues associated with multiple target errors. The methods described here are therefore recommended to improve single-target detection in order to deliver the most representative measures of TS for use in acoustic surveys.