Chipseal is the most common road surface in New Zealand due to its low cost and high performance monitoring. One of the parameters surveyed is the road texture depth which is currently measured by a volumetric technique known as the circular sand patch method, a method that is slow to perform, requires traffic control and depends on operator skill. The objective of our research is to design and build a portable device able to estimate the texture depth and evaluate its performance by correlating the results with the depths obtained with the sand patch method. The method chosen falls under the circular texture meters (CTM) category where a surface profile is acquired using a laser displacement sensor mounted on an arm that is rotated around an axis. The mechanical design follows the specifications from current practice established by a number of standards in Australia, USA and Europe. Profile processing is realized on board by a purpose built embedded system the results are available to the operator on an LCD display. Parameters used in estimation of texture depth are the standardized mean profile depth (MPD) and three parameters borrowed from surface metrology: average roughness (Ra), average square roughness (Rq) and average maximum height (Rz) which was redefined to allow for shorter CTM profiles. Algorithms were first developed into a Matlab toolbox then ported to the embedded system. The efficiency of the parameters was determined by the extent of how parameters correlate with texture depth obtained with the sand patch method characterised by the slope α and the intercept β of the linear relation that result in the best r-square for the parameter. The intercept was derived from mean of the parameters measured over a flat surface. The slope was estimated firstly from a test area of one square meter where sand patch texture depths were obtained and a number of profiles were recorded. Secondly, the slope was estimated from tracks acquired from a variety of sites using the correlation method with the intercept being constrained to the value found from the flat surface measurements. The results show that all parameters have an improved r-square when a second order fit is applied to the profile segments. The slopes estimated with the correlation method are in average 67% percent smaller than the slopes projected with the square meter test. Considering both the coefficient of determination of the correlations and the slope, the circular texture meter returns best estimates from the average maximum height when a second order fit is applied to the profile segments.