The distribution of the marine algae is an important indication of the biodiversity changes in the aquatic ecosystem which algae biologist normally monitors. One of the most prominent instances is the monitoring of the invasive alga, e.g. Caulerpa taxifolia, through conducting regular surveys. It usually involves highly trained algae biologist to annotate the obtained video in order to detect the location where the alga would be likely present within survey area. This may constitute to a lengthy and demanding task which could be prone to observer-induced error. Hence, a framework is proposed herein to automate the analysis of underwater image to deduce if it contains the targeted alga, which is Caulerpa taxifolia. The framework employed HOG feature descriptor for object detection. Its efficiency and reliable was verified by the experiments using our consolidated database. © Springer Science+Business Media Singapore 2017.