Estimating numbers of oil palm (elaeis guineensis) pollen grains using image analysis and processing

Elaeidobius kamerunicus is the most important oil palm pollinator in indonesia and malaysia. however,the mechanism and efficiency of pollen transfer by this weevil are clearly not understood. the lack of study on pollination process in oil palm (elaeis guineensis) is mostly caused by difficulties in pollen counting due to their small size. most of the counting was conducted manually which is prone to mistakes,required extensive training,and time-consuming. the aim of this study is to provide a novel technique for counting pollen that is rapid,consistent,and efficient with a comparable accuracy to manual counting. male and female of e. kamerunicus were collected from male and female oil palm inflorescences (n=60). extracted pollen were placed and distributed in a flat microscope slide separated by designated observation chambers. images of each chamber were captured as a jpeg format and analysed by imagej. multiple macros were constructed for image processing steps to obtain the pollen numbers. comparison with manual counting using paired t-test,pearson's correlation and linear regression showed a high similarity between both methods.