Information on weed distribution within the field is necessary to implement spatially variable herbicide application. This paper deals with the development of near-ground image capture and processing techniques in order to detect broad-leaved weeds in cereal crops under actual field conditions. The proposed methods use colour information to discriminate between vegetation and background, whilst shape analysis techniques are applied to distinguish between crop and weeds. The determination of crop row position helps to reduce the number of objects to which shape analysis techniques are applied. The performance of algorithms was assessed by comparing the results with a human classification, providing an acceptable success rate. The study has shown that despite the difficulties in accurately determining the number of seedlings (as in visual surveys), it is feasible to use image processing techniques to estimate the relative leaf area of weeds (weed leaf area/total leaf area of crop and weeds) while moving across the field and use these data in a stratified manual weed survey of the field.