Printing Digital Recognition Method Based on the Cascade Classification
Hui Pu, Wenxing Bao, Xiaowu Zhang · Advanced materials research · 2012
Milk yield automatic measurement in dairy farms plays a very important role in modern production management. Then, the measurement bottle scale and the reading recognition are key steps. So, this paper presents a new effective recognition algorithm on a printed numeric character of milk measurement bottle. The algorithm uses character image Euler number, the structural shape of printed numeric character, and vertical threading of character to accomplish the cascade classification. The method for character image does not need to be complex refining process, and has come up with a new algorithm of feature extraction from image sunken area. It reduces the complexity of the algorithm, the error recognition and rejection caused by thinning deformation. It greatly improves production efficiency and economic benefits. The experiment proves that the method is effective.