A novel segmentation algorithm for clustered flexional agricultural products based on image analysis

• An algorithm to separate clustered flexional agricultural products has been developed.
• The proposed algorithm achieved a mean accuracy of 92.7% in clustered shrimp dataset.
• The algorithm has potential to segment flexional agricultural products.

The problem for segmentation of clustered flexional agricultural products becomes complex when we perform the duties of counting and classification. A novel algorithm based on concavities and circle fitting is proposed to solve these difficulties. Initially, a circular mask method was applied into the contour images of clustered shrimp, to acquire a series of concavity points. Furthermore, the candidate segmentation lines can be acquired by connecting each two concavity points, and then the correctness for each candidate segmentation line was evaluated by designing four acceptance criterions. Additionally, one new point was acquired by combining adaptive two concavity points together to construct a training model to fit the circle equation, which can transform the erroneous straight segmentation lines into proper curve segmentation lines. Finally, the straight and curve segmentation lines were integrated in one clustered image to achieve the segmentation results. Experimental results revealed that the proposed algorithm achieved a mean accuracy of 92.7% across the clustered shrimp dataset. Other two application examples of flexional agricultural products, such as clustered green pepper and shrimp meat, were also used to test the effectiveness of the proposed algorithm. Segmentation results demonstrated it can successfully segment the images, which indicates the proposed algorithm has the potential to separate clustered flexional agricultural products.