Multilevel Crop Image Segmentation using Bacterial Foraging Optimization Based on Minimum Cross Entropy

Crop images have different color intensities of a pixel as well as complex backgrounds. Hence, multilevel thresholding of crop images is very significant in the field of computer vision. Entropy-based multilevel thresholding is considered a successful enhancement over the bi-level thresholding technique for image segmentation. It is a time-consuming approach for practical uses. In this paper, minimum cross entropy (MCE) has been combined with the bacterial foraging optimization (BFO) algorithm has to enhance the accuracy of the segmented image. The BFO algorithm is a newly constituted evolutionary algorithm, which offers better search capabilities. The accuracy of the proposed method is tested over 10 different crop images with complex backgrounds and compared with an efficient algorithm such as an artificial bee colony (ABC). The experimental result demonstrates that the proposed technique segments the cropped image more accurately and searches multiple thresholds value very efficiently, which are close to the optimal value. The outcome of the proposed techniques shows a high quality of segmented images.