Image segmentation is a process of portion image into regions. From image segmentation schemes available, multilevel thresholding on the histogram is a highly established method. Otsu’s method is a significant multilevel thresholding technique, in this multiple threshold levels selected on histogram and group the pixels of an image into different regions. The optimized threshold levels computed with an Optimized technique by maximizing the inter-class variance. Methods with histograms are incapable to possess spatial details of contextual information for finding optimal threshold levels. As a remedy, a novel method proposed the Energy Curve is used instead of a histogram with Otsu’s method and Harmony Search Algorithm to compute optimized gray levels. The proposed method experimented on several benchmark images, and results compared with various optimization algorithms with histogram by Dunn Index, DB Index, SD Index, mean of fitness and PSNR, comparison clarifies that the proposed method is superior to histogram-based methods.

图像分割是将图像分成多个区域的过程。从可用的图像分割方案来看，对直方图进行多级阈值化是一种高度成熟的方法。Otsu的方法是一种重要的多级阈值技术，它是在直方图上选择多个阈值级别，然后将图像的像素分为不同的区域。通过最大化类间方差，使用优化技术计算出的优化阈值水平。带有直方图的方法无法拥有用于找到最佳阈值水平的上下文信息的空间细节。作为补救措施，提出了一种新的方法，即使用能量曲线代替Otsu方法和Harmony Search Algorithm的直方图来计算优化的灰度级。所提出的方法在几个基准图像上进行了实验，