CAJ | 학술논문

在图像识别问题中，基于Haar特征的图像识别算法已经十分普遍，并且得到了广泛的应用。但是，目前此类算法还存在有时间复杂度高、对图像亮度、尺寸变化敏感、图像识别精度差等缺点。为了提高图像识别的精度，提出了一种新的边缘检测图像识别算法。这种算法首先使用Canny算子将图像的边缘像素识别出来，然后计算每一个有效像素的梯度。通过得到的像素梯度序列建立归一化直方图，对归一化直方图分析后得到最优识别。基于边缘检测的图像识别算法使匹配图像效率得到了提高，图像识别流程得到优化。
재도상식별문제중，기우Haar특정적도상식별산법이경십분보편，병차득도료엄범적응용。단시，목전차류산법환존재유시간복잡도고、대도상량도、척촌변화민감、도상식별정도차등결점。위료제고도상식별적정도，제출료일충신적변연검측도상식별산법。저충산법수선사용Canny산자장도상적변연상소식별출래，연후계산매일개유효상소적제도。통과득도적상소제도서렬건립귀일화직방도，대귀일화직방도분석후득도최우식별。기우변연검측적도상식별산법사필배도상효솔득도료제고，도상식별류정득도우화。
Image Recognition Algorithms based on Harr-features are very popular and widely used in image recognition. However,these algorithms has a lot of shortcomings,such as high time complexity,sensitive to the variation of image luminance and scale, poor Image recognition accuracy and so on. In order to improve the accuracy of image recogni-tion, a new algorithm Based on Edge detection is proposed. This algorithms Canny operator to detect edge pixels of the image,calculated gradient of each pixel which is valid,and build normalized histogram by gradient sequence of pixels. At last, analysis normalized histogram got accurately recognition. The result shows that Image Recognition Algorithm Based on Edge detection improved the image recognition rate,and The image recognition process has been optimized.