基于SIFT算法的多变背景目标识别

Variable Background Target Recognition Based on SIFT Algorithm

针对多变背景下目标识别的复杂性和多样性,利用尺度不变特征变换提取特征点,采用近邻法进行特征匹配,通过调整阈值,提高特征匹配的准确率。通过实验验证,在一定的阈值范围内进行特征匹配,近邻法能够有效保证正确的匹配数量,提高目标识别的稳定性与可靠性,并在多变背景下精确识别目标,具有较好的鲁棒性。基于此,展开具体论述。

Aiming at the complexity and diversity of target recognition in the changeable background,feature points are extracted by scale invariant feature transformation,and feature matching is carried out by using the nearest neighbor method.The accuracy of feature matching is improved by adjusting the threshold.The experimental results show that the nearest neighbor method can effectively guarantee the correct number of matches,improve the stability and reliability of target recognition,and accurately identify the target in the changeable background.It has good robustness.Based on this,the specific discussion is carried out.