One common technology for underground object detection is Ground Penetrating Radar. For landmine detection, it is vital to have a fast and accurate method. It becomes more difficult when it comes to online detection. A fast and low power consuming algorithm needs to be developed for better CPU performances. This paper uses synthetic data from GprMax program and proposes a 3-step method to locate and discriminate underground objects: 1) Pre-processing using n-rows average 2) Image scaling and 3) converting Region of Interest to a feature vector. Proposed method has been tested using 7 methods; 2 classification algorithms; and 3 different image scales. The proposed method has increased Overall Performance from 80.4% to 90.3% for K-Nearest Neighbors (KNN) with K = 5 where Histograms of Oriented Gradients had 91.8%. Although, detection performance for proposed method when KNN is used is slightly lower compared to HOG, proposed method has a good potential with its runtime performance and small representation capacity.

探地物体的一种常用技术是探地雷达。对于地雷检测，拥有一种快速而准确的方法至关重要。在线检测变得更加困难。需要开发一种快速且低功耗的算法以提高CPU性能。本文使用来自GprMax程序的合成数据，并提出了一种三步方法来定位和区分地下物体：1）使用n行平均进行预处理2）图像缩放和3）将关注区域转换为特征向量。建议的方法已使用7种方法进行了测试；2个分类算法；和3种不同的图像比例。该方法增加了总体表现从80.4％至90.3％，为K-近邻（KNN）与ķ= 5，其中“定向梯度直方图”具有91.8％。虽然与HOG相比，使用KNN时该方法的检测性能略低，但该方法具有运行时性能和较小的表示能力，具有很大的潜力。