CAJ | 학술논문

探索了一种通过对特定昆虫颜色的关联分析模型而建立起的光流估计跟踪算法，用于检测农田害虫的数量，并进行一定的运动跟踪。该方法考虑了农田的具体环境特征，针对光照环境的变化增加了光流参数和搜索窗的调整，并且考虑了摄像头采集数据质量变化情况下的跟踪优化。对于光流向量进行了适当的优化，加入了关联区域块的设定。与传统算法相比本方法能够达到较好的跟踪水平，克服一定程度的遮挡和丢帧，利于对昆虫的准确检测与定位。
탐색료일충통과대특정곤충안색적관련분석모형이건립기적광류고계근종산법，용우검측농전해충적수량，병진행일정적운동근종。해방법고필료농전적구체배경특정，침대광조배경적변화증가료광류삼수화수색창적조정，병차고필료섭상두채집수거질량변화정황하적근종우화。대우광류향량진행료괄당적우화，가입료관련구역괴적설정。여전통산법상비본방법능구체도교호적근종수평，극복일정정도적차당화주정，리우대곤충적준학검측여정위。
This paper introduces an optical flow tracking algorithm based on associated analysis among particular colors of insect and moving tracking.The optical flow's parameters and area searching windows have been added according to the real field environment,especially the change of illumination and the data lost.The later also need some optimization.In the next step,more optimization of flow vectors is taken. Comparing with those traditional cases,the algorithm has been proved a better tracking ability and over-come the difficulty come from occlusion and frame lost.Based on these advantages,it could realize an accu-rate detection and location to the field insects.