基于计算机视觉的运砂船超载状态检测

针对采砂监管过程中运砂船超载状态的检测问题,提出一种采用类Haar特征和Gentle Adaboost算法相结合的方法在非特写视频图像中检测吃水线和船舷线,通过手工标定的方式标定出标准载重干舷,然后通过实际干舷值与标准载重干舷值的比较来判断运砂船超载状态。实验表明,该方法能快速准确地在复杂的采砂现场环境中检测到目标运砂船的吃水线和船舷线,经手工标定后能准确地判断运砂船的超载状态,为采砂监管提供技术帮助。

Detection of Overload State for Sand Transporting Boat Based on Machine Vision

Aiming at the problem of sand transporting boat＇s overload in the sand mining supervise, comes up a new way for the detecting of water- line and freeboad using Haar-Like characteristic and Gentle Adaboost algorithm. In order to estimate the overload state, compares the value between waterline and freeboad and the value in the boat＇s certificate. The experiment results show that the waterline and freeboad can be found fast in the complex sand mining environment using this method, the overload state also can be got correctly, this method will provide technical support for the supervise of sand mining.