基於深度資訊之智慧型人形偵測系統設計

Abstract

近年來，由於人形偵測可應用的領域相當廣泛，因此受到重視且被深入的研究與討論，例如居家照護、人機溝通、智慧型汽車等皆是。本篇論文提出以Kinect所產生的深度圖為基礎的智慧型人形偵測系統，除了提高人形偵測率外，同時解決人形遮蔽的問題。整個系統分成三個部分：前景偵測、特徵擷取以及人形識別。雖然人會有許多不同的姿勢，但主要都是以垂直分布的方式呈現並具有一定的高度，根據此特性本系統先去偵測人形可能存在的區域，並且濾掉背景以增快速度；之後藉由邊緣擷取和距離轉換來萃取人形特徵，用以增加辨識率；此外畫面中的人形常因他人或物品之遮擋而只露出部分輪廓，為了解決這種遮蔽問題，本系統並不直接偵測整個人形，而是先利用凹槽匹配法找出各個身體部位，像是頭、身體、腳等，再利用類神經網路把各身體部位加以組合，並依此判斷是否為人形。根據實驗結果，本系統確實可以快速地偵測出人形，同時解決遮蔽問題，使偵測率提高至90%以上，甚至高達95%。

This thesis proposes an intelligent human detection system based on depth information generated by Kinect to find out humans from a sequence of images and resolve occlusion problems. The system is divided into three parts, including region-of-interest (ROI) selection, feature extraction and human recognition. First, the histogram projection and connected component labeling are applied to select the ROIs according to the property that human would present vertically in general. Then, normalize the ROIs based on the distances between objects and camera and extract the human shape feature by the edge detection and distance transformation to obtain the distance image. Finally, the chamfer matching is used to search possible parts of human body under component-based concept, and then shape recognition is implemented by neural network according to the combination of parts of human body. From the experimental results, the system could detect humans with high accuracy rate and resolve occlusion problems.