Sustav za detekciju pješaka

master's thesis

• Title (english): Pedestrian detection system
• Author: Daniel Ćuk
• Mentor: Zoran Kalafatić (mentor)
• Committee member: Zoran Kalafatić (predsjednik povjerenstva)
• Committee member: Siniša Šegvić (član povjerenstva)
• Committee member: Tomislav Hrkać (član povjerenstva)
• Granter: University of Zagreb; Faculty of Electrical Engineering and Computing; Zagreb
• Defense date and country: 2014-07-15, Croatia
• Scientific / art field, discipline and subdiscipline: TECHNICAL SCIENCES; Computing

Abstract

Povećanje sigurnosti u prometu težak je zadatak koji privlači sve veći broj znanstvenika. Sustav za detekciju pješaka uvelike bi pridonio sigurnosti pješaka u prometu. Vozačima bi olakšao uočavanje pješaka te samim time pridonio lakšem i bezbrižnijem sudjelovanju u prometu. U ovome radu opisane su tri vrste detektora pješaka te uspoređene njihove performanse: kaskada klasifikatora s pravokutnim značajkama, klasifikator SVM sa značajkama HoG te kaskada klasifikatora SVM s opisnikom značajki HoG. Kaskade su učene varijacijama algoritma AdaBoost. Algoritam AdaBoost također je upotrijebljen za odabir najrelevantnijih značajki iz velikog skupa značajki. Također je razvijen sustav s grafičkim sučeljem za detekciju pješaka koji koristi sve tri vrste detektora. Sustav radi na slikama ili videu.

Abstract (english)

Improvement of traffic safety is a difficult task that attracts more and more scientists. Pedestrian detection system would considerably contribute to pedestrian traffic safety. It would make the detection of pedestrians easier to drivers and at the same time would contribute to easier and careless traffic engagement.
Three types of pedestrian detection are described in this paper and their characteristics are compared: cascade of classifiers with rectangular features, classifier SVM with HoG features and cascade of SVM classifiers with HoG features descriptor. The cascades are learned with variations of AdaBoost algorithm. The AdaBoost algorithm is also used to select the most relevant features from larger set of features. The system with graphical interface for pedestrian detection using all three types of detector is also developed. The system works on images or video.

• Keywords: detekcija pješaka; detekcija objekata; kaskada klasifikatora; histogram orijentiranih gradijenata; integralni histogram orijentiranih gradijenata; pravokutne značajke; SVM; AdaBoost; integralna slika; odabir značajki
• Keywords (english): pedestrian detection; object detection; cascade of classifiers; histogram of oriented gradients; integral histogram of oriented gradients; rectangular features; SVM; AdaBoost; integral image; feature selection
• Language: croatian
• URN:NBN: urn:nbn:hr:168:384828
• Study programme: Title: Computing; Study programme type: university; Study level: graduate; Academic / professional title: magistar/magistra inženjer/inženjerka računarstva (magistar/magistra inženjer/inženjerka računarstva)
• Type of resource: Text
• File origin: Born digital
• Access conditions: Closed access
• Created on: 2018-02-06 08:22:21