查詢結果分析
相關文獻
- 個人電腦影像處理技術應用於即時車輛面式偵測
- 電腦視覺方法應用於PC廣域即時車輛偵測與交通目標之追蹤
- 高速Lempel Ziv資料壓縮器晶片之設計
- Constant-Time Algorithms for the Area and Perimeter of Image Components on Processor Arrays with Reconfigurable Bus Systems
- 混合式影像處理之即時邊緣偵測
- 媒體處理器視訊濾波應用探討
- 以三維小波轉換進行車輛偵測
- Image Processor and Fuzzy PID Controller Design for Robot-Car Intercept Mission
- 以即時視覺為基礎的夜間車輛偵測與駕駛輔助系統
- 車輛偵測器資料漏失之在線插補技術研究
頁籤選單縮合
題 名 | 個人電腦影像處理技術應用於即時車輛面式偵測=Real-Time Area-Wide Vehicle Detection Using PC-Based Image Processing |
---|---|
作 者 | 黃壬信; 曹壽民; | 書刊名 | 中國土木水利工程學刊 |
卷 期 | 10:3 1998.09[民87.09] |
頁 次 | 頁557-565 |
分類號 | 557.822 |
關鍵詞 | 即時; 面式偵測; 影像處理; 平行處理; 車輛偵測; Real-time; Area-wide detection; Image processing; Parallel processing; Vehicle detection; |
語 文 | 中文(Chinese) |
中文摘要 | 目前各國有關即時車輛偵測的研究係採用點式偵測及線式偵測方法, 模仿迴圈 偵測器之功能,對於交通偵測資料的空間連續性有所漏失。若為顧及空間連續性而採即時面 式偵測,則面臨運算量龐大與演算法複雜兩個問題,此乃交通影像偵測急待突破之課題。本 研究結合面式影像相減法及 OTSU 影像切割法,進行視窗面式偵測,在 PC 架構上,以兩個 中央處理單元作平行處理,速度較過去改進甚多,順利解決運算量龐大的問題,完成 PC 即 時影像偵測試驗。 經實地測試,模擬無人狀態下之自動化即時影像偵測,連續辨認 72,000 格影像,計 1421 輛車, 除外側車道之大車外,辨認率可達 90% 以上,該結果顯示即時交 通影像面式偵測系統的穩定性和未來發展潛力。本研究同時提出絕對精確度之觀念與計算方 式,探討偵測系統產生誤差之原因,並嘗試以偵測結果應用於擁塞判定與事件偵測。 |
英文摘要 | The objective of this paper is to develop a real-time area-wide vehicle detection system by using a PC-based image processing technique. In comparison with point or line detection of vehicles, area-wide detection has the capability of monitoring additional traffic parameters. However, it requires much more data to be processed. And a real-time system requires very high speed of data processing in order to respond instantly. In this study, a parallel processing framework is established to meet the abovementioned requirement. Several algorithms of image segmentation are evaluated. OTSU algorithm is considered the best and therefore is adopted. In a test vehicle classifiaction count, 1421 passenger cars and heavy vehicles are detected by the proposed system. Except for heavy vehicles in the outer traffic lane, accuracy level of 90% can be achieved. A measure of accuracy is developed to show error resulting from false alarms and missing counts. Causes of errors are examined and the use of the vehicle detection system is explored for incident detection. |
本系統中英文摘要資訊取自各篇刊載內容。