TWI601426B - Surveillance system of 3d panoramic images - Google Patents

Surveillance system of 3d panoramic images Download PDF

Info

Publication number
TWI601426B
TWI601426B TW104139564A TW104139564A TWI601426B TW I601426 B TWI601426 B TW I601426B TW 104139564 A TW104139564 A TW 104139564A TW 104139564 A TW104139564 A TW 104139564A TW I601426 B TWI601426 B TW I601426B
Authority
TW
Taiwan
Prior art keywords
image
original
control module
message
monitoring
Prior art date
Application number
TW104139564A
Other languages
Chinese (zh)
Other versions
TW201720141A (en
Inventor
陳智勇
王玉樹
林子涵
楊瑞瑜
徐偉倫
李炫輿
Original Assignee
樹德科技大學
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 樹德科技大學 filed Critical 樹德科技大學
Priority to TW104139564A priority Critical patent/TWI601426B/en
Publication of TW201720141A publication Critical patent/TW201720141A/en
Application granted granted Critical
Publication of TWI601426B publication Critical patent/TWI601426B/en

Links

Description

立體環景影像之監控系統 Stereoscopic panoramic image monitoring system
本發明是有關於一種監控系統,特別是有關於一種用於行動載具,並具有即時監控及通報之立體環景影像之監控系統。 The invention relates to a monitoring system, in particular to a monitoring system for a mobile vehicle and having a stereoscopic panoramic image for real-time monitoring and notification.
傳統的汽車防盜裝置為機械式防盜設備,其主要設計是藉由機械結構來鎖固離合器、制動器、油門、轉向盤或是排擋桿,以達到防盜的目的。但,由於機械結構比較笨重,使用起來也比較麻煩。儘管隨著一些先進材質的開發,機械式防盜器的硬度或韌性大幅提昇,已能更有效的防止被撬開或鋸開,而提高盜賊竊車的難度,然而,在使用上仍嫌不便,並且也無法在竊車行為發生的第一時間,就立即警告車主或通知警察單位,因此已逐漸為市場所淘汰,或做為一種輔助性的防盜設備來使用。 The traditional anti-theft device is a mechanical anti-theft device. Its main design is to lock the clutch, brake, throttle, steering wheel or gear lever by mechanical structure to achieve the purpose of anti-theft. However, due to the cumbersome mechanical structure, it is also troublesome to use. Although the hardness or toughness of the mechanical anti-theft device has been greatly improved with the development of some advanced materials, it has been more effectively prevented from being opened or sawed, and the difficulty of stealing a car by a thief is improved. However, it is still inconvenient in use. It is also impossible to immediately warn the owner or notify the police unit at the first time of the car theft, so it has gradually been eliminated by the market or used as an auxiliary anti-theft device.
近來,汽車上所使用的設備普遍都是偵測震動式的防盜器,當車子受到震動的時候,防盜器警鈴便會響起,提醒使用者前往查看。但若使用者並不在車子附近,無法察覺到警鈴聲已響起,便無法預防悲劇的發生。 Recently, the equipment used in automobiles is generally a shock-detecting type of anti-theft device. When the car is shaken, the alarm bell will sound to remind the user to go to the view. However, if the user is not in the vicinity of the car and cannot detect that the alarm bell has sounded, the tragedy cannot be prevented.
綜觀前所述,本發明之發明人思索並設計一種立體環景影像之監控系統,以期針對習知技術之缺失加以改善,進而增進產業上之 實施利用。 As described above, the inventor of the present invention contemplates and designs a monitoring system for stereoscopic panoramic images in order to improve the lack of conventional techniques, thereby enhancing the industry. Implementation and utilization.
有鑑於上述習知技藝之問題,本發明之目的就是在提供一種立體環景影像之監控系統,以解決習知技術所存在之問題。 In view of the above-mentioned problems of the prior art, it is an object of the present invention to provide a stereoscopic panoramic image monitoring system to solve the problems of the prior art.
根據本發明目的,提出一種立體環景影像之監控系統,其應用於行動載具,監控系統包含複數個攝像模組、控制模組及傳輸模組。複數個攝像模組設置於行動載具之周圍,並持續擷取行動載具之外部環境之影像,以產生複數個原始影像訊號及複數個非原始影像訊號。控制模組電性連接複數個攝像模組,控制模組接收複數個原始影像訊號及複數個非原始影像訊號,並經由影像縫合處理,而分別產生原始環景影像訊息及複數個非原始環景影像訊息,且控制模組經由影像比對處理比對原始環景影像訊息及複數個非原始環景影像訊息,而判斷至少一非原始環景影像訊息異常時,控制模組則據以產生異常訊息。傳輸模組電性連接控制模組,控制模組經由傳輸模組傳送異常訊息至至少一外部電子裝置。 According to the object of the present invention, a stereoscopic panoramic image monitoring system is proposed, which is applied to a mobile vehicle. The monitoring system comprises a plurality of camera modules, a control module and a transmission module. A plurality of camera modules are disposed around the mobile vehicle and continuously capture images of the external environment of the mobile vehicle to generate a plurality of original image signals and a plurality of non-original image signals. The control module is electrically connected to a plurality of camera modules, and the control module receives a plurality of original image signals and a plurality of non-original image signals, and performs image stitching processing to generate original ring image information and a plurality of non-original scenes respectively. Image information, and the control module compares the original panoramic image information and the plurality of non-original panoramic image information through the image comparison, and determines that at least one non-original panoramic image information is abnormal, and the control module generates an abnormality according to the image message. The transmission module is electrically connected to the control module, and the control module transmits an abnormal message to the at least one external electronic device via the transmission module.
較佳地,控制模組依據複數個原始影像訊號而產生複數個原始影像訊息,控制模組比對複數個原始影像訊息中之複數個第一特徵值且判斷符合預設門檻值時,控制模組則據以將複數個原始影像訊息進行影像縫合而產生原始環景影像訊息。 Preferably, the control module generates a plurality of original image messages according to the plurality of original image signals, and the control module compares the plurality of first feature values in the plurality of original image messages and determines that the preset threshold value is met. The group then stitches a plurality of original image messages to produce an original panoramic image message.
較佳地,控制模組依據複數個非原始影像訊號而產生複數個非原始影像訊息,控制模組比對複數個非原始影像訊息中之複數個第二特徵值且判斷符合預設門檻值時,控制模組則據以將複數個非原始影像訊息進行影像縫合而產生非原始環景影像訊息。 Preferably, the control module generates a plurality of non-original image messages according to the plurality of non-original image signals, and the control module compares the plurality of second feature values in the plurality of non-original image messages and determines that the preset threshold value is met. The control module performs image stitching on the plurality of non-original image messages to generate non-original surround image information.
較佳地,控制模組以背景相減法將原始環景影像訊息中之原始背景影像訊息與複數個非原始環景影像訊息中之非原始背景影像訊息進行相減,而產生至少一局部影像訊息,在控制模組經由分類程式判斷至少一局部影像訊息符合預設人形影像訊息時,控制模組則據以產生異常訊息。 Preferably, the control module subtracts the original background image information in the original scene image information from the non-original background image information in the plurality of non-original scene image information by background subtraction to generate at least one partial image message. When the control module determines, by the classification program, that at least one partial image message conforms to the preset humanoid image message, the control module generates an abnormal message accordingly.
較佳地,傳輸模組接收至少一外部電子裝置所傳輸之影像傳輸訊息時,控制模組則據以經由傳輸模組傳送至少一非原始環景影像訊息至至少一外部電子裝置。 Preferably, when the transmission module receives the image transmission message transmitted by the at least one external electronic device, the control module transmits at least one non-original scenery image message to the at least one external electronic device via the transmission module.
較佳地,傳輸模組接收至少一外部電子裝置所傳輸之啟動訊息時,控制模組則據以控制複數個攝像模組進行影像擷取之作動,並依據複數個原始影像訊號及複數個非原始影像訊號進行影像縫合處理與影像比對處理。 Preferably, when the transmission module receives the activation message transmitted by the at least one external electronic device, the control module controls the plurality of camera modules to perform the image capture operation, and according to the plurality of original image signals and the plurality of non- The original image signal is subjected to image stitching processing and image comparison processing.
較佳地,立體環景影像之監控系統更可包含電源供應模組,其電性連接複數個攝像模組、控制模組及傳輸模組,電源供應模組供給電能至複數個攝像模組、控制模組及傳輸模組。 Preferably, the stereoscopic panoramic image monitoring system further comprises a power supply module electrically connected to the plurality of camera modules, the control module and the transmission module, and the power supply module supplies the power to the plurality of camera modules, Control module and transmission module.
較佳地,立體環景影像之監控系統更可包含警示模組,其電性連接控制模組,在控制模組判斷至少一非原始環景影像訊息異常時,控制模組則據以控制警示模組產生警示訊息。 Preferably, the monitoring system of the stereoscopic panoramic image may further comprise a warning module electrically connected to the control module. When the control module determines that at least one non-original panoramic image information is abnormal, the control module controls the warning according to the control module. The module generates a warning message.
承上所述,依本發明之立體環景影像之監控系統,其可藉由複數個攝像模組擷取行動載具四周環景影像,並利用控制模組將環景影像經過影像縫合之處理還原成矩形影像後,就可以透過離散小波轉換(Discrete Wavelet Transform)、影像處理與分割技術、特徵比對、機率式神經網路(Probabilistic Neural Network)等技術,來判斷是否有人接近車子預謀不軌,並且透過4G網路提供矩形影像或各個視角的即時 影像與緊急通知使用者的功能,讓使用者可觀看汽車四周各個視角的畫面,做出最適當的處理,達成安全性極高的一套防盜系統,改善目前汽車監控防盜的不足之處。 According to the present invention, the stereoscopic panoramic image monitoring system of the present invention can capture the surrounding image of the mobile vehicle by using a plurality of camera modules, and use the control module to process the panoramic image through the image stitching. After reverting to a rectangular image, you can use Discrete Wavelet Transform, image processing and segmentation techniques, feature comparison, Probabilistic Neural Network and other techniques to determine whether someone is approaching the car. And provide a rectangular image or instant view of each perspective through the 4G network. The functions of images and emergency notification users allow users to view the various angles of view around the car, make the most appropriate treatment, and achieve a highly secure anti-theft system to improve the current inadequacy of car surveillance.
1‧‧‧監控系統 1‧‧‧Monitoring system
10‧‧‧攝像模組 10‧‧‧ camera module
100‧‧‧原始影像訊號 100‧‧‧ original image signal
101‧‧‧非原始影像訊號 101‧‧‧Non-original image signal
11‧‧‧控制模組 11‧‧‧Control module
110‧‧‧原始環景影像訊息 110‧‧‧ Original Image Information
110A‧‧‧原始影像訊息 110A‧‧‧ original image message
110B‧‧‧原始背景影像訊息 110B‧‧‧ Original background image message
1100‧‧‧第一特徵值 1100‧‧‧ first eigenvalue
111‧‧‧非原始環景影像訊息 111‧‧‧Non-original panoramic image information
111A‧‧‧非原始影像訊息 111A‧‧‧Non-original image messages
111B‧‧‧非原始背景影像訊息 111B‧‧‧ Non-original background image information
1110‧‧‧第二特徵值 1110‧‧‧second eigenvalue
112‧‧‧異常訊息 112‧‧‧Abnormal information
113‧‧‧局部影像訊息 113‧‧‧Partial image message
114‧‧‧預設人形影像訊息 114‧‧‧Preset humanoid image information
12‧‧‧傳輸模組 12‧‧‧Transmission module
13‧‧‧電源供應模組 13‧‧‧Power supply module
14‧‧‧警示模組 14‧‧‧Warning module
140‧‧‧警示訊息 140‧‧‧Warning message
2‧‧‧行動載具 2‧‧‧Mobile Vehicles
20‧‧‧影像 20‧‧‧ images
3‧‧‧外部電子裝置 3‧‧‧External electronic devices
30‧‧‧影像傳輸訊息 30‧‧‧Image transmission message
31‧‧‧啟動訊息 31‧‧‧Start message
第1圖係為本發明之立體環景影像之監控系統之第一實施例之第一示意圖。 1 is a first schematic view of a first embodiment of a stereoscopic panoramic image monitoring system of the present invention.
第2圖係為本發明之立體環景影像之監控系統之第一實施例之第二示意圖。 Figure 2 is a second schematic view of the first embodiment of the stereoscopic panoramic image monitoring system of the present invention.
第3圖係為本發明之立體環景影像之監控系統之第一實施例之方塊圖。 Figure 3 is a block diagram of a first embodiment of a stereoscopic panoramic image monitoring system of the present invention.
第4圖係為本發明之立體環景影像之監控系統之第二實施例之方塊圖。 Figure 4 is a block diagram of a second embodiment of a stereoscopic panoramic image monitoring system of the present invention.
第5圖係為本發明之立體環景影像之監控系統之第二實施例之示意圖。 Figure 5 is a schematic view showing a second embodiment of the stereoscopic panoramic image monitoring system of the present invention.
第6圖係為本發明之立體環景影像之監控系統之第二實施例之離散小波轉換運算流程圖。 Figure 6 is a flow chart of the discrete wavelet transform operation of the second embodiment of the stereoscopic panoramic image monitoring system of the present invention.
第7圖係為本發明之立體環景影像之監控系統之第三實施例之方塊圖。 Figure 7 is a block diagram showing a third embodiment of the stereoscopic panoramic image monitoring system of the present invention.
第8圖係為本發明之立體環景影像之監控系統之第三實施例之示意圖。 Figure 8 is a schematic view showing a third embodiment of the stereoscopic panoramic image monitoring system of the present invention.
第9圖係為本發明之立體環景影像之監控系統之第三實施例之機率式神經網路架構圖。 Figure 9 is a schematic diagram of a probability neural network architecture of a third embodiment of the stereoscopic panoramic image monitoring system of the present invention.
第10圖係為本發明之立體環景影像之監控系統之第四實施例之方塊圖。 Figure 10 is a block diagram showing a fourth embodiment of the stereoscopic panoramic image monitoring system of the present invention.
為利 貴審查員瞭解本發明之技術特徵、內容與優點及其所能達成之功效,茲將本發明配合附圖,並以實施例之表達形式詳細說明如下,而其中所使用之圖式,其主旨僅為示意及輔助說明書之用,未必為本發明實施後之真實比例與精準配置,故不應就所附之圖式的比例與配置關係侷限本發明於實際實施上的專利範圍,合先敘明。 The technical features, contents, and advantages of the present invention, as well as the advantages thereof, can be understood by the present inventors, and the present invention will be described in detail with reference to the accompanying drawings. The subject matter is only for the purpose of illustration and supplementary description. It is not necessarily the true proportion and precise configuration after the implementation of the present invention. Therefore, the scope and configuration relationship of the attached drawings should not be limited to the scope of patent application of the present invention. Narration.
以下將參照相關圖式,說明依本發明之立體環景影像之監控系統之實施例,為使便於理解,下述實施例中之相同元件係以相同之符號標示來說明。 The embodiments of the monitoring system for the stereoscopic panoramic image according to the present invention will be described below with reference to the related drawings. For the sake of understanding, the same components in the following embodiments are denoted by the same reference numerals.
請參閱第1圖至第3圖,其分別為本發明之立體環景影像之監控系統之第一實施例之第一示意圖、第二示意圖及方塊圖。如圖所示,立體環景影像之監控系統1,其應用於行動載具2,監控系統1包含複數個攝像模組10、控制模組11及傳輸模組12。複數個攝像模組10設置於行動載具2之周圍,並持續擷取行動載具2之外部環境之影像20,以產生複數個原始影像訊號100及複數個非原始影像訊號101。控制模組11電性連接複數個攝像模組10,控制模組11接收複數個原始影像訊號100及複數個非原始影像訊號101,並經由影像縫合處理,而分別產生原始環景影像訊息110及複數個非原始環景影像訊息111,且控制模組11經由影像比對處理比對原始環景影像訊息110及複數個非原始環景 影像訊息111,而判斷至少一非原始環景影像訊息111異常時,控制模組11則據以產生異常訊息112。傳輸模組12電性連接控制模組11,控制模組11經由傳輸模組12傳送異常訊息112至至少一外部電子裝置3。 Please refer to FIG. 1 to FIG. 3 , which are respectively a first schematic diagram, a second schematic diagram and a block diagram of a first embodiment of a stereoscopic panoramic image monitoring system according to the present invention. As shown in the figure, a stereoscopic panoramic image monitoring system 1 is applied to a mobile vehicle 2, and the monitoring system 1 includes a plurality of camera modules 10, a control module 11, and a transmission module 12. A plurality of camera modules 10 are disposed around the mobile vehicle 2 and continuously capture the image 20 of the external environment of the mobile device 2 to generate a plurality of original image signals 100 and a plurality of non-original image signals 101. The control module 11 is electrically connected to the plurality of camera modules 10, and the control module 11 receives the plurality of original image signals 100 and the plurality of non-original image signals 101, and respectively generates the original panoramic image information 110 by using image stitching processing. a plurality of non-original surround image messages 111, and the control module 11 compares the original ring image information 110 and the plurality of non-original scenes through the image comparison process When the image message 111 is determined, and the at least one non-original scene image message 111 is abnormal, the control module 11 generates the abnormality message 112 accordingly. The transmission module 12 is electrically connected to the control module 11 , and the control module 11 transmits the abnormality message 112 to the at least one external electronic device 3 via the transmission module 12 .
具體而言,本發明之立體環景影像之監控系統1可應用於行動載具2,如汽車或機車,於本實施例中係以汽車作為示例,但不以此為限。監控系統1包含複數個攝像模組10、控制模組11及傳輸模組12。複數個攝像模組10分別設置於行動載具2主體之四周,以行動載具2之外部環境之影像20,攝像模組10可為車用廣角攝影機,其具有140度以上的視角,且各攝像模組10間平均都有20度左右的重疊區域,攝像模組10數量較佳可為6個。控制模組11可為處理器,並具有記憶體,用於儲存影像處理程式,控制模組11電性連接複數個攝像模組10及傳輸模組12,傳輸模組12可為無線收發裝置,其可藉由wifi或藍芽傳輸訊息。 Specifically, the stereoscopic image monitoring system 1 of the present invention can be applied to the mobile vehicle 2, such as a car or a locomotive. In this embodiment, the automobile is taken as an example, but not limited thereto. The monitoring system 1 includes a plurality of camera modules 10, a control module 11, and a transmission module 12. A plurality of camera modules 10 are respectively disposed around the main body of the mobile vehicle 2 to move the image 20 of the external environment of the mobile vehicle 2. The camera module 10 can be a wide-angle camera for a vehicle, and has a viewing angle of 140 degrees or more, and each The camera module 10 has an average overlap area of about 20 degrees, and the number of the camera modules 10 is preferably six. The control module 11 can be a processor and has a memory for storing an image processing program. The control module 11 is electrically connected to a plurality of camera modules 10 and a transmission module 12, and the transmission module 12 can be a wireless transceiver device. It can transmit messages via wifi or Bluetooth.
因此,駕駛者停放好行動載具2後,啟動監控系統1後,複數個攝像模組10會持續擷取行動載具2之外部環境之影像20,並對應產生複數個原始影像訊號100及複數個非原始影像訊號101,其中原始影像訊號100可為第一次所擷取之影像,非原始影像訊號101則為後續所擷取之所有影像。接著,控制模組11接收複數個原始影像訊號100及複數個非原始影像訊號101,並經由影像縫合處理程序,而分別產生原始環景影像訊息110及複數個非原始環景影像訊息111,其中,原始環景影像訊息110與複數個非原始環景影像訊息111為360度全景鳥瞰環景影像。而後,控制模組11會經由影像比對處理比對原始環景影像訊息110及複數個非原始環景影像訊息111,以判斷是否代表有物體接近車子;而當其中一非原始環景影像訊息111出現物體時,控制模組11會判斷該非原始環景影像訊息111與原始環景影像訊息110比對異常,而 判斷該非原始環景影像訊息111為異常影像,此時,控制模組11會據以產生異常訊息112,如簡訊或郵件等,並經由傳輸模組12傳送異常訊息112至駕駛者之外部電子裝置3,以通知駕駛者有物體接近行動載具2周圍,進而達到警示作用。 Therefore, after the driver has parked the mobile vehicle 2, after the monitoring system 1 is activated, the plurality of camera modules 10 continuously captures the image 20 of the external environment of the mobile vehicle 2, and correspondingly generates a plurality of original image signals 100 and plural. The non-original image signal 101, wherein the original image signal 100 can be the first captured image, and the non-original image signal 101 is the subsequent captured image. Then, the control module 11 receives a plurality of original image signals 100 and a plurality of non-original image signals 101, and generates an original panoramic image message 110 and a plurality of non-original panoramic image messages 111, respectively, through an image stitching processing program. The original panoramic image message 110 and the plurality of non-original panoramic image messages 111 are 360-degree panoramic bird's-eye view images. Then, the control module 11 compares the original panoramic image information 110 and the plurality of non-original panoramic image information 111 through the image comparison to determine whether an object is close to the car; and when one of the non-original panoramic image messages When an object appears, the control module 11 determines that the non-original surround image information 111 is abnormal with the original panoramic image information 110, and The non-original panoramic image information 111 is determined to be an abnormal image. At this time, the control module 11 generates an abnormality message 112, such as a short message or an email, and transmits the abnormality message 112 to the driver's external electronic device via the transmission module 12. 3, to inform the driver that there is an object near the action vehicle 2, and thus achieve the warning effect.
請參閱第4圖至第6圖,其分別為本發明之立體環景影像之監控系統之第二實施例之方塊圖、示意圖及離散小波轉換運算流程圖。並請一併參閱第1圖至第3圖。如圖所示,本實施例中之立體環景影像之監控系統與上述第一實施例之立體環景影像之監控系統所述的相同元件的作動方式相似,故不在此贅述。然而,值得一提的是,在本實施例中,控制模組11可依據複數個原始影像訊號100而產生複數個原始影像訊息110A,控制模組11比對複數個原始影像訊息110A中之複數個第一特徵值1100且判斷符合預設門檻值時,控制模組11則據以將複數個原始影像訊息110A進行影像縫合而產生原始環景影像訊息110。且,控制模組11亦可依據複數個非原始影像訊號101而產生複數個非原始影像訊息111A,控制模組11比對複數個非原始影像訊息111A中之複數個第二特徵值1110且判斷符合預設門檻值時,控制模組11則據以將複數個非原始影像訊息111A進行影像縫合而產生非原始環景影像訊息111。 Please refer to FIG. 4 to FIG. 6 , which are respectively a block diagram, a schematic diagram and a flowchart of discrete wavelet transform operation of the second embodiment of the stereoscopic panoramic image monitoring system of the present invention. Please also refer to Figure 1 to Figure 3. As shown in the figure, the monitoring system of the stereoscopic panoramic image in the present embodiment is similar to the operation of the same component described in the monitoring system of the stereoscopic panoramic image of the first embodiment, and therefore will not be described herein. However, it is worth mentioning that, in this embodiment, the control module 11 can generate a plurality of original image messages 110A according to the plurality of original image signals 100, and the control module 11 compares the plurality of original image messages 110A. When the first feature value 1100 is determined and the preset threshold value is met, the control module 11 performs image stitching on the plurality of original image messages 110A to generate the original surround image information 110. The control module 11 can also generate a plurality of non-original image messages 111A according to the plurality of non-original image signals 101. The control module 11 compares and determines a plurality of second feature values 1110 of the plurality of non-original image messages 111A. When the preset threshold value is met, the control module 11 performs image stitching on the plurality of non-original image messages 111A to generate the non-original surround image information 111.
也就是說,控制模組11可先依據複數個原始影像訊號100與而產生複數個原始影像訊息110A。接著,控制模組11會先於各原始影像訊息110A中找尋相似的第一特徵值1100,例如車輛或盆栽等,當第一特徵值1100匹配的數量及位置符合的預設門檻值時,控制模組11就會以對應之角度及距離進行影像縫合,以產生原始環景影像訊息110。而非原始環景影像訊息111產生之過程亦與上述產生原始環景影像訊息110類似,故不再贅述。 In other words, the control module 11 can generate a plurality of original image messages 110A according to the plurality of original image signals 100. Then, the control module 11 searches for the first first feature value 1100, such as a vehicle or a potted plant, in the original image information 110A. When the first feature value 1100 matches the number and the preset threshold value of the position, the control module 11 controls. The module 11 will perform image stitching at a corresponding angle and distance to generate the original surround image information 110. The process of generating the original ring image information 111 is similar to the generation of the original scene image message 110, and therefore will not be described again.
進一步而言,本發明之控制模組11要將所有攝像模組10所拍到的外部環境影像作物體影像之判別前,必須將所有影像先縫合為一張矩形環景影像,如此一來只須將此環景影像匯入做判斷即可。但由於攝像模組10裝設時很容易因為機構精度的問題,導致影像無法精確縫合。因此,要將四張圖變形校正後之影像縫合(Image Stitching)成一張全景鳥瞰影像(即原始環景影像訊息110與非原始環景影像訊息111),必需知道複數個攝像模組10間在三維空間中的相對關係或是轉換成鳥瞰影像之後的對應關係。再利用鳥瞰轉換投影至相同虛擬影像平面即可完成拼接的動作。在本發明中採用特徵偵測及匹配的演算法,這個方法被稱為尺度不變性特徵轉換(Scale-invariant Feature Transform,SIFT)。SIFT特徵有旋轉不變、尺度不變的特性,並對仿射轉換(Affine Transform)也有良好的抵抗力。本發明運用SIFT特徵匹配與RANSAC演算法將影像對位完成,如第5圖所示。本發明先於兩張分開拍攝的原始影像訊息110A中找尋相似的特徵點,當特徵點匹配的數量及位置符合的門檻值時,就會以對應之角度及距離進行影像縫合。 Further, before the control module 11 of the present invention determines the external environment image crop image captured by all the camera modules 10, all the images must be first stitched into a rectangular ring image, so that only This scene image must be imported for judgment. However, since the camera module 10 is installed, it is easy to accurately stitch the image due to the problem of the accuracy of the mechanism. Therefore, it is necessary to know that a plurality of image capturing modules 10 are image stitching (image Stitching) into a panoramic bird's-eye view image (ie, the original ring image information 110 and the non-original ring image information 111). The relative relationship in the three-dimensional space is either converted into a correspondence after the bird's-eye view image. The stitching action can be completed by using the bird's eye view conversion projection to the same virtual image plane. In the present invention, a feature detection and matching algorithm is adopted, which is called Scale-invariant Feature Transform (SIFT). The SIFT features have the characteristics of rotation invariance and scale invariance, and also have good resistance to affine transformation (Affine Transform). The present invention uses the SIFT feature matching and the RANSAC algorithm to align the images, as shown in Figure 5. The present invention searches for similar feature points in the original image information 110A separately captured. When the number of feature points matches and the threshold value of the position is matched, the image is stitched at a corresponding angle and distance.
再者,本發明之立體環景影像之監控系統亦採用離散型小波轉換中的上提式Lifting5/3的方法做為縮減資料及頻譜切分的工具。Lifting5/3效能是DWT中最好的一種,運算複雜度極低,其運算步驟如第6圖所示,可大致分為兩個步驟: Furthermore, the stereoscopic panoramic image monitoring system of the present invention also uses the lifting Lifting 5/3 method in the discrete wavelet transform as a tool for reducing data and spectrum segmentation. Lifting5/3 performance is the best of DWT, and its computational complexity is extremely low. The operation steps, as shown in Figure 6, can be roughly divided into two steps:
分離步驟 Separation step
在此步驟中,假設具有m筆資料之向量變數d(例如:擷取之影像),可被切分為兩個部分:機數點d 2i+1以及偶數點d 2i ,並分別以以及表示,運算方式(1)說明如下: In this step, a vector variable d (for example, an image captured) having m pen data is assumed to be divided into two parts: the machine point d 2 i +1 and the even point d 2 i , respectively as well as Indicates that the calculation method (1) is as follows:
上提步驟 Step up
此步驟科分離出高頻成分以及低頻成分,運算方式(2)表示如下: This step separates high frequency components And low frequency components The operation mode (2) is expressed as follows:
本發明利用上述簡單的兩個步驟,經過二次的小波轉換後稱為二階小波轉換。低頻的部分會僅剩下原來影像資料量的四分之一大小,更重要的是特徵並無明顯的減少,能夠有效的減少往後的複雜演算法運算時間,同時期高頻的部分也被分離保留下來,可作為後續訊號分析之用。 The present invention utilizes the above two simple steps and is referred to as a second-order wavelet transform after a second wavelet transform. The low-frequency part will only have a quarter of the original image data, and more importantly, there will be no significant reduction in features, which can effectively reduce the complexity of later complex algorithms, while the high-frequency part is also Separation is retained for subsequent signal analysis.
請參閱第7圖至第9圖,其分別為本發明之立體環景影像之監控系統之第三實施例之方塊圖、示意圖及機率式神經網路架構圖。並請一併參閱第1圖至第6圖。如圖所示,本實施例中之立體環景影像之監控系統與上述各實施例之立體環景影像之監控系統所述的相同元件的作動方式相似,故不在此贅述。然而,值得一提的是,在本實施例中,控制模組11可以背景相減法將原始環景影像訊息110中之原始背景影像訊息110B與複數個非原始環景影像訊息111中之非原始背景影像訊息111B進行相減,而產生至少一局部影像訊息113,在控制模組11經由分類程式判斷至少一局部影像訊息113符合預設人形影像訊息114時,控制模組11則據以產生異常訊息112。 Please refer to FIG. 7 to FIG. 9 , which are respectively a block diagram, a schematic diagram and a schematic neural network architecture diagram of a third embodiment of the stereoscopic panoramic image monitoring system of the present invention. Please also refer to Figure 1 to Figure 6. As shown in the figure, the monitoring system of the stereoscopic panoramic image in the present embodiment is similar to the operation of the same component described in the monitoring system of the stereoscopic panoramic image of the above embodiments, and therefore will not be described herein. However, it is worth mentioning that, in this embodiment, the control module 11 can use the background subtraction method to convert the original background image message 110B in the original surround view image message 110 with the non-original of the plurality of non-original surround view image messages 111. The background image message 111B is subtracted to generate at least one partial image message 113. When the control module 11 determines through the classification program that the at least one partial image message 113 conforms to the preset humanoid image message 114, the control module 11 generates an abnormality. Message 112.
舉例而言,本發明之控制模組11可進一步以背景相減法將原始環景影像訊息110中之原始背景影像訊息110B與複數個非原始環景影像訊息111中之非原始背景影像訊息111B進行相減,相互抵消,而將僅出現於非原始環景影像訊息111中而未出現原始環景影像訊息110中的物體影像輸出而產生至少一局部影像訊息113。接著,控制模組11會經由分類程式將至少一局部影像訊息113與預設人形影像訊息114進 行比對;當控制模組11判斷至少一局部影像訊息113符合預設人形影像訊息114時,控制模組11則會據以產生異常訊息112,並經由傳輸模組12傳送至外部電子裝置3。 For example, the control module 11 of the present invention may further perform the background subtraction method on the original background image message 110B in the original surround image message 110 and the non-original background image message 111B in the plurality of non-original surround image messages 111. Subtracting, canceling each other, will occur only in the non-original surround image message 111 without the object image output in the original surround image message 110 generating at least one partial image message 113. Then, the control module 11 enters at least one partial image message 113 and the preset humanoid image message 114 via the classification program. When the control module 11 determines that the at least one partial image message 113 matches the preset humanoid image message 114, the control module 11 generates an abnormality message 112 and transmits it to the external electronic device 3 via the transmission module 12. .
進一步而言,本發明之立體環景影像之監控系統1在判定車子周圍是否有物體靠近,必須將當下之影像與先前建立好的原始影像之不同之處找出,本發明使用多種影像前處理技術,再以背景相減法來分割出與原始影像相異之處的局部影像,再做後續的人形判定之處理。 Further, the stereoscopic panoramic image monitoring system 1 of the present invention determines whether there is an object approaching around the car, and must find out the difference between the current image and the previously established original image. The present invention uses various image pre-processing. The technique then uses the background subtraction method to segment the partial image that is different from the original image, and then performs the subsequent humanoid judgment.
本發明採用機率式神經網路,進行人形判定之處理。機率神經網路的架構基本上,是一種四層神經元結構的網路模型包含:輸入層(Input Layer)、類別層(Pattern Layer)、總和層(Summation Layer)與輸出層(Output Layer),屬於前向式的神經網路架構的一種。機率神經網路主要的理論基礎建立在於貝氏決策(Bayes decision)上,主要的應用是分類器。機率神經網路架構最重要的特色在於網路訓練的即時性,所以機率神經網路適合運用在即時的系統。其決策面的範圍大小可以依據問題的需求不同進行調整,且決策面可逼近貝氏分類器。機率神經網路對於錯誤及雜訊容忍度很高。在稀疏的樣本空間問題上,平滑參數可依據問題的需求,隨時調整參數的大小,而無需進行重新訓練的工作。 The invention adopts a probabilistic neural network to perform the process of human shape determination. The architecture of the probabilistic neural network is basically a network model of a four-layer neuron structure: an input layer, a pattern layer, a summation layer, and an output layer. A type of neural network architecture that is forward-oriented. The main theoretical basis of the probabilistic neural network is based on the Bayes decision. The main application is the classifier. The most important feature of the probabilistic neural network architecture is the immediacy of network training, so the probabilistic neural network is suitable for use in real-time systems. The size of the decision surface can be adjusted according to the needs of the problem, and the decision surface can approach the Bayesian classifier. Probabilistic neural networks are highly tolerant of errors and noise. On the sparse sample space problem, the smoothing parameters can adjust the size of the parameters at any time according to the needs of the problem, without the need for retraining work.
本發明所採用機率神經網路從原始影像經過背景相減法分割出局部影像後,將資料輸入至PNN分類器分類,分類結果為判定是否為人形的基準,(與樣本人形相似之數量),其應用於人形偵測之說明如下: 對於一個可能入侵影像輸入Xx={g 1,g 2,...,g m } (3) The probability neural network used in the present invention divides the partial image from the original image through the background subtraction method, and then inputs the data into the PNN classifier classification, and the classification result is a reference for determining whether it is a human shape (a quantity similar to the sample human form), The description applied to humanoid detection is as follows: For a possible intrusion image input X : x ={ g 1 , g 2 ,..., g m } (3)
其中m為輸入資料的數量,在本研究即為每一筆分割出來的局部影像資料。 Where m is the number of input data, in this study, each segmented partial image data.
假設PNN具有類別向量c(訓練資料):c={c 1,c 2,...,c r }and c i ={y i1,y i2,...,y im } (4) Suppose the PNN has a category vector c (training data): c = { c 1 , c 2 ,..., c r }and c i ={ y i 1 , y i 2 ,..., y im } (4)
其中r為類別c的數量,m為每個類別內的資料數量。 Where r is the number of categories c and m is the number of items in each category.
Parzen是使用一個特徵估測一個類別,對於訓練資料中的每一樣本建立一個以樣本的特徵值為中心的高斯曲線,最後把所有建立的曲線疊加成一個屬於該類別的機率密度函數。若是要用在任意維度的問題上則可以將機率密度函數表示成: Parzen uses a feature to estimate a category, establishes a Gaussian curve centered on the sample's eigenvalues for each sample in the training data, and finally stacks all the established curves into a probability density function belonging to that category. If you want to use it in any dimension, you can express the probability density function as:
其中d為訓練向量的維度,σ為高斯函數之平滑係數(Soothing Parameter)。 Where d is the dimension of the training vector and σ is the Soothing Parameter of the Gaussian function.
實作時部分可視為常數忽略不計,因此最大值(最接近)輸出結果P表示如下: When it is implemented Partially visible as a constant ignore, so the maximum (closest) output P is expressed as follows:
P為判定是否為人形的參數,當輸入之影像資料與訓練樣本資料相似值越高時,其機率密度值就越高,最後當P大於一定的值時,便判定該影像為人形影像。由上述過程可知,本發明所採用之架構預期可以有效成為人形偵測判別演算方法。再者,機率類神經網路的學習過程為零,直接從訓練樣本中讀取所需數據,不需要像傳統類神經網路迭代的學習過程,對於記憶體空間需求較大,對於本發明將PNN實作於微控制器上非常適合。現代微控制器都可外加大量的快閃記憶體,而且成本低廉。完成本發明之架構後只需將訓練資料放入快閃記憶體,以空間換取時間,兼具效能提升與降低成本的優勢。 P is a parameter for determining whether it is a human figure. When the input image data and the training sample data have higher similarity values, the probability density value is higher. Finally, when P is greater than a certain value, the image is determined to be a humanoid image. It can be seen from the above process that the architecture adopted by the present invention is expected to be effectively a humanoid detection discriminant calculation method. Furthermore, the learning process of the probability-like neural network is zero, and the required data is directly read from the training samples, and the learning process like the traditional neural network iteration is not required, and the memory space is relatively large, and the present invention will PNN is very suitable for implementation on a microcontroller. Modern microcontrollers can add a large amount of flash memory at a low cost. After completing the architecture of the present invention, it is only necessary to put the training materials into the flash memory, and exchange space for time, which has the advantages of performance improvement and cost reduction.
請參閱第10圖,其係為本發明之立體環景影像之監控系統之第四實施例之方塊圖。並請一併參閱第1圖至第9圖。如圖所示,本實施例中之立體環景影像之監控系統與上述各實施例之立體環景影像之監控系統所述的相同元件的作動方式相似,故不在此贅述。然而,值得一提的是,在本實施例中,傳輸模組12接收至少一外部電子裝置3所傳輸之影像傳輸訊息30時,控制模組11則據以經由傳輸模組12傳送至少一非原始環景影像訊息111B至至少一外部電子裝置3。也就是說,在駕駛者經由外部電子裝置3傳輸影像傳輸訊息30至傳輸模組12,控制模組11則可據以控制傳輸模組12傳送至少一非原始環景影像訊息111B至至少一外部電子裝置3,抑或控制模組11可控制傳輸模組12與外部電子裝置3進行通訊連結,並將複數個攝像模組10所擷取的即時影像傳輸至外部電子裝置3,以供駕駛者觀看。 Please refer to FIG. 10, which is a block diagram of a fourth embodiment of a stereoscopic panoramic image monitoring system of the present invention. Please also refer to Figures 1 to 9. As shown in the figure, the monitoring system of the stereoscopic panoramic image in the present embodiment is similar to the operation of the same component described in the monitoring system of the stereoscopic panoramic image of the above embodiments, and therefore will not be described herein. However, it is worth mentioning that, in the embodiment, when the transmission module 12 receives the image transmission message 30 transmitted by the external electronic device 3, the control module 11 transmits at least one non-transmission via the transmission module 12. The original panoramic image message 111B is to at least one external electronic device 3. That is, after the driver transmits the image transmission message 30 to the transmission module 12 via the external electronic device 3, the control module 11 can control the transmission module 12 to transmit at least one non-original scene image message 111B to at least one external portion. The electronic device 3 or the control module 11 can control the communication module 12 to communicate with the external electronic device 3, and transmit the instant images captured by the plurality of camera modules 10 to the external electronic device 3 for the driver to watch. .
藉此,當駕駛者收到異常訊息112時,可立即使用智慧型裝置觀看目前車子周遭的影像。本發明可將異常訊息112所拍攝之影 像,經過影像處理方式進行鏡頭校正、幾何校正及影像縫合等數位影像處理,透過立體座標投射演算法將2D影像像素轉換至對應之3D座標位置,且可切換各種不同角度供使用者觀看。因此,駕駛者收到異常訊息112時,便可透過網際網路連線至立體環景影像之監控系統1觀看立體環景影像,再依當下需求查看不同角度的影像,清楚了解當下車子周遭的人、事、物,做出最正確的應對動作。 Thereby, when the driver receives the abnormality message 112, the smart device can be used to immediately view the image of the current car. The invention can take the shadow of the abnormal message 112 For example, image processing is used for digital image processing such as lens correction, geometric correction, and image stitching, and the 2D image pixels are converted to corresponding 3D coordinate positions through a stereo coordinate projection algorithm, and various angles can be switched for the user to view. Therefore, when the driver receives the abnormality message 112, he can view the stereoscopic scene image through the Internet to the monitoring system 1 of the stereoscopic panoramic image, and then view the images of different angles according to the current demand, and clearly understand the surroundings of the current car. People, things, things, make the most correct response.
進一步而言,傳輸模組12可接收至少一外部電子裝置3所傳輸之啟動訊息31時,控制模組11則據以控制複數個攝像模組10進行影像擷取之作動,並依據複數個原始影像訊號100及複數個非原始影像訊號101進行影像縫合處理與影像比對處理。也就是說,在駕駛者停放好車輛後,欲啟動立體環景影像之監控系統1,可經由外部電子裝置3傳輸啟動訊息31,以驅使立體環景影像之監控系統1進行監控作動。 Further, when the transmission module 12 can receive the activation message 31 transmitted by the at least one external electronic device 3, the control module 11 controls the operation of the image capturing module by the plurality of camera modules 10, and according to the plurality of originals. The image signal 100 and the plurality of non-original image signals 101 perform image stitching processing and image comparison processing. That is to say, after the driver parks the vehicle, the monitoring system 1 for initiating the stereoscopic panoramic image can transmit the activation message 31 via the external electronic device 3 to drive the monitoring system 1 of the stereoscopic panoramic image to monitor and act.
進一步而言,本發明之立體環景影像之監控系統1較佳更可包含電源供應模組13,其電性連接複數個攝像模組10、控制模組11及傳輸模組12,電源供應模組13供給電能至複數個攝像模組10、控制模組11及傳輸模組12。也就是說,立體環景影像之監控系統1進一步還可設置電源供應模組13,其可為可重複充放電之電池組件,用於提供立體環景影像之監控系統1各模組運作所需之電能。 Further, the stereoscopic image monitoring system 1 of the present invention preferably further includes a power supply module 13 electrically connected to the plurality of camera modules 10, the control module 11 and the transmission module 12, and the power supply module. The group 13 supplies electric energy to a plurality of camera modules 10, a control module 11, and a transmission module 12. In other words, the monitoring system 1 of the stereoscopic panoramic image can further be provided with a power supply module 13, which can be a rechargeable battery assembly for providing three-dimensional panoramic image monitoring system 1 The power.
值得一提的是,本發明之立體環景影像之監控系統1亦可與行動載具2之電池組件進行電性連結,而可無須再設置電源供應模組13。 It should be noted that the stereoscopic image monitoring system 1 of the present invention can also be electrically connected to the battery component of the mobile carrier 2, and the power supply module 13 can be omitted.
進一步而言,立體環景影像之監控系統1較佳更可包含警示模組14,其電性連接控制模組11,在控制模組11判斷至少一非原始環景影像訊息111異常時,控制模組11則據以控制警示模組14產生警 示訊息140。也就是說,本發明之立體環景影像之監控系統1進一步還可設置警示模組14,因此,在控制模組11判斷至少一非原始環景影像訊息111異常時,即可藉由警示模組14產生警示訊息140,如聲音、光線或其組合,以達到嚇阻竊賊之功效。 Further, the stereoscopic panoramic image monitoring system 1 preferably further includes an alert module 14 electrically connected to the control module 11 for controlling when the control module 11 determines that at least one non-original surround image message 111 is abnormal. The module 11 generates a warning according to the control warning module 14 A message 140 is displayed. In other words, the monitoring system 1 of the stereoscopic panoramic image of the present invention can further be provided with the warning module 14. Therefore, when the control module 11 determines that at least one non-original panoramic image information 111 is abnormal, the warning mode can be used. Group 14 generates a warning message 140, such as sound, light, or a combination thereof, to defeat the thief.
本發明之立體環景影像之監控系統1,以四周裝設之廣角攝影機,將攝影機所拍攝之影像透過多種影像處理方式產生如同由車外觀看汽車本體之全景視角立體影像;再將此環景影像透過處理,使用數位影像處理技術進行影像前處理,再以類神經網路方法判斷影像是否代表有人接近車子。上述情形發生時,監控系統1將會透過行動通訊網路傳送訊息給駕駛者之行動裝置,使其可以即時觀看車體附近360度環景影像,並可以調整各種視角,達到即時防盜監控目的。 The stereoscopic panoramic image monitoring system 1 of the present invention uses a wide-angle camera installed around the camera to generate a panoramic view stereoscopic image of the car body through a plurality of image processing modes through the image taken by the camera; Through processing, digital image processing technology is used for image pre-processing, and then the neural network method is used to determine whether the image represents someone approaching the car. When the above situation occurs, the monitoring system 1 will transmit a message to the driver's mobile device through the mobile communication network, so that it can instantly view the 360-degree panoramic image near the vehicle body, and can adjust various viewing angles to achieve the purpose of real-time anti-theft monitoring.
以上所述僅為舉例性,而非為限制性者。任何未脫離本發明之精神與範疇,而對其進行之等效修改或變更,均應包含於後附之申請專利範圍中。 The above is intended to be illustrative only and not limiting. Any equivalent modifications or alterations to the spirit and scope of the invention are intended to be included in the scope of the appended claims.
1‧‧‧監控系統 1‧‧‧Monitoring system
10‧‧‧攝像模組 10‧‧‧ camera module
100‧‧‧原始影像訊號 100‧‧‧ original image signal
101‧‧‧非原始影像訊號 101‧‧‧Non-original image signal
11‧‧‧控制模組 11‧‧‧Control module
110‧‧‧原始環景影像訊息 110‧‧‧ Original Image Information
111‧‧‧非原始環景影像訊息 111‧‧‧Non-original panoramic image information
112‧‧‧異常訊息 112‧‧‧Abnormal information
12‧‧‧傳輸模組 12‧‧‧Transmission module
3‧‧‧外部電子裝置 3‧‧‧External electronic devices

Claims (8)

  1. 一種立體環景影像之監控系統,其應用於一行動載具,該監控系統包含:複數個攝像模組,其設置於該行動載具之周圍,並持續擷取該行動載具之外部環境之影像,以產生複數個原始影像訊號及複數個非原始影像訊號;一控制模組,電性連接該複數個攝像模組,該控制模組接收該複數個原始影像訊號及該複數個非原始影像訊號,並經由一影像縫合處理,而分別產生一原始環景影像訊息及複數個非原始環景影像訊息,在該控制模組經由一影像比對處理將該原始環景影像訊息與各該非原始環景影像訊息進行比對且該控制模組而判斷至少一該非原始環景影像訊息異常時,該控制模組則據以產生一異常訊息;以及一傳輸模組,其電性連接該控制模組,該控制模組經由該傳輸模組傳送該異常訊息至至少一外部電子裝置。 A stereoscopic panoramic image monitoring system is applied to a mobile vehicle, the monitoring system comprising: a plurality of camera modules disposed around the mobile vehicle and continuously capturing the external environment of the mobile vehicle The image is configured to generate a plurality of original image signals and a plurality of non-original image signals; a control module electrically connecting the plurality of camera modules, the control module receiving the plurality of original image signals and the plurality of non-original images And generating, by an image stitching process, an original panoramic image message and a plurality of non-original panoramic image messages, wherein the original image and the non-original image information are processed by the control module via an image comparison process When the scene image is compared and the control module determines that at least one of the non-original scene images is abnormal, the control module generates an abnormal message; and a transmission module electrically connected to the control module The control module transmits the abnormal message to the at least one external electronic device via the transmission module.
  2. 如申請專利範圍第1項所述之立體環景影像之監控系統,其中該控制模組依據該複數個原始影像訊號而產生複數個原始影像訊息,該控制模組比對該複數個原始影像訊息中之複數個第一特徵值且判斷符合一預設門檻值時,該控制模組則據以將該複數個原始影像訊息進行影像縫合而產生該原始環景影像訊息。 The monitoring system of the three-dimensional panoramic image as described in claim 1, wherein the control module generates a plurality of original image messages according to the plurality of original image signals, and the control module compares the plurality of original image messages The plurality of first feature values are determined to meet a preset threshold value, and the control module performs image stitching on the plurality of original image messages to generate the original surround image information.
  3. 如申請專利範圍第1項所述之立體環景影像之監控系統,其中該控制模組依據該複數個非原始影像訊號而產生複數個非原始影像訊息,該控制模組比對該複數個非原始影像 訊息中之複數個第二特徵值且判斷符合一預設門檻值時,該控制模組則據以將該複數個非原始影像訊息進行影像縫合而產生該非原始環景影像訊息。 The monitoring system of the three-dimensional panoramic image as described in claim 1, wherein the control module generates a plurality of non-original image messages according to the plurality of non-original image signals, and the control module compares the plurality of non-original image information Original image When the plurality of second eigenvalues in the message are determined to meet a preset threshold, the control module performs image stitching on the plurality of non-original image messages to generate the non-original surround image information.
  4. 如申請專利範圍第1項所述之立體環景影像之監控系統,其中該控制模組以一背景相減法將該原始環景影像訊息中之一原始背景影像訊息與該複數個非原始環景影像訊息中之一非原始背景影像訊息進行相減,而產生至少一局部影像訊息,在該控制模組經由一分類程式判斷該至少一局部影像訊息符合一預設人形影像訊息時,該控制模組則據以產生該異常訊息。 The monitoring system of the stereoscopic panoramic image as described in claim 1, wherein the control module uses a background subtraction method to convert the original background image information of the original panoramic image information with the plurality of non-original scenery The non-original background image message is subtracted from the image message to generate at least one partial image message. When the control module determines, by the classification module, that the at least one partial image message conforms to a preset humanoid image message, the control mode The group generates the exception message accordingly.
  5. 如申請專利範圍第1項所述之立體環景影像之監控系統,其中該傳輸模組接收該至少一外部電子裝置所傳輸之一影像傳輸訊息時,該控制模組則據以經由該傳輸模組傳送至少一該非原始環景影像訊息至該至少一外部電子裝置。 The monitoring system of the stereoscopic panoramic image as described in claim 1, wherein the transmission module receives the image transmission message transmitted by the at least one external electronic device, and the control module passes the transmission mode The group transmits at least one of the non-original surround image information to the at least one external electronic device.
  6. 如申請專利範圍第1項所述之立體環景影像之監控系統,其中該傳輸模組接收該至少一外部電子裝置所傳輸之一啟動訊息時,該控制模組則據以控制該複數個攝像模組進行影像擷取之作動,並依據該複數個原始影像訊號及該複數個非原始影像訊號進行該影像縫合處理與該影像比對處理。 The monitoring system of the stereoscopic panoramic image as described in claim 1, wherein the control module controls the plurality of cameras when receiving the activation message transmitted by the at least one external electronic device The module performs image capture operation, and performs image stitching processing and the image comparison processing according to the plurality of original image signals and the plurality of non-original image signals.
  7. 如申請專利範圍第1項所述之立體環景影像之監控系統,其更包含一電源供應模組,其電性連接該複數個攝像模組、該控制模組及該傳輸模組,該電源供應模組供給電能至該複數個攝像模組、該控制模組及該傳輸模組。 The monitoring system of the three-dimensional panoramic image as described in claim 1, further comprising a power supply module electrically connected to the plurality of camera modules, the control module and the transmission module, the power source The supply module supplies power to the plurality of camera modules, the control module, and the transmission module.
  8. 如申請專利範圍第1項所述之立體環景影像之監控系統,其更包含一警示模組,其電性連接該控制模組,在該控制模組判斷至少一該非原始環景影像訊息異常時,該控制模組則據以控制該警示模組產生一警示訊息。 The monitoring system of the stereoscopic panoramic image as described in claim 1, further comprising a warning module electrically connected to the control module, wherein the control module determines that at least one of the non-original panoramic image information is abnormal The control module generates a warning message according to the control module.
TW104139564A 2015-11-27 2015-11-27 Surveillance system of 3d panoramic images TWI601426B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
TW104139564A TWI601426B (en) 2015-11-27 2015-11-27 Surveillance system of 3d panoramic images

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
TW104139564A TWI601426B (en) 2015-11-27 2015-11-27 Surveillance system of 3d panoramic images

Publications (2)

Publication Number Publication Date
TW201720141A TW201720141A (en) 2017-06-01
TWI601426B true TWI601426B (en) 2017-10-01

Family

ID=59687159

Family Applications (1)

Application Number Title Priority Date Filing Date
TW104139564A TWI601426B (en) 2015-11-27 2015-11-27 Surveillance system of 3d panoramic images

Country Status (1)

Country Link
TW (1) TWI601426B (en)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
TWI643497B (en) * 2017-12-01 2018-12-01 中興保全股份有限公司 Self-maintenance system for image monitoring

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
TW201348036A (en) * 2012-05-31 2013-12-01 Yottastor Information Technology Inc Vehicle monitoring system
CN104463778A (en) * 2014-11-06 2015-03-25 北京控制工程研究所 Panoramagram generation method
TW201537977A (en) * 2014-03-21 2015-10-01 Inventec Appliances Corp Panoramic scene capturing and browsing mobile device, system and method
CN103366569B (en) * 2013-06-26 2015-10-07 东南大学 The method and system of real-time grasp shoot traffic violation vehicle
TWM520036U (en) * 2015-11-27 2016-04-11 樹德科技大學 Surveillance system of 3D panoramic images

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
TW201348036A (en) * 2012-05-31 2013-12-01 Yottastor Information Technology Inc Vehicle monitoring system
CN103366569B (en) * 2013-06-26 2015-10-07 东南大学 The method and system of real-time grasp shoot traffic violation vehicle
TW201537977A (en) * 2014-03-21 2015-10-01 Inventec Appliances Corp Panoramic scene capturing and browsing mobile device, system and method
CN104463778A (en) * 2014-11-06 2015-03-25 北京控制工程研究所 Panoramagram generation method
TWM520036U (en) * 2015-11-27 2016-04-11 樹德科技大學 Surveillance system of 3D panoramic images

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
TWI643497B (en) * 2017-12-01 2018-12-01 中興保全股份有限公司 Self-maintenance system for image monitoring

Also Published As

Publication number Publication date
TW201720141A (en) 2017-06-01

Similar Documents

Publication Publication Date Title
TWM520036U (en) Surveillance system of 3D panoramic images
US10885398B2 (en) Joint 3D object detection and orientation estimation via multimodal fusion
CN107792059B (en) Parking control
US10812739B2 (en) Image processing device, image processing method, and image processing system
WO2018021067A1 (en) Image processing device and image processing method
US10684626B1 (en) Handling intersection navigation without traffic lights using computer vision
AU2018311661B2 (en) Method and apparatus for distributed edge learning
US20080158357A1 (en) Apparatus for driver analysis using rearview mirror
JP2015007952A (en) Device and method to detect movement of face to create signal, and computer readable storage medium
WO2018021068A1 (en) Image processing device and image processing method
TWI601426B (en) Surveillance system of 3d panoramic images
JP6944133B2 (en) Image processing device and image processing method
JP6944137B2 (en) Image processing device and image processing method
Yong et al. Real-time vehicle detection based on Haar features and Pairwise Geometrical Histograms
JP2014146267A (en) Pedestrian detection device and driving support device
Soundrapandiyan et al. Robust pedestrian detection in infrared images using rotation and scale invariant-based structure element descriptor
JP6847709B2 (en) Camera devices, detectors, detection systems and mobiles
CN110636263A (en) Panoramic annular view generation method, vehicle-mounted equipment and vehicle-mounted system
Kemsaram et al. A stereo perception framework for autonomous vehicles
Wen et al. Improved Haar wavelet feature extraction approaches for vehicle detection
Satzoda et al. Vision-based front and rear surround understanding using embedded processors
US11193312B1 (en) Child safety lock
Hu et al. A video streaming vehicle detection algorithm based on YOLOv4
Nair et al. Driver assistant system using Haar cascade and convolution neural networks (CNN)
JP2021152966A (en) Seat belt wearing state recognition method, recognition apparatus, electronic device, and recording medium

Legal Events

Date Code Title Description
MM4A Annulment or lapse of patent due to non-payment of fees