201140509 、發明說明： 【發明所屬之技術領域】 μ7~''種危害侧及控制线及其方法，特 _及⑽測Μ制多重危害態樣的系統及方法。 【先前技術】 為口應居豕及城市環境安全，監控系統的設置需求迅 速地增加。在目前的數位監控系統中，主要分為以電腦伺 服^控為基礎之架構，以及單機型操作架構兩大類贺。 i者易方、開叙與更新，加上網路功能及介面擴充功能 料通常具有基礎的内容分析能力 :因1於執行複雜而多樣化的監控作業。反之，後者則 ^ =BB>{、專用電路或數位訊號處理祕配勤體及其他 惊存认備運作，具有結構簡單、效能穩定的特色，較適合 長時間性、毅性高的環境執行歡監控作業。 【發明内容】 、、本發明的目的係為了提供—種多重危害偵測系統及其 方j用以對同一個受偵測的地點，同時判斷多種不同危 =態樣發生與否並加崎示，錢在不增加人力成本的前 提下、’全面性地保障受偵測地點的人身及財物安全。 為了達成上述目的，根據本發明的其中一項方案，提 供種夕重危害债測及控制系統，用以控制防災設施執行 防止危害的措施，防災設施包括一第一防災裝置及一第二 防X裝置。該系統包括：一偵測裝置和一控制中心。偵測 裝置包括：擷取受測地點之偵測影像的一影像擷取單元； 4/28 201140509 一分析模組根據複數種影像分析邏輯分別分 並依照各影像分析賴分_絲赵1像分析3 ; 受：地點是否發生危害態樣;其中，影像分析邏 匕括人體影像分析邏輯，用以分析悄測影像是 人體影像，且影像分析資訊包括人體影像資訊，用以指示 =體影像分析邏輯的分析結果；1林元，根據影^ =貧訊，於受測地點發生危害態樣時，產生包括人體 貝況的-警不訊息；及一傳輸單元輸出警示訊息。控制 心接收及解讀警示訊息’並根據警示訊息啟動防災ς施。 其中，人體影像資訊指示出偵測影像包括人體影像日^,控 制中心啟動第—防災裝置，反之則啟動第二防災裝置。工 為了達成上述目的，根據本發明的另一項方案〜提供 另一種多重危㈣測及控制系統，用以控制—防災設施執 打防止危害的措施，該防災設施包括一第—防災裴 第二防災f置’該系統包括：—_裝置和—控制中心。 r其中i偵測裝置包括：擷取受測地點之谓測影像的-影像擷取單元、❹枝測魅之環賴數的—感測軍元， 以及包括-影像分析模組及—變數分析模組的—分 。影像分析模組根據魏種影像分析邏輯 測景 像，並依照分析的結果產生1像分析資訊； t析邏輯包括—人體影像分析邏輯，用以分析該偵測=像 疋否包括-人體影像’且影像分析f訊包括—人體參像 訊，用以指示人體影像分析邏輯的分析結果；變數^析模 組根據至少—分析分析魏魏 、 結果產生一變數分拚眘1 U 队·、、'刀析的 資⑽㈣八把 分析模組依據影像分析 貝況及义數刀析賴產生—危害分析資訊，用以判斷受測 5/28 201140509 • .· 地點是否發生—危害態樣。偵測裝置還包括：— ，根據危害分析資訊’於受測地點發生危害態樣日。夺:、： 包，人㈣像訊的—警示訊息；及—傳輸單元 到控制中心。控制中心接收及解; _影像包括人體影像時，㈣中心啟動第 反之則啟動第二防災裝置。 人凌置， 為了達成上述目的，根據本發明的另—項方 =:2害_及控制方法’具有叫貞測裝置及—控 、 ^ •偵測裝置接收一受測地點之偵測影像； 測裝置以複數種影像分析邏輯分別分析偵測影像，盆中，、 影像分析邏輯包括—人體影像分析邏輯，用以伯 ^是否包括-人體影像，且影像分析f訊包括—人體影^ 資訊:用以指示人體影像分析邏輯的分析結果；根.二像 分析邏輯分別分析顺得的結果，產生—影像分析資訊，、 用以指出受測地點是否發生一危害態樣，其中，影像分析 貪讯包括人體影像資訊；偵測裝置依據分析的結果判斷受 測地點是否發生危害態樣；當偵測裝置判斷出受測地點ς 生危害態樣時，發送一警示訊息，警示訊息中包括人體^ 像資訊；控制中心接收及解讀警示訊息，並根據人體影^ 資訊而啟動一第一防災裝置或一第二防災裝置，以防止危 害態樣。 關於本發明之技術手段的詳細說明，請參閱以下的實 施方式，並配合所附圖式一併參照。 6/28 201140509 【實施方式】 本發明係為了協助使用者偵測和防範威脅到_受_ 點之安全的危害態樣，而提供—種可同時探測 樣的偵測及控制機制。 ①σ a 第4二圖;示了本發明的1多重危害偵測及控制系統 弟實鈪例方塊圖，所述的多重危害偵測及控制 下簡稱_及控㈣統）丨包括：一侧裝置lQ及一 中心20。偵測裝置1〇可依照需要置每— 二 宝伯偏猫故^ 戈成直万、母個需要接收危 。偵測和預誓的受測地點，以偵測受測地點是否 一 種危害態樣的產生，例如：水患、火災、非法人侵等= 2，置1 〇皆連接到控制中心20，以便於_ ^危害: X &，可通知控制中心20進行必要的應對措施。‘ 偵測裝置10包括：-影像掏取單元1〇〇、—分析模植 :—警示單S104、一傳輸單元106及一記憶單元⑽ 二?麻單元⑽可為—攝影裝置，用以定時或連續操 =雙測地點（如倉庫或廠房）的-偵測影像，傳送給分 =¼组102進行影像的分析辨識，以及將擷取到的偵測影 士己傳，到錢單元⑽中記錄。分析模組102讀取記錄在 =隱早tl 1G8中的複數種影像分析邏輯，分別將偵測影像 7f個影像分析邏輯進行分析和計算處理。影像分析邏 可包括：火災影像分析邏輯、水災影像分析邏輯及入 曰w像刀析邏輯等多種演算邏輯，分別用於辨識受測地點 疋否=生火炎、水災或人員非法人侵等危害態樣。 刀析模組102根據不同的影像分析邏輯一一對接收到 像，測=像進行運算處理後’將產生—影像分析資訊。影 义析貧訊综合所有執行過的分析邏輯的結論，用以判斷 7/28 201140509201140509, invention description: [Technical field of invention] μ7~'' hazard side and control line and its method, special _ and (10) system and method for measuring multiple hazard patterns. [Prior Art] For the safety of the port and the urban environment, the setting requirements of the monitoring system are rapidly increasing. In the current digital monitoring system, it is mainly divided into two types: architecture based on computer servo control and single-machine operation architecture. i-Easy, Open and Update, plus network functions and interface extensions usually have basic content analysis capabilities: due to the complexity and diversity of monitoring operations. On the other hand, the latter is ^ BB> {, dedicated circuit or digital signal processing secret matching body and other shock storage operations, with a simple structure, stable performance characteristics, more suitable for long-term, high-performing environment execution Monitor the job. SUMMARY OF THE INVENTION The purpose of the present invention is to provide a multi-hazard detection system and a method for determining the occurrence of a plurality of different hazards in the same detected location. Under the premise of not increasing the cost of labor, the money comprehensively protects the personal and property safety of the detected locations. In order to achieve the above object, according to one aspect of the present invention, a disaster risk measurement and control system is provided for controlling a disaster prevention facility to perform a hazard prevention measure, and the disaster prevention facility includes a first disaster prevention device and a second anti-X Device. The system includes: a detecting device and a control center. The detecting device comprises: an image capturing unit for capturing the detected image of the tested location; 4/28 201140509 an analysis module separately according to the plurality of image analysis logics and analyzing the image according to each image analysis_丝赵1 image analysis 3; subject to: whether the location is hazardous; the image analysis logic includes human body image analysis logic to analyze the quiet image is human body image, and the image analysis information includes human body image information to indicate = body image analysis logic The results of the analysis; 1 Lin Yuan, according to the shadow ^ = poor news, when the hazard situation occurs at the test site, the generation of the human-British-warm message; and a transmission unit output warning message. The control center receives and interprets the warning message ’ and starts the disaster prevention facility according to the warning message. The human body image information indicates that the detected image includes the human body image day, the control center starts the first disaster prevention device, and vice versa, the second disaster prevention device is activated. In order to achieve the above object, according to another aspect of the present invention, another multi-risk (four) measurement and control system is provided for controlling a disaster prevention facility to perform a hazard prevention measure, and the disaster prevention facility includes a first disaster prevention facility. Disaster prevention f set 'The system includes: -_ device and - control center. r wherein the detecting means comprises: capturing the image of the measured location - the image capturing unit, the sensing element of the litchi measuring sensor, and the image analysis module and the variable analysis Module - points. The image analysis module analyzes the logical scene image according to the Wei image, and generates 1 image analysis information according to the analysis result; the t analysis logic includes - human body image analysis logic for analyzing the detection = image or not including - body image And the image analysis f-including - human body image information, used to indicate the analysis result of the human body image analysis logic; the variable analysis module according to at least - analysis and analysis of Wei Wei, the result produces a variable number of cautious 1 U team ·,, ' The analysis of the tool (10) (four) eight analysis modules based on image analysis and the number of knife analysis - hazard analysis information, used to determine whether the test is carried out 5 / 28 201140509 • .. Whether the location occurs - the hazard. The detecting device further includes: -, according to the hazard analysis information, a hazard state day occurs at the tested location. Capture:,: package, person (four) like the message - warning message; and - transmission unit to the control center. The control center receives the solution; when the image includes the human body image, (4) the center starts, and the second, the second disaster prevention device is activated. In order to achieve the above object, according to the present invention, the other side of the invention has a detection device and a control device, and the detection device receives a detection image of a test site; The measuring device analyzes the detected image by a plurality of kinds of image analysis logics, and the image analysis logic includes - human body image analysis logic, whether to include - human body image, and the image analysis information includes - body shadow ^ information: The analysis result used to indicate the human body image analysis logic; the root image analysis logic separately analyzes the obtained result, generates image analysis information, and indicates whether a dangerous state occurs in the tested place, wherein image analysis is greedy Including human body image information; the detecting device determines whether the tested location is in a hazardous state according to the result of the analysis; when the detecting device determines that the detected location is in a dangerous state, a warning message is sent, and the warning message includes the human body image Information; the control center receives and interprets the warning message, and activates a first disaster prevention device or a second disaster prevention device according to the body shadow information to prevent danger Aspects. For a detailed description of the technical means of the present invention, refer to the following embodiments and refer to the accompanying drawings. 6/28 201140509 [Embodiment] The present invention provides a detection and control mechanism capable of simultaneously detecting a sample in order to assist a user in detecting and preventing a hazard that threatens to be safe. 1σ a Figure 4 2 shows a block diagram of a multi-hazard detection and control system of the present invention. The multiple hazard detection and control referred to as _ and control (four) system includes: one side device lQ and a center 20. The detection device 1 can be placed as needed - the two treasures are too cats, the mothers need to receive the danger. Detecting and pre-empting the test site to detect whether a test site is a hazard, such as: flood, fire, illegal person, etc. = 2, set 1 is connected to the control center 20, so as to facilitate _ ^ Hazard: X &, can notify the control center 20 to carry out the necessary countermeasures. The detecting device 10 comprises: an image capturing unit 1 , an analysis module: a warning sheet S104, a transmission unit 106 and a memory unit (10). The second unit (10) can be a photographing device for timing or Continuous operation = detection of double-measured locations (such as warehouses or factories), transmission to points = 1⁄4 group 102 for image analysis and identification, and capture of detected videographers have been transmitted to the money unit (10) . The analysis module 102 reads a plurality of image analysis logics recorded in = hidden early tl 1G8, and analyzes and calculates the image analysis logic of the detected images. Image analysis logic can include: fire image analysis logic, flood image analysis logic, and input logic, such as knife analysis logic, which are used to identify the location under test 疋 no = fire, flood or illegal person invasion kind. The knife analysis module 102 receives images according to different image analysis logics, and the image processing information is generated after the image processing. The analysis of the conclusions of all the analytical logics that have been implemented by the Poverty Alleviation is used to judge 7/28 201140509
圃诼―置囬退仃夕禋演算， )2係利用影像分析的技術，配合 ，可對一張偵測影像中所呈現的 以偵知不同危害態樣；相較於已 知用來_單-態樣的探測I置，更能周延地維護受測地 點的安全。 控制中心20可連結同一區域（例如同一楝大樓）中設 置在不同受測地點的偵測裝置1Q、以及設置在各個侧裝 置10附近的防災設備30。控制中心2〇根據警示訊息中所 P付的訊號及資訊内容’得知該區域巾何處之侧裝置探測 到危害的發生，以便啟動設置在該探測裝置附近的防災設 施30，例如用於撲滅火災的灑水器或滅火噴霧、防堵水災 的防水閘門、或是蜂鳴器及警示燈等設備。此外，控制中 心20也經常連接到一保全中心的監控主機4〇，以便保全人 員也接收到相關的危害發生資訊，以進行必要的處置。 本例中，控制中心20包括一處理模組2〇〇、一設施控 制模組202及一通訊單元204。處理單元200接收偵測裝置 10的傳輸單元丨〇6所傳送的警示訊息，並解讀警示訊息中 所指出的危害態樣的訊號、偵測裝置10的識別資訊等訊息 8/28 201140509圃诼 置 置 置 置 禋 禋 ) ) ) ) ) ) ) ) ) 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 - The detection of the state I can more closely maintain the safety of the tested location. The control center 20 can connect the detecting devices 1Q disposed at different tested locations in the same area (e.g., the same building), and the disaster prevention device 30 disposed near each of the side devices 10. The control center 2 detects the occurrence of the hazard based on the signal and the information content of the P message in the warning message, so as to activate the disaster prevention facility 30 disposed near the detecting device, for example, for extinguishing Fire sprinklers or fire-fighting sprays, waterproof gates to prevent flooding, or buzzers and warning lights. In addition, the control center 20 is often connected to the monitoring host 4 of a security center so that the security personnel can also receive relevant hazard information for necessary disposal. In this example, the control center 20 includes a processing module 2, a facility control module 202, and a communication unit 204. The processing unit 200 receives the warning message transmitted by the transmission unit 丨〇6 of the detecting device 10, and interprets the signal of the hazard state indicated in the warning message, the identification information of the detecting device 10, etc. 8/28 201140509
内容後’產生控制命令，以使設施控制模組2〇2啟動設置 在受測地點附近的防災設施30 ’開始執行終止危害所需的 行動，例如：啟動滅火器喷撒二氧化石炭以撲滅火勢、關閉 防水閘門阻檔水流，或是關閉受測地點的出入口以防止入 侵人員逃脫。處理模組200解讀警示訊息完畢後，還可將 所解讀的危害訊息透過有線或無線的通訊單元2 監控主機40以通知保全人員。 辱 J 在本實施例中’分析模組102還可在控制警示單元ι〇4 ^生警示訊息關時’將記憶單元1Q8記_偵測影像直 接經由傳輸單元廳發送到監控主機。舉例來說：傳輸單 疋106可包括有線或無線網路通訊單元1〇6〇， ==彡像從記憶單元⑽中讀取出來，讀位訊號的型 態直接透過網路通訊單元麵傳送到監控主機40 :監控主機㈣保全人以視覺辨觸方式，從偵測影像 直接说測到叉測地點所發生的危害情狀，以利於相關人 j縱危《樣的發展或_危_嚴重程 取 有效的應對措施。 ％扯取 原理在=]中：爾分析邏輯分析各種危害態樣的 原里係_在-個㈣地點發生火災時， 使得光線的穿透率降低；或發生水 的地面上’水流的波紋造成光線的折射角=3水 _’在軸像的畫面中會反映出：= 是否有符合上述職^f料騎，計❾貞測影像中 遮蔽裎产特徵，以及該等特徵的分佈情形、 ^度寻’《判斷影像中是否有符合發生特定危害的情 9/28 201140509 況。 此種手段所達到的偵測效果，用於鮮少有人監控或看 管的地下室、車庫或倉庫等地點時，可有效地 成的財物損失和災害。 心 當然’有關火警和人員入侵的偵測，也有許多的濟算 倾可進行分析’如上述根據影像中的光線亮度、煙霧遮 蔽物件的程度等條件，來判斷火警的發生與否。 人員的人侵則已經有更多的演算邏輯可供分析模組 1〇2進行分析和辨識，例如根據人體影像的頭、身、四肢的 比例’甚至是偵測影像中擷取到的人類五官的分佈比例（ 辨》戠）等荨’皆可用來判斷受測地點中的人體影像的 身刀’甚至了比對預先^己錄在記憶單元1〇8中的資料，以 判斷人體f彡像是否為可被允許出現在舰㈣人員。例如 適用，家庭巾時’用以分辨被偵測到的人員是否為該家庭 的成員；或是適用在辦公處所時’姻被侧到的人員是 否在其被允許出現的時段内出現（如公司員工僅於上班時 間出現）。 此外，除了可從偵測影像中判斷出受測地點所發生的 危害態樣外，更可利用影像分析技術精確地判斷受測地點 、，么生危的區域，並將區域資訊也包含在警示訊息中傳 U到技制中心20,以利控制中心20針對受測地點進行區域 性的防災措施。例如發生火災時，根據影像分析可獲知在 該地點中僅有局部區域發生火災時，控制中心20可單獨針 對杳生火災的區域灑水，防止其他未受火焰波及的部分也 被淋澄。 由於上述各種偵測及辨識的影像分析邏輯係為該技術 10/28 201140509 =或中之技術人員所可知悉之手段，因此即不再對詳細的 :::式進行贅述。然而’必須特別說明的是，上述所說 $種*像分析邏輯所分析的危害態樣、對每—種危宝 =樣係根_㈣徵進行分m所適㈣分析演^ 僅為貫施方式的例示，本實施例中分析模、组1〇2所 進行的分析作業，並不限於上述範例；設計者自然可視偵 測的需求·更多的影像簡邏輯和演算方法，以使影像 偵測的技術更臻至完美。After the content, a control command is generated to cause the facility control module 2〇2 to start the disaster prevention facility 30' disposed near the tested location to start the action required to terminate the hazard, for example, starting the fire extinguisher to spray the carbon dioxide to extinguish the fire, Close the waterproof gate to block the flow of water, or close the entrance and exit of the site under test to prevent intruders from escaping. After the processing module 200 interprets the warning message, the interpreted hazard message can also be monitored by the wired or wireless communication unit 2 to notify the security personnel. In this embodiment, the analysis module 102 can also send the memory unit 1Q8_detection image directly to the monitoring host via the transmission unit hall when the control alert unit ι〇4^the warning message is turned off. For example, the transmission unit 106 may include a wired or wireless network communication unit 1〇6〇, == the image is read from the memory unit (10), and the type of the read signal is directly transmitted to the network communication unit surface. Monitoring host 40: monitoring host (4) The security person directly detects the hazard situation of the fork test location from the detected image in order to facilitate the related person's development or _ crisis _ serious process Effective response measures. The principle of % pulling is in =]: the analysis logic analyzes the original system of various hazard patterns _ when the fire occurs in a place (four), the penetration rate of light is reduced; or the ripple of the water flow on the ground where water occurs The refraction angle of the light = 3 water _' will reflect in the picture of the axis image: = Is there any fit in the above-mentioned occupation, the characteristics of the shadowing in the image, and the distribution of these features, ^ "Finding" "Determining whether there is a situation in the image that meets the specific hazard 9/28 201140509. The detection effect achieved by such means is effective for loss of property and disasters when used in places such as basements, garages or warehouses where few people are monitored or supervised. Of course, there are also many economic calculations for the detection of fire and personnel intrusion. As mentioned above, the occurrence of fire alarms is judged based on the brightness of the light in the image and the extent to which the smoke obscures the objects. There are more calculation logics for personnel intrusion for analysis and identification. For example, according to the ratio of the head, body and limbs of the human body image, even the human facial features captured in the image are detected. The distribution ratio (recognition) and so on can be used to judge the body image of the human body in the test site', and even compare the data recorded in the memory unit 1〇8 to determine the human body image. Whether it is allowed to appear in the ship (four) personnel. For example, when a family towel is used, it is used to distinguish whether the detected person is a member of the family; or when it is applied to the office, whether the person who is married is present during the time period in which it is allowed to appear (such as a company) Employees only appear during working hours). In addition, in addition to judging the hazard pattern of the tested location from the detected image, it is also possible to use image analysis technology to accurately determine the location to be tested, the area that is at risk, and include the regional information in the warning. The message is transmitted to the technical center 20 to facilitate the control center 20 to carry out regional disaster prevention measures for the tested location. For example, in the event of a fire, according to the image analysis, it is known that when there is only a local area fire in the location, the control center 20 can separately sprinkle the area where the fire is generated, and prevent other parts that are not affected by the flame from being drenched. Since the image analysis logic of the above various detection and identification is a means that can be known by the technician in the technology, the detailed ::: is no longer described. However, it must be specially stated that the above-mentioned $ species* is like the analysis of the hazard pattern analyzed by the analysis logic, and the classification of each type of dangerous treasure = sample root _ (four) is appropriate. The exemplification of the method, the analysis operation performed by the analysis module and the group 1〇2 in the embodiment is not limited to the above example; the designer naturally needs the visual detection, more image simple logic and calculation method to make the image detection The technology of measurement is even more perfect.
考里到叉測地點的性質、架設防災設施的成本等因素 ，每個受測地點所設置的各種防災設施及其啟用的優先順 序可能都不同。例如：若受測地點為少有人員出入的倉庫 :^了防止火災的發明’同時在該地點設置了撒水裝置及 -氧化兔灌注裝置。當彳貞測到所述受測地點發生火災時， 為了避，貨品被水淋澄而損壞，控制中心、2G可能設定優先 啟，二氧化碳灌注I置’讓火源缺乏助燃物而熄滅。但若 ^災發生時’受測地點正好有人員在内，使用此種防災設 也雖可、、止危害’但部造成人員的生命危險。控制中心烈 ί此種狀況下，則應適用對人體無害的防災設備來執行防 火的措施。 因此，在另-實施例中，複數影像分析邏輯當 匕括-人體影像分析賴，用时析彳貞測影像巾是否 ^體影像，藉以判斷偵測影像被擷取的當時，是否人 ί於受測地點之内。同時，影像分析資訊中也可更進一ΐ “-人體景彡„訊，其料 Q ς :析邏輯後所產生的資訊，用以指示出受測地二 肢影像。當分析模組⑽根據影像分析資訊_出受測= 11/28 201140509 點發生至少―種危錢樣而使警示單元 時，警示訊息中亦包括所述的人體费傻毐匕 心誠據此資訊為不_ = ^像贫訊，以供控制中 受?制k 2。控_ m設施3。 300及弟—防災裝置3〇2。其中第一 二執 對人體無害的防災作業ΪThe nature of the test site to the location of the fork test, the cost of setting up the disaster prevention facility, etc., the various disaster prevention facilities set at each site tested and the priority order of their activation may be different. For example, if the site to be tested is a warehouse with few people entering and exiting: ^Invented to prevent fires' At the same time, a sprinkler device and an oxidized rabbit infusion device were installed at the site. When it is detected that there is a fire in the tested location, in order to avoid the damage caused by the water being drenched, the control center, 2G may set priority, and the carbon dioxide perfusion I shall cause the fire source to be extinguished without the combustion aid. However, if the disaster occurred, the location of the test site was just the presence of personnel, and the use of such disaster prevention facilities could be used to stop the hazard. In the case of such a situation, the fire prevention measures that are harmless to the human body should be applied to perform fire prevention measures. Therefore, in another embodiment, the plurality of image analysis logics are used to analyze whether the image is captured or not, and whether the image is captured at the time of the detection, whether or not the image is captured. Within the test location. At the same time, the image analysis information can also be further enhanced by the "-human body scene" message, which is Q ς: the information generated after the logic is used to indicate the image of the measured limbs. When the analysis module (10) causes the warning unit to generate at least one kind of dangerous money according to the image analysis information_11/28 201140509, the warning message also includes the human body fee, which is based on the information. For the sake of not _ = ^ like a poor news, for control in the process of making k 2 . Control _ m facility 3. 300 and brother - disaster prevention device 3〇2. The first two of them are harmless to the human body.
' ^ ^ ^ 0 ^ f ；Ff' I =行不易損壞受測地點財物_災_ H 石反灌注裝置或氧氣柚取裝置等。 、 羊 當控制中心20接收到包含有人體 訊號細裝置丨。識別資訊的警示訊4 :先:斷3 ”偵測影像中有無人體影像，若二 時有人貝位在發生危害的受測地點，此時處理模組二二 發出控制命令，使設施控制模組2 動 ' 3〇中的第-防災裝置，以線止产啟動防災設施 ，、、止危告態樣的持續並保全 人身女王，反之，从體影像資訊指出並無人體 ，設施控制模組逝則可根據控制命令啟動第二防災裝= 3〇2 夬速終止危害態樣，控制財物損失到最小程度。 第二圖展示了本發明提供的—種多重危害仙及控制 系統第二實施例的方塊圖。第二實施例中，_及控制系 統la的偵測裝置10a更包括一感測單元11〇 ;而分析模組 l〇2a則包括一影像分析模組1〇2〇及_變數分析模組1⑽2 〇 、感測單元110可為煙霧感測單元、光線感測單元、溫 度感測單元或紅外線感測單元等，分別用於感測受測地點 的煙霧濃度或是溫度高低變化，而產生包括光線穿透率、 12/28 201140509 溫度等環境變數。 影像分析模組1020係用於執行多種影像分析邏輯，以 分析影像操取單元議所娜到的偵測影像及產生影像分 析資訊，細節請參照第-實施例所迷。變數分賴組1〇22 則接收感測單元110所感測到的-到多個環境變數’並從 • 雜早元⑽讀取相對應的變數分析邏輯來分析環境變數 * =斷受測地點是否發生感測單元]]G所用以_的危 :二羡。’交數分析邏輯可包括溫度變數分析邏輯、溼度變 • 數分析邏輯、煙霧濃度變數分析邏輯等。 舉例來說，變數分析模組體可執行該等變數分析邏 Μ利用—預設值和接收到的環境變數相比，根據環境 ’支數與預設值的大小作為判斷是否發生危害態樣的基準。 -種感測單元11G通㈣以_—種環數，在實 作上可設置多個不同的感測單元，以供變數分析模組】〇22 分別以不同的變數分析邏輯分析不同感測單元所感測到的 環境變數，以供偵測不同態樣的危害。 φ 變數分析极組1022利用變數分析邏輯將相對應的環境 ’交數分析7G成後，亦會產生一變數分析資訊，用以指示出 • 變數分析模組1022的分析結果。 對於影像分析模組1020與變數分析模組1〇22分別產 生的影像分析資机與變數分析資訊，分析模組1〇2&可採取 不同的作法。 其中一個貫她方式中，影像分析模組丨〇2〇與變數分析 , 模組1〇22为別分析相同的危害態樣，分析模組l〇2a更比 對影像分析資減魏分析資訊，以確認利用影像分析的 結果和感測單兀110感測的結果，是否都得到相同的結論 13/28 201140509 .， « • t t ’藉此以避免產生誤判的情形。 。例如：感測單元11Q係為用於偵測煙霧濃度的煙霧感 齡元’魏分龍組順執行濃度變數分析祕後，在 變數/刀析資訊中包含受測地點是否發生火災的結論；而影 像分析模組嶋也在執行火災影像分析邏輯後，於影像分 析寊訊中記錄受測地點是否發生火災的結論。分析模組 職將影像和變數分析資訊中對於是否發生火災的結論進 行比對，再產生最後的危害分析資訊。 田一個刀析資訊皆指示出受測地點發生火炎時，分析 模組购才觸受測地點確有火災發生；#二個分析資訊 皆扎出文測地點未發生火災時，分析模組】〇2汪則判斷受測 地點並無火災發生。 /刀析模組102a比對影像分析資訊和變數分析資訊後， 根據比對的結果產生危害分析資訊，警示單元]〇4則根據 1害分析資騎指出的分析結果，於相有危錢樣產生 日寸’發迗警不訊息到控制中心2〇。控制中心2〇的處理模組 200接—收及解讀警示訊息中的危害態樣訊號、镇測襄置他 識別資料及人體影像資訊後，再控制設施控讎組啟 動適當的防災設施30，以終止危㈣樣的持續。 田刀析模組102a比對影像分析資訊和變數分析資訊， 發現對於同-種危害態樣的判斷結論*同時，分析模組 102a可將偵卿驗由網路通訊單元删輸㈣監控主機 或遠端電子裝置42 (如行動電話、筆記型電腦），以供接收 到偵測影像的人員目視判斷。 而在另一個實施方式當中，影像分析模組1020及變數 分析模組1022分別用以分析不同的危害態樣。分析模組 14/28 201140509 l〇2a除了接收偵測影像進行影像分析，以 態樣之外，同時也接收感測單元nQ所===害 再^別根據相對應的影像及變數分析= 的务生。利用影像分析模組]〇2〇和 …、° 別判斷不㈣轉紐，減料賴;^ ‘ ^ 102a^^, 綱密度高的感測單元„〇獲取準確的環境：= 於對危害發生與否的判斷。 文數有助 第，實施射所未詳述的其他元件其特徵 一貫施例之說明。 月 > …乐 第三圖顯示-種多重危害偵測及控制方者 =程圖。請同時參照第-圖所示之方塊圖。本 7裝置1G以影像擷取單㈣或連續操取 、 = = 分析模組1〇2讀取記憶單^⑽ 肀的衫像分析邈輯分析偵測影像。 其中，分析模組1〇2執行一人體影像分析邏輯 軸象’用以判斷是否有人體影像存在偵測影像當中⑽3 以及執行其他影像分㈣輯，以觸受測地點是否發生 重態樣的危害（S505)，例如火災影像分析邏輯、水 二=象分析，或人侵影像分析邏輯。複數種影像分析邏 白執仃完畢後，分析模組1Q2產生—.影像分析資訊（咖 i傻-Γ包括㈣指利貞·像妓轉有人體影像的人體 二像貝訊，以及執行其他影像分析邏輯之後的結果，用以 才曰不出受測地點有無危害態樣發生。 分析模組逝依據影分析資訊判斷受測地點是 刀析的危害態樣發生⑽9):若影像分析資訊指出受測地 15/28 201140509 點並任何受分析的危害態樣產生，即可返回步驟伽以 ’擷取=張_影像⑽續分析；若影像 艾分析出受測地點有危害態樣產生，= 早7C丨04則產生一婆不句自，爲、"2；丨± s 丁汛心，傳达到控制中心20 (S5】n ^示^包括危害的態樣、人體影料訊，Μ 1 置以供控制中心2。進行適置裝 控制中心20接收到馨示自砗 1 福前要m ='存（3 )，並根據危害的態樣與 = 訊’判斷應適用的防災設施％類別 2點°虽人體影像訊息指出受測地點無人員存在時，控 中^〇控制第二防災裝置302執行終止危害的作業（ S5! 5 )，反之’若當時有人員存在受_ 防災裝置期執行終止危害的作#(s = 災裝置將執行對人體無害的防災措施。 第四關顯衫f危害勤】及控制方法第二實施例的 ^壬圖、’ μ併參閱第二圖所示之系統方塊圖 。本實施例 :由偵測裝置l〇a同時接收受測地點的侧影像及感測單 兀m感測到的環境變數（_);偵測裝置伽的影像分 斤柄組1_根據偵測影像執行人體影像分析邏輯，產生一 =影像資i 影像U否包括人體影像（ B外退分取其他的影像分觸輯分析偵測影像， S60O文數刀析模組1022執行至少一種變數分析邏輯（ ’以分職得影像分析f訊及變數分析資訊（S607) 中，影像分析資訊包括步驟S603所產生的人體影像資 刀析板組lG2a更進—步比對影像分析資訊及變數分析 16/28 201140509 二者對同種危害態樣是否有相同的' ^ ^ ^ 0 ^ f ; Ff' I = line does not easily damage the property at the site under test _ disaster _ H stone reverse osmosis device or oxygen pomelo device. , Sheep When the control center 20 receives a device containing a human body signal. Warning message for identifying information 4: First: Break 3 "Detects whether there is any human body image in the image. If there is a shell in the damaged location, the processing module sends a control command to make the facility control module. 2 The first - disaster prevention device in the '3 〇 , 启动 启动 启动 启动 启动 启动 启动 启动 启动 启动 启动 启动 启动 启动 启动 启动 启动 启动 启动 启动 启动 启动 启动 启动 启动 启动 启动 启动 启动 启动 启动 启动 启动 启动 启动 启动 启动 启动 启动 启动 启动 启动 启动 启动Then, the second disaster prevention device can be started according to the control command = 3〇2, the hazard state is terminated at an idle speed, and the property loss is controlled to a minimum. The second figure shows the second embodiment of the multi-hazard and control system provided by the present invention. In the second embodiment, the detecting device 10a of the control system la further includes a sensing unit 11A; and the analyzing module 102a includes an image analyzing module 1〇2〇 and _variable analysis The module 1 (10) 2 〇, the sensing unit 110 can be a smoke sensing unit, a light sensing unit, a temperature sensing unit or an infrared sensing unit, etc., respectively, for sensing the smoke concentration or the temperature change of the tested location, and Produce including light Line penetration rate, environment variable such as 12/28 201140509 temperature. Image analysis module 1020 is used to perform a variety of image analysis logic to analyze the image detected by the image manipulation unit and generate image analysis information. Referring to the first embodiment, the variable group 1 22 receives the -to-multiple environment variables sensed by the sensing unit 110 and reads the corresponding variable analysis logic from the impurity element (10) to analyze the environment. Variable * = Whether the sensing unit is detected at the location under test]] The risk used by G is: 羡. The intersection analysis logic may include temperature variable analysis logic, humidity variable number analysis logic, smoke concentration variable analysis logic, and so on. For example, the variable analysis module body can perform the variable analysis logic utilization - the preset value is compared with the received environment variable, and the size of the environment and the preset value are used as a judgment to determine whether a hazardous situation occurs. Benchmark - The sensing unit 11G is connected (4) with _-the number of rings, and in practice, a plurality of different sensing units can be set for the variable analysis module. 〇22 respectively analyzes the logic with different variables. The environmental variables sensed by the sensing unit are used to detect the hazards of different aspects. φ The variable analysis pole group 1022 uses the variable analysis logic to analyze the corresponding environment 'the intersection number of 7G, and also generates a variable analysis information. For indicating the analysis result of the variable analysis module 1022. For the image analysis machine and the variable analysis information generated by the image analysis module 1020 and the variable analysis module 1〇22, the analysis module 1〇2& Different methods. Among them, the image analysis module 丨〇2〇 and the variable analysis, the module 1〇22 analyzes the same hazard, and the analysis module l〇2a compares the image analysis. Wei analyzes the information to confirm whether the results of the image analysis and the results of the sensing unit 110 are used to obtain the same conclusion. 13/28 201140509 ., « • tt ' to avoid misjudgment. . For example, the sensing unit 11Q is a smoke sensory element for detecting the concentration of smoke. Weilonglong Group performs the analysis of the concentration variable, and the variable/knife analysis information includes the conclusion that the fire is detected in the tested location; The image analysis module 嶋 also performs the fire image analysis logic and records the conclusion of whether the fire occurred at the tested location in the image analysis. The analysis module compares the conclusions of the image and the variable analysis information about whether a fire has occurred, and then generates the final hazard analysis information. When a piece of knife analysis information indicates that there is a fire in the site under test, the analysis module purchases the site to be tested, and there is a fire; #二分析信息的出出出处处处有检测的消防的分析模块〇 2 Wang judged that no fire occurred at the site under test. After the tool analysis module 102a compares the image analysis information and the variable analysis information, the damage analysis information is generated according to the comparison result, and the warning unit 〇4 is based on the analysis result indicated by the trait analysis, and the dangerous sample is in the phase. Produce a day's squad's message to the control center 2〇. The processing module 200 of the control center 2 receives and interprets the hazard state signal in the warning message, and then controls the facility control group to activate the appropriate disaster prevention facility 30 to Termination of the end of the crisis (four). The field knife analysis module 102a compares the image analysis information and the variable analysis information, and finds the judgment conclusion for the same kind of harm state. At the same time, the analysis module 102a can delete the detection module from the network communication unit (4) the monitoring host or The remote electronic device 42 (such as a mobile phone, a notebook computer) is visually judged by a person who receives the detected image. In another embodiment, the image analysis module 1020 and the variable analysis module 1022 are respectively configured to analyze different hazard patterns. Analysis module 14/28 201140509 l〇2a In addition to receiving the detected image for image analysis, in addition to the aspect, it also receives the sensing unit nQ === harm and then according to the corresponding image and variable analysis = Health. Use image analysis module] 〇2〇 and..., ° Do not judge (4) turn, reduce material lag; ^ ' ^ 102a^^, high density of sensing unit „〇Get accurate environment:= Judgment of the or not. The number of texts helps the implementation of the other elements of the unspecified details of the shooting. The characteristics of the consistent application. Month > ... Le the third picture shows - a variety of multi-hazard detection and control = Cheng map Please also refer to the block diagram shown in the figure - Figure 7: 1G image capture unit (4) or continuous operation, = = analysis module 1〇2 read memory sheet ^ (10) 衫 shirt image analysis analysis The image is detected by the analysis module 1〇2, and a human body image analysis logic axis image is used to determine whether there is a human body image in the detected image (10) 3 and other image segmentation (4) is performed to determine whether the touched site is in a heavy state. Such hazards (S505), such as fire image analysis logic, water two = image analysis, or human intrusion image analysis logic. After a plurality of image analysis logics are completed, the analysis module 1Q2 generates - image analysis information Silly - Γ includes (four) refers to Li Wei The two images of the human body, such as Beixun, and the results of other image analysis logics, can not be used to detect the presence or absence of harm in the tested location. The analysis of the module is based on the analysis of the image to determine the location of the damage. (10)9): If the image analysis information indicates that the 15/28 201140509 point is detected and any of the analyzed hazard patterns are generated, you can return to the step 伽 撷 撷 张 张 张 影像 影像 影像 影像 影像 影像 ; ; ; ; ; ; ; ; ; ; ; ; ; The location of the test has a harmful situation, = 7C丨04 is generated by a woman, for, "2; 丨± s Ding Xin, conveyed to the control center 20 (S5) n ^ show ^ including the state of harm Sample, human body shadow information, Μ 1 set for the control center 2. The appropriate control center 20 receives the sinus from the 砗 1 before the blessing m = ' deposit (3), and according to the hazard and = 'Determination of the disaster prevention facility % category to be applied 2 points. Although the human body image message indicates that no person exists in the test site, the control unit 2 controls the second disaster prevention device 302 to perform the operation to terminate the hazard (S5! 5), otherwise Some people existed due to _ disaster prevention device period execution termination risk The work #(s = the disaster device will perform the disaster prevention measures that are harmless to the human body. The fourth level shows the hazard] and the control method of the second embodiment, 'μ and refer to the system block shown in the second figure In the embodiment, the detection device l〇a simultaneously receives the side image of the tested location and the environmental variable (_) sensed by the sensing unit ;m; the image of the detecting device gamma is divided into groups 1_ Detecting images and performing human body image analysis logic to generate a = image i image U or not including human body images (B external retreat to take other image sub-touch analysis to detect images, S60O text analysis module 1022 to perform at least one variable The analysis logic (in the image analysis and the variable analysis information (S607), the image analysis information includes the human image processing tool group lG2a generated in step S603, and the image analysis information and the variable analysis are further advanced. 16/28 201140509 Does the two have the same kind of hazard?
一者判斷為有危害發生、另一 資訊的内容，用以判斷二 分析結論（S609)，包括： 於同一種危害態樣都判斷 判斷出發生火災；若否；， ’即對於 者則判斷 為無危害發生時’分析模組咖則透過網路通訊單元川6〇 ’將偵測縣發翻遠端電子裝置Μ，以供接 識（S615)。 ，捕 右在步驟S611 +，分析模組臟判斷出二個分析資 訊都分析出發生特定危害態樣，警示單元1Q4則產生包括 人體影像資關警示訊息，傳送到控制中心2G⑽7)。控 制中心20先根據人體影像資訊判斷受測地點是否有人員在 内（S619)，若無，則啟動設置在受測地點附近、可終止危 害態樣的第二防災裝置302以終止危害（S621);但若受測 地點有人員在内，控制中心2〇則啟動設置在受測地點附近 、可終止危害態樣且對人體無害的第一防災裝置3〇〇以終 止危害（S623)。 第五圖再提供多重危害偵測及控制方法第三實施例的 流程圖，亦請同時參照第二圖。第三實施例亦由偵測裝置 10a擷取偵測影像及接收環境變數（S7〇i)，旅由影像分析 模組1020以人體影像分析邏輯判斷受測地點有無人員存在 (S703) ° 17/28 201140509 影像分析模組102〇與變數分析模組1022分別根據倘 測影像與環境變數，分析不同態樣的危害是否在受測地點 發生（S705 )，例如影像分析模組]020利用水災影像分析 邏輯與入侵影像分析邏輯分析分別判斷受測地點有無水災 或人員入侵’另外由變數分析模組】022以煙霧濃度變數分 析邏輯判斷受測地點有無火災；再分別產生影像分析資訊 及變數分析資訊（S707)。 分析模組102a根據影像及變數分析資訊的内容，判斷 受測地點是否有任何一種危害態樣發生（S7〇9):若無則返 回步驟S70]以下繼續執行偵測；若確實有危害態樣產生， 則發送包括人體影像資訊的警示訊息到控制中心( π]】 )’控制中心20先識別受測地點有無人員在内（S713)，若 無，則啟動第二防災裝置302 (S715);若有人員在内則啟 動第一防災裝置300 (S717)，以終止危害的發生。 本把例中與苐一貫施例相同之處即不再重述，請參 照第二實施例之說明。 / 綜合上述，已揭露出本發明所提供的多重危害偵測及 ，制系統和方法所使用的手段，利用影像分析的技術，同 蚪可對一地點進行多重的危害偵測及警示，並且可彈性地 调整俄測的方式、順序及對應㈣結果而提供不 裝置終止危害。 〜同時，還可與現行的各種探測系統整合，不但無需另 探測系統的硬體規格，更可配合現有的感測裝置做 重的制絲確認’提高制結果醉確度、降低誤判 ’此外，利用此種多重危害偵測的手段，更可長時間自 監控少有人煙的地點，減少以人力監控所需花費喊本 18/28 201140509 ^而，上述各實施例當中之項目及元件，僅係為闡述 本發明所舉之例示’並無自限所請求保護之範圍之意圖。 凡遵循本發明之精神及根據本發明所揭示之技術手段，而 進行微幅之修飾或改變者，亦屬本發明所保護之範疇。 【圖式簡單說明】 ^第一圖：本發明所提供的一種多重危害偵測及控制系 統第一實施例之方塊圖； —第二圖：本發明所提供的一種多重危害偵測及控制系 統第二實施例之方塊圖； 第二圖.本發明所提供的一種多重危害偵測及,控制方 法第一實施例之流程圖； 、第四圖：本發明所提供的一種多重危害偵測及控制方 法第二實施例之流程圖；及 第五圖：本發明所提供的一種多重危害侧及控制方 法第三實施例之流程圖。 【主要元件符號說明】 1，la多重危害偵測及控制系統 10, 10a偵測裝置 1〇〇影像擷取單元 102, 102a分析模組 1020影像分析模組 1022變數分析模組 104警示單元 19/28 201140509 106傳輸單元 1060網路通訊單元 108記憶單元 110感測單元 20控制中心 .200處理模組 202設施控制模組 204通訊單元 30防災設施 300第一防災裝置 302第二防災裝置 40監控主機 42遠端電子裝置 S501-S517流程步驟 S601-S623流程步驟 S701-S717流程步驟 20/28One judges that there is a hazard, and the content of another information is used to judge the second analysis conclusion (S609), including: judging that a fire has occurred in the same hazard; if not; When no harm occurs, the 'analysis module coffee will send the remote electronic device to the county through the network communication unit Chuan 6〇' for the access (S615). In the step S611+, the analysis module dirtyly determines that the two analysis information analyzes the specific hazard situation, and the warning unit 1Q4 generates a warning message including the human body image and transmits it to the control center 2G(10)7). The control center 20 first determines whether there is a person at the tested location according to the human body image information (S619), and if not, starts the second disaster prevention device 302 that is disposed near the measured location and can terminate the hazardous aspect to terminate the hazard (S621) However, if there is a person at the test site, the control center 2 starts the first disaster prevention device that is placed near the test site and can end the hazard and is harmless to the human body to terminate the hazard (S623). The fifth figure further provides a flowchart of the third embodiment of the multiple hazard detection and control method, and also refers to the second figure. The third embodiment also captures the detected image and the receiving environment variable (S7〇i) by the detecting device 10a, and the traveling image analyzing module 1020 determines the presence or absence of the person at the tested location by the human body image analysis logic (S703). 28 201140509 The image analysis module 102〇 and the variable analysis module 1022 respectively analyze whether the hazards of different aspects occur at the measured location according to the measured image and the environmental variables (S705), for example, the image analysis module 020 uses the flood image analysis. Logic and intrusion image analysis logic analysis to determine whether there is no water disaster or personnel invasion in the tested location. In addition, the variable analysis analysis module 022 uses the smoke concentration variable analysis logic to determine whether there is fire in the tested location; and then separately generates image analysis information and variable analysis information ( S707). The analysis module 102a determines whether any of the hazard scenes are generated according to the content of the image and the variable analysis information (S7〇9): if not, return to step S70] to continue the detection; if there is a harmful situation If yes, send a warning message including human body image information to the control center (π]]) 'the control center 20 first identifies the presence or absence of the person at the test site (S713), and if not, activates the second disaster prevention device 302 (S715); If there is a person, activate the first disaster prevention device 300 (S717) to terminate the occurrence of the hazard. In the example, the same as the conventional embodiment is not repeated, please refer to the description of the second embodiment. / In summary, the methods used in the multi-hazard detection and system and method provided by the present invention have been disclosed, and the techniques of image analysis are utilized, and multiple hazard detection and warning can be performed on a location, and Elastically adjust the method, sequence and corresponding (4) results of the Russian test to provide no device termination hazard. ~ At the same time, it can be integrated with the current various detection systems, not only does not need to detect the hardware specifications of the system, but also can be used with the existing sensing devices to make heavy wire confirmations to improve the results of drunkenness and reduce false positives. This multi-hazard detection method can also monitor the location of people with less smoke for a long time, and reduce the cost of manpower monitoring. The items and components in the above embodiments are only The exemplification of the invention is not intended to limit the scope of the claimed invention. It is also within the scope of the present invention to modify or modify the micro-frames in accordance with the spirit of the present invention and the technical means disclosed herein. BRIEF DESCRIPTION OF THE DRAWINGS FIG. 1 is a block diagram of a first embodiment of a multiple hazard detection and control system provided by the present invention; FIG. 2 is a multi-hazard detection and control system provided by the present invention. A block diagram of a second embodiment; a second diagram of a first embodiment of a multi-hazard detection and control method provided by the present invention; and a fourth diagram: a multi-hazard detection provided by the present invention A flow chart of a second embodiment of the control method; and a fifth diagram: a flow chart of a third embodiment of the multi-hazard side and control method provided by the present invention. [Main component symbol description] 1, la multi-hazard detection and control system 10, 10a detection device 1 image capturing unit 102, 102a analysis module 1020 image analysis module 1022 variable analysis module 104 warning unit 19/ 28 201140509 106 transmission unit 1060 network communication unit 108 memory unit 110 sensing unit 20 control center 200 processing module 202 facility control module 204 communication unit 30 disaster prevention facility 300 first disaster prevention device 302 second disaster prevention device 40 monitoring host 42 Remote electronic device S501-S517 process steps S601-S623 process steps S701-S717 process step 20/28