技术领域technical field
本公开涉及一种移动物侦测方法；具体而言，本公开是用于驾驶辅助系统的一种移动物侦测方法。The disclosure relates to a method for detecting a moving object; in particular, the disclosure is a method for detecting a moving object used in a driving assistance system.
背景技术Background technique
驾驶安全性一直是运输工业的重要课题之一。随着电子技术的进展，有许多主动式安全驾驶方案导入运输工业中。例如，车道偏离警示、盲点侦测、前方物体侦测、驾驶状态监测等等。以前方物体侦测方案为例，其接收并侦测驾驶人前方的场景数据，当侦测到有物体接近时会向驾驶人发出警示。Driving safety has always been one of the important topics in the transportation industry. With the advancement of electronic technology, many active safety driving solutions are introduced into the transportation industry. For example, lane departure warning, blind spot detection, front object detection, driving status monitoring, etc. Taking the front object detection solution as an example, it receives and detects the scene data in front of the driver, and sends a warning to the driver when an object is detected approaching.
然而，现有前方物体侦测方案的侦测方式将接收影像中所有位置及所有大小的区域做比对，此一作法将耗费较多的运算资源，造成运算效能不佳。因此，现有前方物体侦测方案的侦测方式仍有待改进。However, the detection method of the existing front object detection scheme compares all positions and areas of all sizes in the received image. This method consumes more computing resources and results in poor computing performance. Therefore, the detection method of the existing front object detection scheme still needs to be improved.
发明内容Contents of the invention
本公开的一目的在于提供一种移动物侦测方法，可根据侦测结果调整侦测范围。An object of the present disclosure is to provide a method for detecting a moving object, which can adjust the detection range according to the detection result.
本公开的一目的在于提供一种移动物侦测方法，可提高侦侧效率。An object of the present disclosure is to provide a mobile object detection method, which can improve detection efficiency.
移动物侦测方法用于驾驶辅助系统，驾驶辅助系统包括摄影机，用以持续取得侦测范围的多个视格，多个视格至少包括第一视格及第二视格，移动物侦测方法包括下列步骤：根据侦测范围取得第一视格，并于第一视格执行全区域侦测；根据第一视格的全区域侦测结果定义侦测范围包含侦测区及非侦测区；根据侦测范围取得第二视格，于第二视格的侦测区执行部分区域侦测。The moving object detection method is used in the driving assistance system. The driving assistance system includes a camera to continuously obtain multiple frames of the detection range. The multiple frames include at least the first frame and the second frame. The moving object detection The method includes the following steps: obtaining the first visual frame according to the detection range, and performing full area detection in the first visual frame; defining the detection range including the detection area and the non-detection area according to the full area detection result of the first visual frame area; get the second visual frame according to the detection range, and perform partial area detection in the detection area of the second visual frame.
附图说明Description of drawings
图1为本公开移动物侦测方法的一实施例流程图；FIG. 1 is a flow chart of an embodiment of the mobile object detection method of the present disclosure;
图2为本公开驾驶辅助系统的一实施例方框图；FIG. 2 is a block diagram of an embodiment of the disclosed driving assistance system;
图3为本公开移动物侦测方法执行不同部分区域侦测的实施例流程图；FIG. 3 is a flow chart of an embodiment of performing detection of different parts of the mobile object detection method of the present disclosure;
图4为未侦测到移动物时执行部分区域侦测的上视图；Fig. 4 is a top view of performing partial area detection when no moving object is detected;
图5A及图5B为侦测到移动物时执行部分区域侦测的上视图；FIG. 5A and FIG. 5B are top views of performing partial area detection when a moving object is detected;
图6为边缘区域组成的示意图；Fig. 6 is a schematic diagram of the composition of the edge region;
图7为本公开移动物侦测方法具有侦测周期设定的实施例流程图；FIG. 7 is a flow chart of an embodiment of the disclosed mobile object detection method with detection period setting;
图8为本公开移动物侦测方法判断全区域侦测过程的实施例流程图；FIG. 8 is a flow chart of an embodiment of the process of judging the whole area detection process of the mobile object detection method of the present disclosure;
图9A为采用移动物后方为阻隔区的上视图；Figure 9A is a top view using the rear of the moving object as the blocking area;
图9B至9D为说明图9A设定方式的示意图。9B to 9D are schematic diagrams illustrating the setting method of FIG. 9A.
具体实施方式detailed description
本公开提供一种移动物侦测方法，可在设定的侦测范围内根据对移动物的侦测结果来调整后续侦测方式，以提高系统运算效能。移动物例如为行人、车辆、野生物等。本公开的移动物侦测方法可用于车辆的驾驶辅助系统。The present disclosure provides a method for detecting a moving object, which can adjust the subsequent detection method according to the detection result of the moving object within a set detection range, so as to improve the computing performance of the system. Mobile objects are, for example, pedestrians, vehicles, and wild animals. The moving object detection method of the present disclosure can be used in a driving assistance system of a vehicle.
整体而言，本公开的移动物侦测方法包含两个阶段：(1)全区域侦测；以及(2)部分区域侦测。请参考图1。图1为本公开移动物侦测方法的一实施例流程图。如图1所示，在步骤S100：执行全区域侦测。例如，以拍摄前方30米，张开水平视角100度所围成的范围作为侦测范围。摄影机取得侦测范围的多个视格，这些视格的内容涵盖侦测范围。上述视格可组成影像串流，例如每秒二十张视格(20f/s)。多个视格可包含至少第一视格及第二视格。根据侦测范围取得第一视格，并在第一视格执行全区域侦测。例如，将侦测范围分成多个区域，以摄影机正面的方向为前方，采由左至右、由近至远的方式对侦测范围中每个位置逐一侦测。全区域侦测结束后，得到全区域侦测结果。Generally speaking, the mobile object detection method of the present disclosure includes two stages: (1) whole area detection; and (2) partial area detection. Please refer to Figure 1. FIG. 1 is a flowchart of an embodiment of the mobile object detection method of the present disclosure. As shown in FIG. 1 , in step S100 : perform full-area detection. For example, take a shot 30 meters ahead and open the area surrounded by a horizontal viewing angle of 100 degrees as the detection range. The camera gets a number of viewframes of the detection range whose content covers the detection range. The above-mentioned frames can form an image stream, for example, twenty frames per second (20f/s). The plurality of frames may include at least a first frame and a second frame. Get the first frame according to the detection range, and perform full area detection on the first frame. For example, the detection range is divided into multiple areas, and the front of the camera is taken as the front, and each position in the detection range is detected one by one from left to right and from near to far. After the detection of the whole area is completed, the detection result of the whole area is obtained.
在得到全区域侦测结果后，接收第二视格。在步骤S200：定义侦测范围并于第二视格执行部分区域侦测。根据全区域侦测结果定义侦测范围包含侦测区及非侦测区，并在第二视格的侦测区执行部分区域侦测。例如，在第一视格的全区域侦测完成后侦测到移动物，在接下来的第二视格中将移动物所在位置定义为侦测区，而移动物以外的位置则为非侦测区。亦即，在第二视格中仅以移动物所在位置作侦测，由此提升运算效率。After getting the detection result of the whole area, receive the second visual frame. In step S200 : define a detection range and perform partial area detection in the second frame. Define the detection range including the detection area and the non-detection area according to the detection result of the whole area, and perform partial area detection in the detection area of the second visual frame. For example, if a moving object is detected after the detection of the entire area in the first frame is completed, in the next second frame the location of the moving object is defined as the detection area, while the location outside the moving object is defined as a non-detection area measuring area. That is, only the position of the moving object is used for detection in the second viewing frame, thereby improving the computing efficiency.
图2为本公开驾驶辅助系统的一实施例方框图。如图2所示，驾驶辅助系统100包含CPU110、影像获取模块120、处理模块130、移动物侦测模块140、速度感测模块150、定位模块160、周期设定模块170。此外，驾驶辅助系统100与摄影模块200及显示模块300连接。FIG. 2 is a block diagram of an embodiment of the disclosed driving assistance system. As shown in FIG. 2 , the driving assistance system 100 includes a CPU 110 , an image acquisition module 120 , a processing module 130 , a moving object detection module 140 , a speed sensing module 150 , a positioning module 160 , and a cycle setting module 170 . In addition, the driving assistance system 100 is connected to the camera module 200 and the display module 300 .
举例而言，摄影模块200包括摄影机及镜片组，以拍摄场景数据。影像获取模块120可为CMOS传感器或CCD传感器，配合摄影机进行视格获取，以持续取得多个视格。处理模块130可为数字信号处理电路，以进行前述侦测区/非侦测区的区分。移动物侦测模块140可为数字图像处理单元，以及配合超声波雷达或红外线雷达。利用数字图像处理单元对视格内容进行特征辨识，利用雷达对视格中各位置进行侦测，并测量与移动物的相对距离、移动物速度等数据。速度感测模块150可为加速度传感器或车轮转速传感器，CPU110通过速度感测模块150可取得当前的驾驶速度。定位模块160可为GPS或转角传感器，CPU110通过定位模块160可获知驾驶方向。周期设定模块170可为频率单元，CPU110可根据视格接收速率变更周期设定模块170的设定值，以调整移动物侦测模块140的侦测方式。显示模块300可为实体的显示器或是抬头显示器。驾驶辅助系统100可将取得的视格于显示模块300呈现，另外，当侦测到移动物时，驾驶辅助系统100也可通过显示模块300警示驾驶人。For example, the camera module 200 includes a camera and a lens set to capture scene data. The image acquisition module 120 can be a CMOS sensor or a CCD sensor, and cooperates with a camera to acquire frames to continuously acquire multiple frames. The processing module 130 can be a digital signal processing circuit to perform the aforementioned detection region/non-detection region distinction. The moving object detection module 140 can be a digital image processing unit, and cooperate with ultrasonic radar or infrared radar. Use the digital image processing unit to identify the features of the view frame content, use the radar to detect each position in the view frame, and measure the relative distance to the moving object, the speed of the moving object and other data. The speed sensing module 150 can be an acceleration sensor or a wheel speed sensor, and the CPU 110 can obtain the current driving speed through the speed sensing module 150 . The positioning module 160 can be a GPS or a rotation angle sensor, and the CPU 110 can obtain the driving direction through the positioning module 160 . The period setting module 170 can be a frequency unit, and the CPU 110 can change the setting value of the period setting module 170 according to the frame receiving rate to adjust the detection method of the moving object detection module 140 . The display module 300 can be a physical display or a head-up display. The driving assistance system 100 can present the acquired visual frame on the display module 300 . In addition, when a moving object is detected, the driving assistance system 100 can also alert the driver through the display module 300 .
图3为本公开移动物侦测方法执行不同部分区域侦测的实施例流程图。在图3中，部分区域侦测阶段是针对全区域侦测结果是否有移动物的出现而对第二视格提供不同侦测方式。如图3所示，在步骤S100：执行全区域侦测。例如，在第一视格全区域侦测完成后，得到全区域侦测结果。接着接收第二视格，在步骤S202确认前一视格是否有移动物。承上述的实施例，若第一视格的全区域侦测结果未侦测到移动物，则为步骤S220：侦测边缘区域。换言之，选取侦测范围中的边缘区域作为侦测区，对第二视格的边缘区域执行部分区域侦测。FIG. 3 is a flow chart of an embodiment of performing detection of different parts of the mobile object detection method of the present disclosure. In FIG. 3 , in the part area detection stage, different detection methods are provided for the second visual frame according to whether there is a moving object in the whole area detection result. As shown in FIG. 3 , in step S100 : perform full-area detection. For example, after the detection of the entire area in the first viewing frame is completed, the detection result of the entire area is obtained. Next, the second frame is received, and in step S202, it is confirmed whether there is a moving object in the previous frame. According to the above-mentioned embodiment, if no moving object is detected in the whole area detection result of the first visual frame, it is step S220: detect the edge area. In other words, the edge area in the detection range is selected as the detection area, and partial area detection is performed on the edge area of the second visual frame.
请参考图4。图4为未侦测到移动物时执行部分区域侦测的上视图。如图4所示，车400在前方场景中设定有侦测范围R。在执行全区域侦测时，即以侦测范围R内的所有位置进行侦测。当全区域侦测完成后未侦测到移动物时，选取边缘区域P作为侦测区。边缘区域P包括接近车400的位置(近端)、沿视角自两侧延伸的边缘带以及远程的水平边缘带。考虑到未侦测移动物时，后续移动物自视线边缘处进入侦测范围，因此选取边缘区域P作为侦测区。此时边缘区域P即为部分区域侦测的范围，而中央区域Q则为非侦测区。由此局部侦测的方式可省下系统运算的时间。Please refer to Figure 4. FIG. 4 is a top view of performing partial area detection when no moving object is detected. As shown in FIG. 4 , the car 400 has a detection range R set in the front scene. When performing full-area detection, all positions within the detection range R are used for detection. When no moving object is detected after the whole area detection is completed, the edge area P is selected as the detection area. The edge region P includes a position close to the car 400 (near end), edge bands extending from both sides along the viewing angle, and a remote horizontal edge band. Considering that when no moving object is detected, the subsequent moving object enters the detection range from the edge of the line of sight, so the edge area P is selected as the detection area. At this time, the edge area P is the detection range of some areas, and the central area Q is the non-detection area. In this way, the local detection method can save the time of system operation.
反之，若第一视格的全区域侦测结果侦测到移动物，如图3所示，则为步骤S210：产生目标区域。当侦测到移动物时，定义移动物在侦测范围的位置，并在移动物周围产生目标区域。在步骤S212：侦测边缘区域及目标区域。选取侦测范围中的边缘区域及目标区域作为侦测区，在第二视格的边缘区域及目标区域执行部分区域侦测。On the contrary, if the whole area detection result of the first visual frame detects a moving object, as shown in FIG. 3 , it is step S210 : generating a target area. When a moving object is detected, define the position of the moving object within the detection range and generate a target area around the moving object. In step S212 : detecting the edge area and the target area. Select the edge area and the target area in the detection range as the detection area, and perform partial area detection in the edge area and the target area of the second visual frame.
请参考图5A及图5B。图5A及图5B为侦测到移动物时执行部分区域侦测的上视图。如图5A所示，在侦测范围R内除了可设定边缘区域P，还可以针对移动物的位置于移动物500周围产生目标区域M。当全区域侦测完成后侦测到移动物时，选取边缘区域P及目标区域M作为侦测区。此时边缘区域P及目标区域M涵盖的位置即为部分区域侦测执行的范围，而中央区域Q(扣除边缘区域P及目标区域M后的部分)则为非侦测区。利用目标区域的设定追踪已侦测到的移动物，以及利用边缘区域的设定追踪可能出现的移动物。由此局部侦测的方式可省下系统运算的时间。Please refer to FIG. 5A and FIG. 5B . FIG. 5A and FIG. 5B are top views of partial region detection performed when a moving object is detected. As shown in FIG. 5A , in addition to setting the edge area P within the detection range R, a target area M can also be generated around the moving object 500 according to the position of the moving object. When a moving object is detected after the whole area detection is completed, the edge area P and the target area M are selected as the detection area. At this time, the area covered by the edge area P and the target area M is the area where partial area detection is performed, and the central area Q (the part after deducting the edge area P and the target area M) is a non-detection area. Use the target area setting to track detected moving objects, and use the edge area setting to track possible moving objects. In this way, the local detection method can save the time of system operation.
在一实施例中，目标区域是根据移动物速度、驾驶速度或驾驶方向所产生。例如前述的CPU通过移动物侦测模块取得移动物距离与速度，并通过处理模块在移动物四周产生目标区域作为侦测区。In one embodiment, the target area is generated according to the speed of the moving object, the driving speed or the driving direction. For example, the aforementioned CPU obtains the distance and speed of the moving object through the moving object detection module, and generates a target area around the moving object as the detection area through the processing module.
此外，目标区域的范围大小可根据移动物速度、视格速率来调整。此处视格速率是指每秒视格的数量。如图5B所示，传感器600通过镜片602于侦测范围R内侦测到移动物500。其中传感器600例如为前述的影像获取模块，镜片602例如为前述的摄影模块。在移动物500周围设定有目标区域M。目标区域M具有边长m大于移动物500宽度w，并涵盖移动物500所在位置。In addition, the size of the target area can be adjusted according to the speed of the moving object and the frame rate. The viewframe rate here refers to the number of viewframes per second. As shown in FIG. 5B , the sensor 600 detects the moving object 500 within the detection range R through the lens 602 . The sensor 600 is, for example, the aforementioned image acquisition module, and the lens 602 is, for example, the aforementioned camera module. A target area M is set around the moving object 500 . The target area M has a side length m greater than the width w of the moving object 500 and covers the location of the moving object 500 .
具体而言，例如移动物速度为10m/s，对于10f/s的视格速率而言，每个视格间移动物的移动距离则为1m。在一实施例中，目标区域的边长不大于移动物宽度的3倍(亦即，目标区域的边长不大于移动物速度与视格速率的比值的3倍)。根据上述计算方式可设定目标区域具有边长3m大小的范围。由此可推估目标区域的大小。Specifically, for example, if the speed of the moving object is 10m/s, for a viewing frame rate of 10f/s, the moving distance of the moving object between each viewing frame is 1m. In one embodiment, the side length of the target area is not greater than 3 times the width of the moving object (that is, the side length of the target area is not greater than 3 times the ratio of the speed of the moving object to the viewing frame rate). According to the above calculation method, the target area can be set to have a range with a side length of 3 m. From this, the size of the target area can be estimated.
需补充的是，上述移动物速度的数据亦可为预设的，例如用于侦测行人的情形，可预设行人的步行速度上限值作为设定目标区域的根据。在其它实施方式中，可针对不同的辨识特征建立不同移动物与移动速度的对应表。通过特征辨识找出对应的移动物速度的数据以作为设定目标区域的根据。What needs to be added is that the speed data of the above moving objects can also be preset, for example, in the case of detecting pedestrians, the upper limit of walking speed of pedestrians can be preset as the basis for setting the target area. In other implementation manners, a correspondence table between different moving objects and moving speeds may be established for different identification features. Find the corresponding moving object speed data through feature recognition as the basis for setting the target area.
应理解，所述目标区域的边长是指各边加总后的平均边长。另外，目标区域的形状并不以图5B所示的正三角框为限。根据后续视格中对移动物移动方向的侦测结果可进一步沿移动物的移动方向调整目标区域的形状及位置。例如，在第一视格侦测结束后针对移动物产生目标区域，接着在第二视格侦测结束后，对比第一视格而判断出移动物的移动方向，根据移动物的移动方向调整目标区域的形状。对于设定为圆形或近似圆形的目标区域而言，设定范围则是指径向长度的平均值(例如长轴与短轴长度的平均)。It should be understood that the side length of the target area refers to an average side length after summing up the sides. In addition, the shape of the target area is not limited to the regular triangle box shown in FIG. 5B . The shape and position of the target area can be further adjusted along the moving direction of the moving object according to the detection result of the moving object's moving direction in the subsequent view frame. For example, after the detection of the first frame is completed, the target area is generated for the moving object, and then after the detection of the second frame is completed, the moving direction of the moving object is determined by comparing with the first frame, and adjusted according to the moving direction of the moving object The shape of the target area. For a target area set to be circular or approximately circular, the setting range refers to the average value of the radial length (for example, the average of the lengths of the major axis and the minor axis).
图6为边缘区域组成的示意图。如图6所示，边缘区域包含第一侧缘P1、第二侧缘P2及远程边缘P3。第一侧缘与第二侧缘夹一角度且延伸连接远程边缘的两端。换言之，边缘区域为水平视角两侧的外缘，以及两侧朝水平面延伸至远程所包围的带状区。Fig. 6 is a schematic diagram of the composition of the edge region. As shown in FIG. 6 , the edge area includes a first side edge P1 , a second side edge P2 and a remote edge P3 . The first side edge forms an angle with the second side edge and extends to connect two ends of the remote edge. In other words, the edge area is the outer edge on both sides of the horizontal viewing angle, and the band-shaped area surrounded by the two sides extending toward the horizontal plane to the far end.
在一实施例中，上述边缘区域的选取方式可根据驾驶速度于侦测范围内调整远程边缘的大小。例如在图6中，当车速增加时，在侦测范围R内增加远程边缘P3的大小。在另一实施例中，可以同时调整第一侧缘P1、第二侧缘P2及远程边缘P3的范围。具体而言，例如侦测距离(深度)r1为30m，远程水平距离r2为40m。侦测得到移动物500速度为10m/s，对于10f/s的视格速率而言，每个视格间移动物500的移动距离则为1m。可设定中央区域Q深度q1为28m(两端较r1缩减1m)，中央区域Q远程水平距离q2为38m(两端较r2缩减1m)。由此可推估边缘区域的大小。In one embodiment, the selection method of the above-mentioned edge area can adjust the size of the remote edge within the detection range according to the driving speed. For example, in FIG. 6 , when the vehicle speed increases, the size of the remote edge P3 within the detection range R increases. In another embodiment, the ranges of the first side edge P1 , the second side edge P2 and the remote edge P3 can be adjusted simultaneously. Specifically, for example, the detection distance (depth) r1 is 30m, and the remote horizontal distance r2 is 40m. The detected speed of the moving object 500 is 10 m/s, and for a viewing frame rate of 10 f/s, the moving distance of the moving object 500 between each viewing frame is 1 m. The depth q1 of the central area Q can be set to 28m (the two ends are reduced by 1m compared with r1), and the remote horizontal distance q2 of the central area Q is 38m (the two ends are reduced by 1m compared to r2). From this, the size of the edge region can be estimated.
移动物侦测方法除前述两个阶段以外，还可包含：(3)周期设定。请参考图7。图7为本公开移动物侦测方法具有侦测周期设定的实施例流程图。如图7所示，在进入步骤S100及其后续的侦测步骤前，进行步骤S300：设定侦测周期。例如，设定侦测周期三张视格，则第一视格执行全区域侦测后，接着在第二视格与第三视格执行部分区域侦测，在第四视格会再次执行全区域侦测。In addition to the aforementioned two stages, the moving object detection method may also include: (3) Period setting. Please refer to Figure 7. FIG. 7 is a flow chart of an embodiment of the mobile object detection method with detection period setting in the present disclosure. As shown in FIG. 7 , before entering step S100 and subsequent detection steps, proceed to step S300 : set the detection cycle. For example, if the detection cycle is set to three frames, after the first frame performs full-area detection, then performs partial area detection in the second and third frames, and executes full-scale detection again in the fourth frame Area detection.
在步骤S302确认周期是否结束。若否，则重复进行部分区域侦测(如上述的实施例，当第二视格的部分区域侦测结束后由于周期尚未结束，因此于第三视格进行部分区域侦测)，即重复步骤S202、S210、S212；或是重复步骤S202、S220。应理解，在上述的实施例中，第二视格是根据全区域侦测结果来决定侦测区与非侦测区；第三视格是根据第二视格的侦测结果来决定侦测区与非侦测区。In step S302, it is confirmed whether the period is over. If not, repeat the partial area detection (as in the above-mentioned embodiment, after the partial area detection of the second visual frame is completed, because the cycle has not yet ended, the partial area detection is performed in the third visual frame), that is, the steps are repeated S202, S210, S212; or repeat steps S202, S220. It should be understood that, in the above-mentioned embodiment, the detection area and the non-detection area are determined according to the detection result of the whole area in the second visual frame; the detection is determined in the third visual frame according to the detection result of the second visual frame area and non-detection area.
反之，当步骤S302判断侦测周期结束，则执行全区域侦测。对于上述的实施例，在接收第四视格时执行全区域侦测。类似地，在第五视格，则是根据第四视格的全区域侦测结果定义侦测区与非侦测区。换言之，对于第一视格之后的视格，侦测周期与执行全区域侦测具有以下关系：n＝mx+1，其中，n为视格序数(2,3,4……)、m为侦测周期(例如3)、x为周期序数(1,2,3……)。亦即，每m张视格之后取得的一张视格进行全区域侦测。若将第一视格纳入上式一并考虑，则为x＝0的情形。所述侦测周期(m值)为大于或等于2的整数。由此可利用一段时间内执行全区域侦测以确保原先设定的未侦测区中出现突发状况。On the contrary, when it is determined in step S302 that the detection period is over, the whole area detection is performed. For the embodiments described above, full area detection is performed when receiving the fourth frame. Similarly, in the fifth frame, the detection area and the non-detection area are defined according to the detection result of the whole area in the fourth frame. In other words, for the viewframes after the first viewframe, the detection cycle has the following relationship with the execution of full-area detection: n=mx+1, where n is the viewframe number (2, 3, 4...), m is Detection cycle (eg 3), x is the cycle number (1,2,3...). That is to say, one visual frame obtained after every m visual frames is used for the whole area detection. If the first visual lattice is taken into consideration in the above formula, it is the case of x=0. The detection period (m value) is an integer greater than or equal to 2. Therefore, it is possible to use a period of time to perform full-area detection to ensure that emergencies occur in the originally set undetected area.
图8为本公开移动物侦测方法判断全区域侦测过程的实施例流程图。在图8中，全区域侦测阶段可针对移动物的出现于第一视格提供不同侦测方式。如图8所示，在步骤S300：设定侦测周期。接着进入全区域侦测阶段。在步骤S100：执行全区域侦测。在步骤S102中执行全区域侦测时确认是否有移动物。若未侦测到移动物，则为步骤S103：持续完成全区域侦测，并于步骤S108判断全区域侦测是否完成，接着进行后续的部分区域侦测(步骤S202～S220)。另举一例，当设定侦测周期为每两张视格执行一次全区域侦测，在第一视格执行全区域侦测，并于全区域侦测进行期间持续判断是否有移动物。若全区域侦测完成，在侦测范围中的每个位置皆侦测过而仍未侦测到移动物，产生全区域侦测结果，接着取得第二视格进行后续步骤(部分区域侦测)。FIG. 8 is a flowchart of an embodiment of the process of judging the whole area detection of the mobile object detection method of the present disclosure. In FIG. 8 , the whole-area detection phase can provide different detection methods for the presence of moving objects in the first frame. As shown in FIG. 8 , in step S300 : set the detection period. Then enter the whole area detection stage. In step S100: Execute full-area detection. Whether there is a moving object is confirmed when the whole area detection is performed in step S102. If no moving object is detected, it is step S103: continue to complete the whole area detection, and determine whether the whole area detection is completed in step S108, and then carry out subsequent partial area detection (steps S202-S220). As another example, when the detection period is set to perform full-area detection every two frames, the full-area detection is performed in the first frame, and whether there is a moving object is continuously judged during the full-area detection. If the detection of the whole area is completed, every position in the detection range has been detected but no moving object has been detected, and the detection result of the whole area is generated, and then the second visual frame is obtained for subsequent steps (partial area detection) ).
反之，在步骤S102若判断侦测到移动物，则为步骤S104：定义阻隔区。承上例，若在第一视格的全区域侦测进行期间侦测到移动物，定义移动物后方的部分区域为阻隔区。举例来说，当相距摄影机近处侦测到移动物，将移动物正后方(相距摄影机较远处)的范围定义为阻隔区。On the contrary, if it is determined in step S102 that a moving object is detected, then proceed to step S104: define a blocking area. Continuing from the above example, if a moving object is detected during the whole-area detection of the first viewing frame, define the part of the area behind the moving object as the blocking area. For example, when a moving object is detected close to the camera, the range directly behind the moving object (farther away from the camera) is defined as the blocking area.
接着于步骤S106：忽略阻隔区并继续执行其他部分的全区域侦测。当判断全区域侦测完成，得到全区域侦测结果，接着取得第二视格进行后续步骤。换言之，全区域侦测可随侦测到移动物的出现而设定略过的区域，由此增进全区域侦测的效率。Then in step S106 : ignoring the blocking area and continuing to perform other parts of the whole area detection. When it is judged that the detection of the whole area is completed, the result of the detection of the whole area is obtained, and then the second visual frame is obtained for subsequent steps. In other words, the whole-area detection can set skipped areas when the presence of moving objects is detected, thereby improving the efficiency of the whole-area detection.
在步骤S302确认周期是否结束。依上述的实施例，由于设定侦测周期为每两张视格执行一次全区域侦测，当第二视格的部分区域侦测结束，接着根据侦测范围取得第三视格，在第三视格执行全区域侦测。类似地，根据第三视格的全区域侦测结果可重新定义侦测区及非侦测区，并根据侦测范围取得第四视格。然后于第四视格的侦测区执行部分区域侦测。简言之，依前述设定，取得视格与侦测方式的关系为：第一视格(全区域)、第二视格(部分区域)、第三视格(全区域)、第四视格(部分区域)，利用局部侦测的方式可省下系统运算的时间。同时，在上例中，第一视格及第三视格的侦测过程中皆可随侦测到移动物而定义略过的区域，进一步省下系统运算时间。In step S302, it is confirmed whether the period is over. According to the above-mentioned embodiment, since the detection cycle is set to perform a full-area detection every two frames, when the detection of the partial area of the second frame is completed, then the third frame is obtained according to the detection range. Triple-view performs full-area detection. Similarly, the detection area and non-detection area can be redefined according to the detection result of the whole area in the third frame, and the fourth frame can be obtained according to the detection range. Then perform partial area detection in the detection area of the fourth frame. In short, according to the above settings, the relationship between the obtained view frame and the detection method is: the first view frame (full area), the second view frame (partial area), the third view frame (whole area), and the fourth view frame Grid (partial area), the use of local detection can save the time of system operation. At the same time, in the above example, during the detection process of the first frame and the third frame, skipped regions can be defined as moving objects are detected, further saving system computing time.
在一实施例中，当执行全区域侦测时判断有移动物时，选取位于移动物后方的范围为阻隔区。请参考图9A至图9D。图9A为采用移动物后方为阻隔区的上视图。如图9A所示，在侦测范围R内侦测到移动物500，并选取移动物500后方(延伸至侦测范围R远程)的范围为阻隔区N。阻隔区N的两侧可依移动物500宽度所张开的视角朝侦测范围R远程延伸，使阻隔区N自上视角度整体略呈梯形区域。In one embodiment, when it is determined that there is a moving object during the full-area detection, the area behind the moving object is selected as the blocking area. Please refer to FIG. 9A to FIG. 9D . FIG. 9A is a top view of using the rear of the moving object as the blocking area. As shown in FIG. 9A , a moving object 500 is detected within the detection range R, and the area behind the moving object 500 (extending to the far end of the detection range R) is selected as the blocking area N. Both sides of the blocking area N can extend far away toward the detection range R according to the viewing angle opened by the width of the moving object 500 , so that the blocking area N is generally a trapezoidal area viewed from above.
图9B至9D为说明图9A设定方式的示意图。以图9B正视图为例，驾驶前方侦测范围内有不同的移动物，标示于虚线框500A与虚线框500B之中。当侦测到位于较接近驾驶人的移动物500，后方(远程)虚线框500B所示处则忽略而不进行侦测。亦即，阻隔区涵盖虚线框500B所在位置。9B to 9D are schematic diagrams illustrating the setting method of FIG. 9A. Taking the front view of FIG. 9B as an example, there are different moving objects within the detection range in front of the driver, which are marked in the dotted line frame 500A and the dotted line frame 500B. When the moving object 500 located closer to the driver is detected, the rear (far) dotted frame 500B is ignored and not detected. That is, the blocking area covers the location of the dotted box 500B.
具体而言，如图9C及图9D所示，侦测范围R中，在位置T侦测到移动物500。自感测端(镜片602)依移动物500宽度所张开的视角朝侦测范围R远程延伸构成一自上视整体略呈三角形范围SCD。其中，于范围SVD内可根据移动物500距离ST、移动物500半宽度ET及侦测距离SV的比例关系计算出DV的长度。邻近移动物500后方的水平距离AU可依上述方式计算得到。类似地，在范围SVC内可计算出CV及BU的值。根据上述计算结果可设定范围ABCD为阻隔区。由此可在全区域侦侧时省略部分区域不执行侦测，以进一步提升运算效率。Specifically, as shown in FIG. 9C and FIG. 9D , in the detection range R, a moving object 500 is detected at a position T. As shown in FIG. The viewing angle opened from the sensing end (lens 602 ) according to the width of the moving object 500 extends far away toward the detection range R to form a slightly triangular range SCD viewed from above. Wherein, within the range SVD, the length of DV can be calculated according to the proportional relationship between the distance ST of the moving object 500 , the half-width ET of the moving object 500 , and the detection distance SV. The horizontal distance AU behind the adjacent moving object 500 can be calculated in the above manner. Similarly, the values of CV and BU can be calculated within the range SVC. According to the above calculation results, the range ABCD can be set as the barrier area. Therefore, it is possible to omit part of the region from performing detection when the whole region is detected, so as to further improve the calculation efficiency.
综上所述，本公开的移动物侦测方法可根据侦测结果调整侦测范围。此外，亦可考虑全区域侦测的过程是否出现移动物而决定是否部分忽略不侦测。由此可提高侦侧效率。To sum up, the mobile object detection method of the present disclosure can adjust the detection range according to the detection result. In addition, it is also possible to consider whether there is a moving object in the whole area detection process and decide whether to partially ignore it or not to detect it. As a result, detection efficiency can be improved.
本公开已由上述相关实施例加以描述，然而上述实施例仅为实施本公开的范例。必需指出的是，已公开的实施例并未限制本公开的范围。相反地，包含于权利要求的精神及范围的修改及均等设置均包含于本公开的范围内。The present disclosure has been described by the above-mentioned related embodiments, but the above-mentioned embodiments are only examples for implementing the present disclosure. It must be noted that the disclosed embodiments do not limit the scope of the present disclosure. On the contrary, modifications and equivalent arrangements included in the spirit and scope of the claims are included in the scope of the present disclosure.
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| TW105117072ATWI571830B (en) | 2016-05-31 | 2016-05-31 | Moving object detecting method |
| TW105117072 | 2016-05-31 |
| Publication Number | Publication Date |
|---|---|
| CN107451519Atrue CN107451519A (en) | 2017-12-08 |
| CN107451519B CN107451519B (en) | 2021-02-05 |
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710103633.9AActiveCN107451519B (en) | 2016-05-31 | 2017-02-24 | Moving object detection method |
| Country | Link |
|---|---|
| CN (1) | CN107451519B (en) |
| TW (1) | TWI571830B (en) |
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112288774A (en)* | 2020-10-22 | 2021-01-29 | 深圳市华宝电子科技有限公司 | Movement detection method and device, electronic equipment and storage medium |
| TWI761047B (en)* | 2021-01-22 | 2022-04-11 | 明泰科技股份有限公司 | Method for verifying detection range of image detector |
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| TWI676152B (en)* | 2018-05-08 | 2019-11-01 | 中華電信股份有限公司 | A moving object detection method |
| TWI783390B (en)* | 2021-02-26 | 2022-11-11 | 圓展科技股份有限公司 | Image processing system and method for generating dynamic image segmentation |
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101080003A (en)* | 2006-05-24 | 2007-11-28 | 深圳市景阳数码技术有限公司 | Mobile detection alert of vidicon and digital amplification linkage method |
| CN101161524A (en)* | 2006-10-12 | 2008-04-16 | 财团法人车辆研究测试中心 | Method and device for detecting vehicle distance |
| US20080100709A1 (en)* | 2006-10-27 | 2008-05-01 | Matsushita Electric Works, Ltd. | Target moving object tracking device |
| CN101415103A (en)* | 2007-10-19 | 2009-04-22 | 英华达（南京）科技有限公司 | Method for regulating video signal picture |
| CN101469987A (en)* | 2007-12-29 | 2009-07-01 | 财团法人工业技术研究院 | Moving object detection device and method |
| CN101605209A (en)* | 2008-05-26 | 2009-12-16 | 三洋电机株式会社 | Camera head and image-reproducing apparatus |
| CN101807131A (en)* | 2009-02-13 | 2010-08-18 | 华信光电科技股份有限公司 | Detection module and optical detection system including the detection module |
| CN101894378A (en)* | 2010-06-13 | 2010-11-24 | 南京航空航天大学 | Method and system for visual tracking of moving target based on dual regions of interest |
| CN102131076A (en)* | 2011-01-19 | 2011-07-20 | 中兴通讯股份有限公司 | Video monitoring method and monitoring terminal |
| CN102170558A (en)* | 2010-12-30 | 2011-08-31 | 财团法人车辆研究测试中心 | Obstacle detection warning system and method |
| US20110254955A1 (en)* | 2010-04-18 | 2011-10-20 | Peter Chi-Chen Shen | Detachable Universal Electronic Door Viewer |
| CN102438122A (en)* | 2010-09-29 | 2012-05-02 | 鸿富锦精密工业（深圳）有限公司 | Camera device and method for dynamically detecting monitoring object by using same |
| CN103043000A (en)* | 2011-10-13 | 2013-04-17 | 能晶科技股份有限公司 | Obstacle detection system and obstacle detection method thereof |
| CN203005249U (en)* | 2012-10-29 | 2013-06-19 | 华创车电技术中心股份有限公司 | Vehicular omnidirectional warning system |
| US20130156269A1 (en)* | 2011-12-20 | 2013-06-20 | Fujitsu Limited | Information processing apparatus and method |
| CN204432515U (en)* | 2014-12-22 | 2015-07-01 | 合肥晟泰克汽车电子有限公司 | A kind of vehicle blind zone dynamic monitoring system |
| CN105015408A (en)* | 2014-04-25 | 2015-11-04 | 帷享科技有限公司 | Obstacle Warning System |
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101080003A (en)* | 2006-05-24 | 2007-11-28 | 深圳市景阳数码技术有限公司 | Mobile detection alert of vidicon and digital amplification linkage method |
| CN101161524A (en)* | 2006-10-12 | 2008-04-16 | 财团法人车辆研究测试中心 | Method and device for detecting vehicle distance |
| US20080100709A1 (en)* | 2006-10-27 | 2008-05-01 | Matsushita Electric Works, Ltd. | Target moving object tracking device |
| CN101415103A (en)* | 2007-10-19 | 2009-04-22 | 英华达（南京）科技有限公司 | Method for regulating video signal picture |
| CN101469987A (en)* | 2007-12-29 | 2009-07-01 | 财团法人工业技术研究院 | Moving object detection device and method |
| CN101605209A (en)* | 2008-05-26 | 2009-12-16 | 三洋电机株式会社 | Camera head and image-reproducing apparatus |
| CN101807131A (en)* | 2009-02-13 | 2010-08-18 | 华信光电科技股份有限公司 | Detection module and optical detection system including the detection module |
| US20110254955A1 (en)* | 2010-04-18 | 2011-10-20 | Peter Chi-Chen Shen | Detachable Universal Electronic Door Viewer |
| CN101894378A (en)* | 2010-06-13 | 2010-11-24 | 南京航空航天大学 | Method and system for visual tracking of moving target based on dual regions of interest |
| CN102438122A (en)* | 2010-09-29 | 2012-05-02 | 鸿富锦精密工业（深圳）有限公司 | Camera device and method for dynamically detecting monitoring object by using same |
| CN102170558A (en)* | 2010-12-30 | 2011-08-31 | 财团法人车辆研究测试中心 | Obstacle detection warning system and method |
| CN102131076A (en)* | 2011-01-19 | 2011-07-20 | 中兴通讯股份有限公司 | Video monitoring method and monitoring terminal |
| CN103043000A (en)* | 2011-10-13 | 2013-04-17 | 能晶科技股份有限公司 | Obstacle detection system and obstacle detection method thereof |
| US20130156269A1 (en)* | 2011-12-20 | 2013-06-20 | Fujitsu Limited | Information processing apparatus and method |
| CN203005249U (en)* | 2012-10-29 | 2013-06-19 | 华创车电技术中心股份有限公司 | Vehicular omnidirectional warning system |
| CN105015408A (en)* | 2014-04-25 | 2015-11-04 | 帷享科技有限公司 | Obstacle Warning System |
| CN204432515U (en)* | 2014-12-22 | 2015-07-01 | 合肥晟泰克汽车电子有限公司 | A kind of vehicle blind zone dynamic monitoring system |
| Title |
|---|
| SIDIB´E D等: "Using visual saliency for object tracking with particle filters", 《PROCEEDINGS OF THE 18TH EUROPEAN SIGNAL PROCESSING CONFERENCE》* |
| 黎万义等: "引入视觉注意机制的目标跟踪方法综述", 《自动化学报》* |
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112288774A (en)* | 2020-10-22 | 2021-01-29 | 深圳市华宝电子科技有限公司 | Movement detection method and device, electronic equipment and storage medium |
| CN112288774B (en)* | 2020-10-22 | 2024-01-30 | 深圳市华宝电子科技有限公司 | Mobile detection method, mobile detection device, electronic equipment and storage medium |
| TWI761047B (en)* | 2021-01-22 | 2022-04-11 | 明泰科技股份有限公司 | Method for verifying detection range of image detector |
| Publication number | Publication date |
|---|---|
| CN107451519B (en) | 2021-02-05 |
| TWI571830B (en) | 2017-02-21 |
| TW201742002A (en) | 2017-12-01 |
| Publication | Publication Date | Title |
|---|---|---|
| JP5783243B2 (en) | Periodic stationary object detection apparatus and periodic stationary object detection method | |
| JP5870908B2 (en) | Vehicle collision determination device | |
| CN102016921B (en) | image processing device | |
| US20170242117A1 (en) | Vision algorithm performance using low level sensor fusion | |
| JP6361592B2 (en) | Vehicle control device | |
| CN107451519A (en) | Moving object detection method | |
| CN106485233A (en) | Drivable region detection method, device and electronic equipment | |
| JP2017207973A (en) | Detector and detection method | |
| CN102414715A (en) | Object detection device | |
| CN112507862A (en) | Vehicle orientation detection method and system based on multitask convolutional neural network | |
| JP2014160322A (en) | Lane boundary deviation suppression device | |
| JP6747389B2 (en) | Collision estimating device and collision estimating method | |
| JP2018097765A (en) | Object detection device and object detection method | |
| CN111052201B (en) | Collision prediction device, collision prediction method, and storage medium | |
| JP2008117073A (en) | Interrupting vehicle detection device | |
| CN107516423B (en) | Video-based vehicle driving direction detection method | |
| US9076034B2 (en) | Object localization using vertical symmetry | |
| JP2018147399A (en) | Target detection device | |
| WO2017043358A1 (en) | Object detecting device and object detecting method | |
| Nakamura et al. | Real-time monocular ranging by Bayesian triangulation | |
| Di et al. | Forward Collision Warning system based on vehicle detection and tracking | |
| JP2012164275A (en) | Image recognition apparatus | |
| JP7327992B2 (en) | Three-dimensional object detection device and three-dimensional object detection method | |
| JP2013142972A (en) | Travel path recognition device | |
| US9811744B2 (en) | Fast and robust stop line detector |
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |