Risk assessment support device

Abstract

To solve the problems of conventional arts, that is, to provide a risk assessment support device that can assess a risk of a work area while allowing for false detection, rather than issuing an alert immediately.SOLUTION: A risk assessment support device is a device that assesses a risk of a work area set in a target area based on on-site images of the target area periodically captured by imaging means installed at a certain location, and comprises area setting means, risk indicator setting means, worker detection means, entry determination means, and assessment value calculation means. The assessment value calculation means is means for calculating a risk assessment value based on a product of the number of entries for the work area and a risk indicator for the work area. The risk of the work area is then assessed based on the risk assessment value.SELECTED DRAWING: Figure 3

Description

[Technical field]

[0001] The present invention relates to technology related to safety management in construction work, and more specifically, to a risk assessment support device that can assess the risk of a work area based on an index calculated from the risk of the work area and the number of times workers enter the area. [Background technology]

[0002] According to the "Current Status of Safety and Health in the Construction Industry" compiled by the Ministry of Health, Labor and Welfare and the Ministry of Land, Infrastructure, Transport and Tourism in February 2023, fatal accidents in the construction industry have decreased significantly over the past 50 years up to 2021. Looking at the detailed breakdown, while the decrease was 1 / 7 (from 6,208 to 867) across all industries, the decrease in the construction industry was 1 / 9 (from 2,492 to 288).

[0003] The background to this significant reduction in industrial accidents is largely due to thorough safety management at construction sites.Specifically, various initiatives for safe work are implemented at each construction site, such as the use of safety harnesses when working at height, the installation of safety pathways, prohibition of entry within the working radius of heavy machinery or under suspended loads, and KY (hazard prediction) activities and confirmation of construction procedures at toolbox meetings.

[0004] However, we have not yet completely eliminated industrial accidents at construction sites, and efforts are being made to further improve safety. For example, workers may inadvertently enter the working radius of heavy machinery or under the load hoisted by a crane, and in order to prevent this, a system that issues an alert immediately is sometimes introduced. By installing detection devices such as laser barriers in dangerous areas, the intrusion of a worker is detected and an alert is immediately output to the worker using sound or light.

[0005] In recent years, technology has also come into use that uses images instead of laser barriers to detect the entry of workers. For example, Patent Document 1 proposes a technology in which a photographing means capable of photographing the entire construction site is installed, the contours (areas) of construction machinery and workers are captured from the images obtained by the photographing means, and an alert is issued when the construction machinery area and the worker area come into contact with each other. [Prior art documents] [Patent documents]

[0006] [Patent Document 1] JP 2012-203677 A Summary of the Invention [Problem to be solved by the invention]

[0007] As a technology for preventing workers from entering a dangerous area, the mainstream technology has been one that issues an alert after detecting an entry, such as that described in Patent Document 1. However, whether it is a laser barrier or a detection means using an image, neither method can detect 100%. In particular, in a case where an image is used to detect the entry of a worker in a vast construction site, the detection target (i.e., the worker) in the image becomes small, so that false detections tend to increase. In addition, in a situation where false detections occur frequently, the alerts become unreliable, and as a result, the workers gradually stop responding to the alerts. In other words, the conventional technology that issues an alert after detecting an entry is not sufficient to prevent workers from entering a dangerous area.

[0008] The object of the present invention is to solve the problems associated with the conventional technology, that is, to provide a risk assessment support device that can assess the risk level of a work area while, so to speak, accepting erroneous detections, rather than issuing an immediate alert. [Means for solving the problem]

[0009] The present invention was made with a focus on evaluating the degree of danger of a work area based on an index calculated from the danger of the work area and the number of times workers enter the area, and is an invention based on an unprecedented idea.

[0010] The risk assessment support device of the present invention is a device that evaluates the risk of a work area set in a target area based on local images of the target area captured periodically (or intermittently) by a photographing means installed at one location, and includes an area setting means, a risk indicator setting means, a worker detection means, an entry determination means, and an evaluation value calculation means. The area setting means is a means for setting one or more work areas within the field of view of the local image, the risk indicator setting means is a means for setting a risk indicator for the work area according to the work content related to the work area, and the worker detection means is a means for detecting a worker by performing image processing on the local image. The entry determination means is a means for determining whether or not a worker has entered the work area based on the worker and work area detected by the worker detection means and for counting the number of entries, and the evaluation value calculation means is a means for calculating a risk assessment value based on the product of the number of entries related to the work area and the risk indicator related to the work area. The risk of the work area is then evaluated based on the risk assessment value.

[0011] The risk assessment support device of the present invention may further include a feature point extraction means and a tracking means. The feature point extraction means is a means for extracting feature points from a part of the worker detected by the worker detection means, and the tracking means is a means for tracking the worker included in the successive on-site images based on the feature points. In this case, the entry determination means counts the number of entries as 1 when the worker exits the work area after determining that the worker has entered the work area.

[0012] The risk assessment support device of the present invention may further include a stay time calculation means. This stay time calculation means is a means for calculating the stay time that a worker stays in the same work area. In this case, the entry determination means acquires an entry time at which it is determined that the worker has entered the work area, and an exit time at which the worker exits the work area. Then, the stay time calculation means calculates the stay time in the work area based on the entry time and the exit time.

[0013] The risk assessment support device of the present invention may also extract a work area from an on-site image and perform image processing on the extracted work area from the on-site image.

[0014] The risk assessment support device of the present invention may further include a process input means for inputting the work content for each time period related to the work area. In this case, the risk index setting means sets a risk index for each time period according to the work content input by the process input means. Effect of the Invention

[0015] The risk assessment support device of the present invention has the following effects. (1) Since a comprehensive evaluation is performed, so to speak, using detection results over a specified period of time without issuing an immediate alert, it is possible to prevent the adverse effects of false detection, i.e., to reduce the decrease in reliability caused by operators. (2) Since the work area and hazard indicators can be set arbitrarily, it is possible to flexibly respond to changes in the crane layout, for example, and also to changes in the work content. (3) The danger level of the work area can be objectively evaluated based on the calculated danger evaluation value. As a result, it is possible to determine the appropriateness of the work area, review the placement of heavy machinery such as cranes, or install new safety passages, thereby creating a safer construction environment. [Brief description of the drawings]

[0016] [Figure 1] (a) is an image of a target area captured by a photographing means installed at one location, and (b) is an enlarged image of the target area including the work area. [Diagram 2] 1 is a block diagram showing a main configuration of a risk assessment support device according to the present invention; [Diagram 3] A screenshot showing a "work area" that has been set with a buffer around the construction machine. [Figure 4] A model diagram showing a "risk indicator table" with the elements "severity" and "likelihood." [Diagram 5] 1 is a model diagram showing a worker in which a worker frame is set, feature points are extracted, and a worker is tracked based on the feature points. [Figure 6] (a) is a model diagram that shows a schematic diagram of a worker frame that is completely contained within a work area, and (b) is a model diagram that shows a schematic diagram of a worker frame with part of it entering the work area. [Figure 7] FIG. 11 is a graph showing a risk assessment value that changes over time. [Figure 8] 3 is a flowchart showing a main process flow until the risk assessment support device of the present invention calculates a risk assessment value. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0017] An embodiment of the risk assessment support device of the present invention will be described with reference to the drawings. The risk assessment support device of the present invention can be used in situations where people enter areas that are considered to be dangerous, and can be used in a variety of cases, such as construction work, various facility construction work, and the establishment of event venues, for example. For convenience, however, an example of use in construction work will be described here.

[0018] 1.Overview FIG. 1(a) is an image of a target area (hereinafter simply referred to as the "target area") in a construction site captured by a photographing means (a so-called fixed camera) installed at one location, and FIG. 1(b) is an image of an enlarged portion of the target area. As shown in this figure, a "work area" is set in the target area. A work area here refers to an area that is dangerous for workers to enter, such as the area around a backhoe during excavation, the area around a crane during lifting work, or the area around temporarily stacked materials. Therefore, one work area or two or more work areas can be set in the target area.

[0019] As shown in Figure 1, the image captured by the image capture means (hereinafter referred to as "on-site image") includes the target area, which includes one or more work areas. Then, using various conventional image processing techniques, a worker is detected, and whether or not the worker has entered the work area is determined, and the number of times the worker has entered within a predetermined period (several hours, one day, one week, etc.) (hereinafter simply referred to as "number of entries") is recorded.

[0020] Each work area is assigned an indicator of danger depending on the type of work (hereinafter simply referred to as a "danger indicator"). For example, a danger indicator indicating "particularly dangerous" is set for a work area where there is a risk of being caught in heavy machinery or a suspended load falling, while a danger indicator indicating "slightly dangerous" is set for a work area related to welding work or temporary storage of lightweight materials.

[0021] When a danger index is set for a work area and the number of entries into that work area is recorded, a "risk evaluation value" is calculated, which is the product of the danger index and the number of entries. This risk evaluation value is a kind of KPI (Key Performance Indicator) in safety management, and is a value that can evaluate the danger of a work area. As a result, for work areas with high risk evaluation values, the number of entries by workers can be reduced by reviewing the location (e.g., placement of heavy machinery), installing safety fences, and ensuring appropriate safety passages, thereby reducing the risk evaluation value (KPI).

[0022] 2. Risk assessment support device Next, the risk assessment support device 100 of the present invention will be described in detail. Fig. 2 is a block diagram showing the main configuration of the risk assessment support device 100 of the present invention. As shown in this figure, the risk assessment support device 100 of the present invention includes area setting means 101, risk index setting means 102, worker detection means 103, entry determination means 104, and evaluation value calculation means 105, and can also include process input means 106, feature point extraction means 107, tracking means 108, stay time calculation means 109, photography means 110, on-site image storage means 111, etc.

[0023] Among the main elements constituting the risk assessment support device 100, the area setting means 101, risk index setting means 102, worker detection means 103, entry determination means 104, evaluation value calculation means 105, process input means 106, feature point extraction means 107, tracking means 108, and stay time calculation means 109 can be manufactured as dedicated devices, or a general-purpose computer device can be used. This computer device is equipped with a processor such as a CPU, memories such as ROM and RAM, input means such as a mouse and a keyboard, and a display, and can be configured as a personal computer, a server, a tablet PC such as an iPad (registered trademark), a mobile terminal including a smartphone, or the like. In other words, the processing of these various means is performed by making the computer device execute arithmetic processing according to a predetermined program. The photographing means 110 can also be prepared as a dedicated device, or one included in the computer device can be used.

[0024] The local image storage means 111 can use a storage device of a general-purpose computer, or can be constructed in a database server. When constructed in a database server, it can be placed in a local network (LAN: Local Area Network) or can be a cloud server that stores data via the Internet.

[0025] Hereinafter, each of the main elements constituting the risk assessment support device 100 of the present invention will be described.

[0026] (Method of filming) The photographing means 110 can acquire "local images" periodically, intermittently, or continuously, and can be, for example, a digital camera or a digital video camera. When a digital camera is used as the photographing means 110, it is controlled to acquire so-called "continuous still images", and when a digital video camera is used, each single image constituting a video is treated as a local image. The frequency at which the photographing means 110 acquires local images can be designed to be, for example, 24 to 60 frames per second (i.e., 24 to 60 fps), but of course it can also be designed to be less than 24 frames per second, or 61 frames per second or more. The local images acquired by the photographing means 110 are stored in the local image storage means 111 (FIG. 2).

[0027] The imaging means 110 is installed at one location and used as a so-called "fixed-point camera." The imaging means 110 is also positioned so that it can capture the target area including the work area. For example, in FIG. 1, the imaging means 110 is installed at a high position so that it can overlook the target area.

[0028] (Area setting method) The area setting means 101 is a means for setting one or more "work areas" within the field angle of the on-site image (i.e., the target area). This area setting means 101 can be designed so that an operator can set it using a pointing device (mouse, touch panel, pen tablet, touch pad, track pad, track ball, etc.), a keyboard, etc. Specifically, while checking the on-site image displayed on a display or the like, the operator operates the pointing device to generate a polygon surrounding, for example, a crane, and sets this polygon as the work area.

[0029] When setting up a work area, it is a good idea to leave a certain amount of margin (buffer). For example, in Figure 3, "Work Area 1" is set based on the backhoe excavating, but a buffer is provided around the backhoe. Similarly, "Work Area 2" is set up with a buffer around the crane, and "Work Area 3" is set up with buffers around the dump truck for transporting soil and the backhoe loading soil.

[0030] (Means for setting risk indicators) The risk indicator setting means 102 is a means for setting a "risk indicator" for the work area set by the area setting means 101. When two or more work areas are set in the target region, a unique risk indicator is set for each work area. As described above, this risk indicator is a value that is set according to the work content performed in the work area, and is, so to speak, an index that indicates the risk of the work content.

[0031] A number of levels of risk indexes are prepared in advance, and a correspondence table (hereinafter referred to as a "risk index table") with elements of "severity" and "possibility" can be prepared as shown in FIG. 4. For example, for a work area where excavation work is being performed by a backhoe, the severity can be set to "fatal / serious" and the possibility can be set to "possible", so the risk index for that work area can be set to "3". Alternatively, for a work area where welding work is being performed, the severity can be set to "mild" and the possibility can be set to "possible", so the risk index for that work area can be set to "1". Of course, risk indexes can be prepared in any number of stages, not limited to the three stages shown in FIG. 4, and can also be prepared based on elements other than (or in addition to) "severity" and "possibility".

[0032] When setting a danger indicator for a work area, the danger indicator can be input directly while checking the work related to the work area. Alternatively, the danger indicator can be set according to a previously prepared "work-indicator table (a correspondence table showing the relationship between work content and danger indicator)." In this case, the operator inputs the work content as text by operating the keyboard, or selects an appropriate work content from multiple types of work content presented by operating the pointing device, and reads out and sets the danger indicator by referring to the work content in the work-indicator table.

[0033] However, even in the same work area, the work content may differ depending on the time of day. Reinforcement work may be performed within the formwork in the morning, and a pump truck may be set up to pour concrete in the afternoon. Or, lifting work using a crane may be scheduled from the start of work that day until 3:00 p.m., but the crane may be removed after that and no work is scheduled to be performed. If the work content differs even in the same work area, it is desirable to set hazard indicators according to the respective work content, in other words, according to the time period during which the work content is performed.

[0034] In setting risk indicators according to time periods, the process input means 106 may be used. The process input means 106 is a means for inputting a combination of work content and time periods scheduled to be performed in the work area (hereinafter referred to as "scheduled work data"), and more specifically, a means for an operator to input multiple pieces of scheduled work data by operating a keyboard or the like. In this case, the risk indicator setting means 102 reads out the scheduled work data, consults the work-indicator table for each time period, and sets a risk indicator according to the time period. The entry determination means 104, which will be described later, counts the "number of entries" for a predetermined period such as several hours, one day, or one week (hereinafter referred to as the "counting period"). Therefore, it is preferable to input multiple pieces of scheduled work data using the process input means 106, with the counting period being one unit.

[0035] (Worker detection means) The worker detection means 103 is a means for detecting a worker from an on-site image acquired by the photographing means 110. When detecting a worker from an on-site image, various conventionally used image processing techniques, such as automatic detection by machine learning, can be used. The worker detection means 103 can be configured to detect a worker from the entire on-site image, or to extract a work area and detect a worker only from that work area (or the work area and its surroundings). By detecting only the extracted work area, it is possible to reduce the calculation time and calculation cost involved in detecting a worker. Furthermore, when the worker detection means 103 detects a worker from an on-site image, a screenshot of the on-site image can be stored in the on-site image storage means 111.

[0036] When the worker detection means 103 detects a worker, a frame (hereinafter referred to as a "worker frame") that surrounds the worker can be set as shown in FIG. 5 (top). In addition, by using the feature point extraction means 107, feature points can be extracted from the worker as shown in FIG. 5 (middle). When the feature point extraction means 107 extracts the feature points of the worker, the operator can manually assign the feature points while checking the worker included in the on-site image, or the feature points can be extracted by various conventional image processing techniques such as automatic detection by machine learning. When the feature points can be extracted in this way, the same worker can be tracked between temporally consecutive on-site images by using the tracking means 108 as shown in FIG. 5 (bottom). When the tracking means 108 tracks the worker, various conventional image processing techniques such as optical flow can be used.

[0037] (Entry judgment means) The entry determination means 104 is a means for determining whether or not a worker detected by the worker detection means 103 has entered the work area set by the area setting means 101 by comparing the worker, and for counting the number of entries (number of times that it has been determined that an entry has occurred) during the collection period. When determining that "a worker has entered the work area" (hereinafter simply referred to as "entry determination"), it is possible to use specifications that require the worker frame to be completely contained within the work area as shown in Fig. 6(a), or to use specifications that require the worker frame to be partially within the work area as shown in Fig. 6(b).

[0038] (Means for calculating staying time) The on-site image captured by the image capturing means 110 has the time of capture recorded therein, and therefore it is possible to know the time of capture of the on-site image used for the entry judgment, i.e., the time at which the entry judgment was made by the entry judgment means 104. Similarly, it is possible to know the time at which the entry judgment means 104 judged that "the worker has not entered the work area" (hereinafter simply referred to as "outside area judgment").

[0039] The staying time calculation means 109 is a means for calculating the time each time the worker stayed in the work area (hereinafter simply referred to as "staying time") by using the time determined by the entry determination means 104. Specifically, the staying time is calculated from the time when the entry determination means 104 performed the entry determination (hereinafter referred to as "entry time") to the time when the entry determination means 104 performed the outside area determination (hereinafter referred to as "exit time"), and more specifically, the time when the first entry determination was made after the exit time related to the worker was recorded is set as the "entry time", and the time when the first outside area determination was made after successive entry determinations is set as the "exit time", and then the "staying time" is calculated.

[0040] The stay time calculated by the stay time calculation means 109 means the time that the worker continuously stayed in the same work area. Therefore, the entry determination means 104 should count the time from the entry time to the exit time as one entry count. Furthermore, if the stay time calculated by the stay time calculation means 109 is extremely short, it is possible that the worker actually stopped by the work area for just a moment, but it is also possible that the worker was erroneously detected by the worker detection means 103. Therefore, it is possible to set a threshold value for the stay time (hereinafter simply referred to as the "stay threshold"), and to specify that if the stay time does not exceed this stay threshold, it is not counted as an entry count.

[0041] (Evaluation value calculation means) The evaluation value calculation means 105 is a means for calculating a "risk evaluation value KPI" for each work area. Note that this risk evaluation value KPI can be calculated by "Formula 1" using the risk indicator RK and the number of entries EN. KPI=RK×EN...Formula 1 Since the number of entries EN changes over time, the risk assessment value KPI also changes over time, as shown in FIG.

[0042] When multiple types of scheduled work data are input during the aggregation period by the process input means 106, it is advisable to find a risk assessment value for each scheduled work data (hereinafter, specifically referred to as "hourly risk assessment value") and then calculate the risk assessment value by aggregating these hourly risk assessment values. Specifically, the hourly risk assessment value is calculated as the product of the risk index related to the scheduled work data and the number of entries per time period related to that scheduled work data, and all the hourly risk assessment values ​​related to the input scheduled work data are aggregated to calculate the risk assessment value. That is, in this case, the assessment value calculation means 105 can calculate the risk assessment value by "Equation 2." Note that the "subscript" in the equation i " indicates the number of types of planned work data, and is a KPI i is the hourly risk assessment value. KPI = Σ(KPI i )=Σ(RK i ×EN i )...Equation 2

[0043] In addition, when counting the time from entry time to exit time as one entry, the evaluation value calculation means 105 can calculate the risk evaluation value after setting a coefficient according to the stay time (hereinafter referred to as "stay coefficient CF"). That is, the longer the stay time of an entry count, the larger the stay coefficient CF is set, and the risk evaluation value is calculated by "Equation 3". Risk evaluation value = CF × RK × EN Formula 3

[0044] The risk evaluation value calculated by the evaluation value calculation means 105 indicates the "risk" of the work area. In other words, the higher the risk evaluation value, the more dangerous the work area is, and conversely, the lower the risk evaluation value, the safer the work area is. For work areas with high risk evaluation values, measures can be taken to reduce the number of times workers enter the area by reviewing the location (e.g., placement of heavy machinery), installing safety fences, and ensuring appropriate safety passages. In addition, the effectiveness of the measures can be determined by monitoring the risk evaluation value after the measures have been taken.

[0045] (Processing flow) Below, the main processing of risk assessment support device 100 of the present invention will be described in detail with reference to Fig. 8. Fig. 8 is a flow diagram showing the flow of main processing up to the calculation of a risk assessment value by risk assessment support device 100 of the present invention, in which the central column shows the processing to be executed, the left column shows information necessary for that processing, and the right column shows information resulting from that processing.

[0046] When the risk assessment support device 100 calculates a risk assessment value, first, an on-site image is acquired by the photographing means 110 (Step 201 in FIG. 8). Then, one or more work areas are set within the target area by the area setting means 101 (Step 202 in FIG. 8). The photographing means 110 periodically (or intermittently or continuously) acquires on-site images, and it is advisable to use a representative on-site image among them when setting the work area. Once the work area is set, the work content in the work area is input (Step 203 in FIG. 8). At this time, as already mentioned, multiple types of scheduled work data can be input during the aggregation period by the process input means 106. Once the work content in the work area is input, a risk indicator corresponding to the work content is set for the work area by the risk indicator setting means 102 (Step 204 in FIG. 8).

[0047] Once a series of preparations have been completed up to this point, the worker detection means 103 detects a worker from on-site images captured periodically (or intermittently or continuously) by the imaging means 110 (Step 205 in FIG. 8). When a worker is detected from the on-site images, the feature point extraction means 107 extracts feature points of the worker (Step 206 in FIG. 8), and the tracking means 108 tracks the same worker between temporally consecutive on-site images (Step 207 in FIG. 8).

[0048] When a worker is detected by the worker detection means 103 and tracked by the tracking means 108, the entry determination means 104 performs an "entry determination / outside area determination" and counts the number of entries (Step 208 in FIG. 8). The stay time calculation means 109 calculates the stay time for one stay (Step 209 in FIG. 8), and sets a stay coefficient according to the stay time (Step 210 in FIG. 8). Then, when the necessary information has been obtained by the processes up to this point, the evaluation value calculation means 105 calculates a risk evaluation value for the work area (Step 211 in FIG. 8). [Industrial Applicability]

[0049] The risk assessment support device of the present invention can be used in various cases where people may enter dangerous areas, such as civil engineering and architectural works, various facility construction works, construction of event venues, etc. According to the present invention, various construction works can be carried out while ensuring the safety of workers, which leads to a reduction in industrial accidents, and therefore the invention can be said to be an invention that can be expected to not only be used industrially but also to make a great contribution to society. [Explanation of symbols]

[0050] 100 Risk assessment support device of the present invention 101 (Risk assessment support device) area setting means 102 (Risk assessment support device) risk index setting means 103 (Hazard Assessment Support Device) Worker Detection Means 104 (Danger Assessment Support Device) Entry Judgment Means 105 (Risk assessment support device) assessment value calculation means 106 (Hazard Assessment Support Device) Process Input Means 107 (Risk assessment support device) feature point extraction means 108 (Risk Assessment Support Device) Tracking Means 109 (Risk assessment support device) Stay time calculation means 110 (Risk Assessment Support Device) Photographing Means 111 (Risk Assessment Support Device) On-site Image Storage Means

Claims

1. An apparatus for evaluating a degree of danger of a work area set in a target area based on on-site images of the target area periodically or intermittently captured by an imaging means installed at a certain location, comprising: an area setting means for setting one or more of the work areas within a field angle of the on-site image; a risk indicator setting means for setting a risk indicator for the work area according to the work content related to the work area; a worker detection means for detecting a worker by performing image processing on the on-site image; an entry determination means for determining whether or not the worker has entered the work area based on the worker detected by the worker detection means and the work area, and for recording the number of entries that are determined to have been made; an evaluation value calculation means for calculating a risk evaluation value based on a product of the number of entries into the work area and the risk index for the work area, The danger level of the work area can be evaluated based on the danger level evaluation value. A risk assessment support device comprising:

2. a feature point extraction means for extracting feature points from a part of the worker detected by the worker detection means; a tracking means for tracking the worker included in the successive on-site images based on the feature points, the entry determination means counts the number of entries as 1 when the worker exits the work area after determining that the worker has entered the work area; 2. The risk assessment support device according to claim 1,

3. A staying time calculation means for calculating a staying time during which the worker stays in the same work area is further provided, The entry determination means acquires an entry time when it is determined that the worker has entered the work area, and acquires an exit time when the worker has exited the work area, The staying time calculation means calculates the staying time in the work area based on the entry time and the exit time, the evaluation value calculation means calculates the risk evaluation value based on a product of a coefficient corresponding to the staying time, the number of entries, and the risk index; 2. The risk assessment support device according to claim 1,

4. the worker detection means extracts the work area from the on-site image, and performs image processing on the extracted work area from the on-site image; 2. The risk assessment support device according to claim 1,

5. A process input means for inputting the work content for each time period related to the work area is further provided, the risk index setting means sets the risk index for each time period according to the work content inputted by the process input means; 2. The risk assessment support device according to claim 1,

Images