A forklift was moving steel coils from a dock into a marine-terminal warehouse. Inside the building, a coil slipped off a crane hook. The bridge-crane operator left his control station to see what had happened, walked among the stacked coils, and stepped into the area the forklift used to turn. The forklift operator lost sight of him. While reversing, the forklift struck and killed him.
OSHA recorded the fatality in Longshoring and Marine Terminals Fatal Facts, Section I, Summary No. 12. Its account describes a warehouse 500 feet long in which only the main aisle had a fixed width. The other aisles varied with how many coils were in storage, and no aisle anywhere in the building was marked or designated. OSHA lists the absence of designated safe aisle ways and forklift operating areas among the relevant factors, and notes that a forklift safety manual kept on site already called for marking forklift paths.[1]
The trip that put him in the turning area was unplanned. Nobody forecasts a dropped coil. What made that trip fatal was already in place: aisle positions that moved with the stock, sightlines closed by stacked coils, and no marked boundary anywhere in the building telling a person on foot where a thirty-foot forklift would reverse. The upset was the trigger. The standing condition was the exposure.
The incident is sudden. The condition that makes it possible usually is not.
The safe route can disappear without anyone changing the plan
Terminal warehouses, transit sheds and cargo-handling areas are not static. Steel, lumber, pallets and project cargo move through them. Staging boundaries shift. Loads narrow aisles and block sightlines. A route that was clear at the start of a shift may no longer be the route people actually use several hours later.
Picture one cross-aisle beside an inbound cargo area. A new load occupies part of the pedestrian path. Workers begin walking around it. The detour takes them through the space a forklift uses to reverse and turn. Nothing is formally changed. No incident is reported. But the same interface forms again each time cargo arrives.
The United Kingdom's Health and Safety Executive lists unsegregated vehicle and pedestrian access, reversing vehicles, movement of vehicles in cargo storage areas and the use of vehicles with limited visibility among the typical workplace transport hazards in ports.[2]
A safe aisle on the plan is not necessarily the path people are using. That is the operating reality a camera can help make visible.
Use the camera to see the pattern, not only the incident
For terminal safety and operations leaders, the practical question is not whether a camera can detect a forklift. It is whether existing views can reveal a recurring shared-space condition early enough for the site to act.
Most terminal cameras are used for live viewing or investigation after something happens. Edgentik uses selected existing views to answer a different question:
Where do forklifts and pedestrians keep entering the same space before an incident occurs?
Starts after something happens
A reported incident, complaint or manual search sends someone back through footage.
Shows the condition while it is recurring
Repeated observations become reviewed episodes that show where and when the same exposure keeps returning.
What the terminal team receives
Receiving cross-aisle - Aisle 7
Illustrative worked example using recorded third-party industrial footage. Baseline counts and timing were measured; the site context and follow-up are illustrative. Not a customer result.
Baseline rate: 0.35 reviewed episodes per usable hour.
12 episodes across 14 usable hours, versus 0.27 per usable hour during the other 82 measured hours.
Estimated 95% range: 1.6x-6.5x, based on 12 peak-window episodes.
Follow-up: 17 / 96 hours = 0.18
The same camera, area, condition and review method were used. This compares one baseline week with the following week; it cannot separate the site action from normal week-to-week operating variation.
This is the output: one area, one recurring pattern and one illustrative matched-basis follow-up after action.
Edgentik measures the condition in a defined area. It does not identify individuals or score operators and workers. The unit of a finding is a place and a time, not the person in the frame.
From a changing route to a measurable safety decision
Staged cargo narrows the normal route. Pedestrians begin crossing the forklift turning area.
Reviewed episodes keep returning at the same cross-aisle during inbound handling.
The site changes staging, separation, access or timing, then measures the same area again.
This is Edgentik
The site condition overview brings the work into three operational questions: what is happening in the latest measured period, what keeps recurring, and what was measured after the site acted.
A follow-up measurement is not proof that the site action caused the difference. Cargo mix, volume, staffing and season can also affect the measured rate, and Edgentik reports the result whether it falls, holds or rises. Measuring a condition does not prevent an incident, and nothing here suggests a camera would have saved the worker in the OSHA case.
The customer does not buy a stream of AI alerts. The customer gets a site condition, reviewed evidence, a decision point and a follow-up.
What a focused terminal pilot delivers
A terminal does not need a port-wide deployment to begin. Start with a warehouse doorway, cross-aisle, staging boundary or loading interface already covered by a suitable fixed camera.
Suitable, unsuitable or a recommendation for a better view.
Reviewed forklift–pedestrian episodes across agreed shifts, hours or types of work.
The location and operating period where the shared-space pattern is strongest.
Reviewable episodes supporting the finding.
The control, staging change or operating decision selected by the site.
The area measured again under similar operating conditions after action.
The response may be a protected walkway, a pedestrian gate, a designated crossing, revised staging, one-way travel, tighter access control or a change in when the work is performed. A feature in WorkSafeBC's WorkSafe Magazine describes B.C. sites layering controls of this kind through engineering separation, protected routes and controlled crossings.[3]
Edgentik records the site's decision and reports what the follow-up measured. The site decides what caused the change and what to do next.
Camera fit is decided before the pilot begins. If the selected view cannot reliably see the condition, Edgentik recommends a better view or stops there.
Look at one terminal camera differently
Choose a camera covering a warehouse doorway, cross-aisle, staging boundary or loading interface. Then ask:
- Where do people step around staged cargo or leave the intended route?
- Where do forklifts reverse or turn with limited sightlines?
- Does the same shared-space condition return during inbound handling, loading or shift change?
- After the site changes the route or separation, does the condition still form?
That is the practical opportunity. Use cameras that already see the work to find a recurring condition before the next incident forces the site to investigate it.
Start with one area where forklifts and people repeatedly share space.
Send us one existing camera view. Edgentik will assess whether the condition is measurable and return a focused pilot scope covering the area, reviewed clips, baseline and same-area follow-up.
Sources
- OSHA: Longshoring and Marine Terminals Fatal Facts, Section I: Vehicular Accidents, Summary No. 12, Forklift Truck Operations
- Health and Safety Executive: Workplace Transport in Ports and Docks
- WorkSafeBC: Working Safely Around Mobile Equipment, WorkSafe Magazine, Spring 2026
Ashish Syal
Founder of Edgentik. Edgentik uses suitable existing cameras to reveal recurring forklift–pedestrian conditions and measure the same condition again after the site acts.
About Edgentik →