Background
The image to the right is that of Luton Borough Police Force taken in November 1920. My great grandfather Detective Sergeant Arthur Bacon is sitting to the far right in the first row, with walking stick, overcoat and hat.
Many years back I remember my grandfather George Bacon telling me that the photograph was taken in the garden of a house at the bottom of Stockingstone Road in Luton. He wasn’t much more specific than that. My mother also echoed the same opinion about the photographs location. Both were not really sure of where it was taken, only that the house backed onto Wardown Park, somewhere.
In about 2015 I attended a local lecture by an ex-Luton police officer who had on display various memorabilia and photographs that he had collected over the years, including this particular photo. I asked him where he thought the location was. He told me that all the police photo’s were usually taken at the cricket pavilion in Wardown Park. Not really convinced by his answer and armed with the comments by my family about the photo, I set about attempting to geolocate the image myself.
Initial clues
I examined the photo quite closely in an attempt to gather some clues in order to get started. There are principally three interesting features in the image:
-
To the upper far left there is what appears to be a corrugated wooden shed.
-
The policemen seem to be sitting over a slight left-right downward gradient. The photograph itself is not misaligned in terms of framing. That being the case I would be able to do a similar analysis as I used with the Unknown Luton House.
-
To the mid far right is a feature that looks like the top of a chimney stack. If the photograph had been taken from the bottom of Stockingstone Road, then the camera would be facing Southwards out over Wardown Park area. The most significant landmark, which was just to the South of the park, could only have been Lye’s dye works chimney situated on Old Bedford Road.
Verifying the clues
The wooden shed
This was fairly easy. The candidate location, i.e. “A house at the bottom of Stockingstone Road”, could only be North Lodge, which is situated on the corner of Old Bedford Road and Stockingstone Road. A quick look at Google Street View produced the image to the right.
Lo and behold, in the garden of North Lodge is a small wooden building with a corrugated roof on it. I tried panning around in Street View to see if I could find any evidence of finials on the roof ends as there is in the original photograph. No luck really as the end of the shed is covered with a thick layer of Ivy that has grown over it. That being said, it was a promising start, and the shed situated in approximately the correct position within the property boundary.
Finding this small wooden building, assuming it is the same one as in the original photo, was going to be vital for me in terms positioning the photograph within the garden of North Lodge.
Meanwhile back to the original photograph
My next set of thoughts moved to what was the length of ground covered by the photograph, would it even fit into the garden of North Lodge? For that I need not have worried, the garden is ample sized.
The next step was to scale the photograph in order to make an approximation of the total on-the-ground width of the scene. Fortunately, I actually have the walking stick that my great grandfather has in the photo and measuring it gave the length of 0.88m. Scaling the image given this known length, the total width of the scene in the photograph is 8.18m.
I did try to remove the lens distortion present in the photograph, although in this particular instance, did not succeed. That being said I’d have to live with the 8.18m length, give or take due to the curvature of the lense.
Armed with the ground image width of the scene and the position of the wooden building, it was time to position the photograph on the 1922 map of Luton.
Clearly, I had no option than to position the photograph by eye. In this type of work there often is little option. It is a question of making best use of the relatively small amount of information available at the time.
The ground gradient in the photograph
I extracted the 1m LiDAR elevation data along the section line of the photograph (the yellow line on the map above). I did the same for the photo itself, using the nose positions of the police men in the first row to give a set of sharp well defined points over the image.
The results from the elevation comparison is shown to the right. The slope of the two linear regression lines are similar. Doing a formal Bayesian statistical comparison between the two regression model slope parameters yielded a difference of 0.298. The 95% confidence interval for the slope estimate is between -2.298 and 3.086, hence the 0.298 difference between the two elevation slope’s is not statistically significant from zero.
To rephrase the above (in English), the net result of this analysis suggests that the apparent slope that the policemen are on in the photograph is the same as the land gradient obtained from the LiDAR data, i.e. it is probably the same location.
The chimney stack
Next I needed to locate where Lye’s dye works chimney used to stand. The chimney and works vanished a very long time ago. The present site has been redeveloped into housing and the road running through it is called Earls Mead.
By georeferencing a historical aerial photograph of the dye works site, I would be able to determine the coordinates of the base of the chimney, and project it onto the modern map.
My computed position for the Chimney is TL 08842 22291. The chimney is 905m to the South of North Lodge.
I was a little bit concerned when first I saw the composite map, the chimney looked a lot too far to the west than I had originally imagined, and I started to get second thoughts regarding my initial hypothesis about it being the dye works chimney in the photograph. That being said a further stage needed to be done.
The 3D scene reconstruction model
What I needed to do next was to create a simple 3D model of the photograph scene, so that I could gauge how the chimney position fitted in relation to the original photograph as viewed in perspective. So roughly the view that the camera man would have seen.
Much to my surprise, given the effects of perspective, the chimney did not look overly too far to the West as I originally suspected. Yes, it does not exactly overlap the chimney feature in the photograph and this is due to my positioning of the photograph by hand in the grounds of North Lodge, plus and distortion present in the photograph. Given the ill conditioned problem that I had set out to solve, I think the end result is a pretty good match.
Conclusion
After my analysis of this situation, I feel pretty confident that the police photograph was indeed taken in the grounds of North Lodge, Wardown Park, as originally suggested my relatives.