how to normalise latitude into the range of -90 to 90?

Question

Using piexif I got the DMS of Latitude and longitude which I am trying to convert in decimal degree, but for some images I am getting lattitude value as 184.62583333333333 which is out of the [-90,90] range.

check out the code below,

exif_dict = piexif.load('images/DJI_0026.JPG')
long = 0
latt = 0
value = exif_dict['GPS']
if value:
    lat = value[2]
    lon = value[4]
    for i in range(3):
        if i == 1:
            latt += lat[i][0]/60.0
        elif i == 2:
            latt += lat[i][0]/3600.0
        else:
            latt += lat[i][0]
    for i in range(3):
        if i == 1:
            long += lon[i][0]/60.0
        elif i == 2:
            long += lon[i][0]/3600.0
        else:
            long += lon[i][0]
print(latt, long)

value = {0: (2, 3, 0, 0), 1: b'N', 2: ((19, 1), (8, 1), (595773, 10000)), 3: b'E', 4: ((73, 1), (0, 1), (131775, 10000)), 5: 0, 6: (70989, 1000)}

I am concerned with latitude and longitude, that is stored in value of key 2 and 4.

latitude = 19+8/60.0+595773/3600.0

longitude = 73+0/60.0+131775/3600.0

that is what the output is.

OutPut: 184.62583333333333 109.60416666666666

Please let me know how to normalise the latitude in the range of [-90,90].

Answer 1

The data as returned by piexif for the GPS coordinates has the following format:

exif_data = {0: (2, 3, 0, 0),
             # Latitude: b'N' or b'S'
             1: b'N', 
             # deg, min, sec as (numerator,denominator) of rationals
             2: ((19, 1), (8, 1), (595773, 10000)),
             # Longitude: b'E' or b'W'
             3: b'E', 
             4: ((73, 1), (0, 1), (131775, 10000)), 
             5: 0, 6: (70989, 1000)}

Latitude and longitude are given as positive, rational values, towards N or S (resp. E or W).

We need to convert the positive values from DMS to decimal, then give them the right sign depending on the direction:

def convert_DMS_tuple(tup):
    d, m, s = [t[0]/t[1] for t in tup]
    degrees = d + m/60 + s/3600
    return degrees

def EXIF_to_lat_long(exif_data):
    # Latitude is counted positive towards N
    lat_sign = 1 if exif_data[1] == b'N' else -1
    lat = lat_sign*convert_DMS_tuple(exif_data[2])

    # Longitude is counted positive towards E
    long_sign = 1 if exif_data[3] == b'E' else -1
    long = long_sign*convert_DMS_tuple(exif_data[4])

    return lat, long

EXIF_to_lat_long(exif_data)
# (19.149882583333333, 73.00366041666666)

You can use this like:

exif_dict = piexif.load('images/DJI_0026.JPG')
value = exif_dict['GPS']
if value:
   print(EXIF_to_lat_long(value))

Answer 2

From the example values you gave, it looks obvious that you should divide val[0] by val[1].

For example, 595773 / 10000 = 59, which makes perfect sense for a second.

By the way, you don't need the for loops. They make your code longer than necessary.

There's probably a reason why value[1] == b'N' . Your current code does not evaluate it though. This means that your code only works for 1/4 of the Earth's surface.

See https://sno.phy.queensu.ca/~phil/exiftool/TagNames/GPS.html for a good overview. Your code should be able to interpret South and West as well.