Restricted　by　the　environment　and　hardware　equipment　resources,　existing　chaotic　systems　have　shortcomings　such　as　low　complexity,　low　randomness,　and　chaotic　degradation　phenomena,　which　in　turn　cause　the　security　risks　of　chaotic　image　encryption　algorithms.　To　overcome　these　issues,　this　paper　proposes　a　method　for　the　construction　of　a　LE-controlled　four-dimensional　(4D)　non-degenerate　discrete　chaotic　system.　Numerical　analysis　has　demonstrated　that　the　developed　system　possess　high　complexity　and　unpredictability.　Based　on　the　developed　chaotic　system,　an　image　compression　encryption　algorithm　is　proposed.　Wherein,　semi-tensor　product　compressed　sensing　is　applied　to　allow　data　compression　sampling　in　different　dimensions　resulting　in　reducing　the　data　transfer　load　and　storage　cost.　Subsequently,　the　positions　and　values　of　the　image　pixels　are　secretly　altered　during　the　algorithmic　encryption　process　using　two-dimensional　cat　confusion　and　finite　field　diffusion.　The　simulation　results　show　that　the　proposed　encryption　algorithm　effective　enough　to　offer　great　encryption　quality.　The　performance　comparison　analysis　indicates　that　the　proposed　encryption　algorithm　has　good　furnishes　better　security,　compression,　as　well　as　resistance　to　diverse　data　attacks.